Add mapped dataset README files

Upload README files mapped in the local coverage report to dataset-root README.md paths.

Surgical/cmr_surgical/cholecystectomy/README.md

@@ -0,0 +1,247 @@
+<!--
+Open-H Embodiment Dataset README Template (v1.0)
+Please fill out this template and include it in the ./metadata directory of your LeRobot dataset.
+This file helps others understand the context and details of your contribution.
+-->
+
+# CMR Surgical Versius Surgical Robot Dataset - README
+
+---
+
+## 📋 At a Glance
+
+*This dataset contains a variety of in-vivo surgeries performed with a Versius surgical system. The Versius system by CMR Surgical is a modular soft-tissue surgical robotic system.*
+*Versius can be used with up to 4 robotic arms simultaneously, and supports a variety of surgical instruments.* 
+<div align="center">
+    <img src="./images/CMR_Versius_System.png"  alt="Versius system used in surgery." width="600" />
+</div>
+
+*During surgery, each robotic arm is distinguishable by its LED band color. The arm with a white LED band is a Visualisation Arm and holds the endoscopic camera.*
+
+<div align="center">
+    <img src="./images/CMR_Versius_System_Console_Arms.png"  alt="Versius system used in surgery." width="600" />
+</div>
+
+*Versius is controlled from the Surgeon Console, where two 7 degrees of freedom hand controllers capture the surgeon's hand movements to enable precise control of the instruments attached to the robot arm.*
+
+<div align="center">
+<img src="./images/CMR_Surgeon_Control.png" alt="Versius console controlled by a surgeon." width="600"/>
+</div>
+
+
+
+*Each hand controller has an electro-surgery activation button (1), a clutch button (2), a thumbstick (3) and a trigger (also known as pince) (4).*
+
+
+<div align="center">
+    <img src="./images/CMR_Hand_Controller_labeled.png"  alt="Versius Hand Controller labeled." width="600" />
+</div>
+
+
+---
+
+## 📖 Dataset Overview
+
+This dataset contains video and telemetry for 399 surgeries, split into 99 Cholecystectomies, 100 Hysterectomies, 100 Hernia repairs, and 100 Prostatectomies. The video is taken from an endoscopic camera attached to one of the Versius arms. The video is only recorded when inside the patient to ensure it is anonymous.
+
+
+| | |
+| :--- | :--- |
+| **Total Trajectories** | `N/A` |
+| **Total Hours** | `485 hours` |
+| **Data Type** | `Clinical` |
+| **License** | CC BY 4.0 |
+| **Version** | `[1.0]` |
+
+---
+
+## 🎯 Tasks & Domain
+
+### Domain
+
+- [X] **Surgical Robotics**
+- [ ] **Ultrasound Robotics**
+- [ ] **Other Healthcare Robotics** (Please specify: `[Your Domain]`)
+
+### Demonstrated Skills
+
+The videos in this dataset demonstrate succesful completions of cholecystectomy, prostatectomy, hysterectomy, and inguinal hernia repair surgeries.
+
+---
+
+## 🔬 Data Collection Details
+
+### Collection Method
+
+- [X] **Human Teleoperation**
+- [ ] **Programmatic/State-Machine**
+- [ ] **AI Policy / Autonomous**
+- [ ] **Other** (Please specify: `[Your Method]`)
+
+
+### Operator Details
+
+| | Description |
+| :--- | :--- |
+| **Operator Count** | `N/A` |
+| **Operator Skill Level** | `[X] Expert Surgeon` <br> `[ ] Intermediate (e.g., Trained Researcher)` <br> `[ ] Novice (e.g., ML Researcher with minimal experience)` <br> `[ ] N/A` |
+| **Collection Period** | From `[2025-01-01]` to `[2025-11-05]` |
+
+### Recovery Demonstrations
+
+This dataset is provided unlabeled, but will include instances of failure and recovery when completing a surgical task.
+
+- [X] **Yes**
+- [ ] **No**
+
+---
+
+## 💡 Diversity Dimensions
+
+- [ ] **Camera Position / Angle**
+- [ ] **Lighting Conditions**
+- [X] **Target Object** (Different procedure types)
+- [X] **Spatial Layout** (e.g., placing the target suture needle in various locations)
+- [ ] **Robot Embodiment** (if multiple robots were used)
+- [X] **Task Execution** (e.g., different techniques for the same task)
+- [X] **Background / Scene**
+- [ ] **Other** (Please specify: `[Your Dimension]`)
+
+The target objects and tasks were varied by selecting surgeries of different procedure types. Camera angles and background naturally change depending on the patient and procedure. The background will mostly be patient anatomy, except when the 
+endoscope is close to its trocar - in these situations the view will be occluded by the trocar.
+
+The spatial layout is mostly constant, with at least one surgical instrument to either side of the camera. Most often two instruments will be in the field of view of the camera.
+
+
+---
+
+## 🛠️ Equipment & Setup
+
+### Robotic Platform(s)
+
+
+- **Robot:** `Versius Surgical System`
+
+Depending on the surgery, anywhere between 1-4 robotic arms are used simultaneously.
+
+
+### Sensors & Cameras
+
+| Type | Model/Details |
+| :--- | :--- |
+| **Primary Camera** | `Endoscopic Camera, 1920x1080 @ 60fps` |
+
+---
+
+## 🎯 Action & State Space Representation
+
+The actions in the dataset map directly to all the inputs available to a surgeon on a Versius system.
+The actions map to the buttons in two hand controllers (left and right), and include:
+1. Orientation quaternion (a,b,c,w)
+2. Position (x,y,z)
+2. Pince
+3. Clutch Button
+4. Energy Button
+5. Thumbstick Button
+6. Joystick X and Joystick Y
+
+The movement of the hand controllers maps to the camera frame.
+
+### Action Space Representation
+
+**Primary Action Representation:**
+- [X] **Absolute Cartesian** (position/orientation relative to robot base)
+- [ ] **Relative Cartesian** (delta position/orientation from current pose)
+- [ ] **Joint Space** (direct joint angle commands)
+- [ ] **Other** (Please specify: `[Your Representation]`)
+
+**Orientation Representation:**
+- [X] **Quaternions** (x, y, z, w)
+- [ ] **Euler Angles** (roll, pitch, yaw)
+- [ ] **Axis-Angle** (rotation vector)
+- [ ] **Rotation Matrix** (3x3 matrix)
+- [ ] **Other** (Please specify: `[Your Representation]`)
+
+**Reference Frame:**
+- [ ] **Robot Base Frame**
+- [ ] **Tool/End-Effector Frame**
+- [ ] **World/Global Frame**
+- [X] **Camera Frame**
+- [ ] **Other** (Please specify: `[Your Frame]`)
+
+**Action Dimensions:**
+```
+action: [x_left, y_left, z_left, quat_x_left,... quat_z_left, clutchBtn_left, energyBtn_left, thumbstickBtn_left, pince_left, thumbstick_x_left, thumbstick_y_left, ... (repeated for right)]
+
+- x,y,z: [float[3]] Absolute position of the hand controller (left or right) (metres)
+- quat_(x,y,z): [float[4]] Absolute orientation of the hand controller (left or right) (radians)
+- pince: [float from 0-1] value of the pince on the hand controller. Maps to the instrument jaws.
+- clutchBtn: [bool] Button input to clutch in/out. Engages and disengages a selected arm.
+- energyBtn: [bool] Button to activate electrosurgery.
+- thumbstickBtn: [bool] Button to activate ICG (Indocyanine green) visualisation.
+- thumbstick_x/y: [float[2]] Value of thumbstick - used to control the endoscope.
+```
+<div align="center">
+    <img src="./images/CMR_Versius_System_Dual_Hand_Controller.png"  alt="Versius Hand Controllers." width="600" />
+</div>
+
+*All actions come in `_left` and `_right` variants, which map to the left and right hand controllers.*
+
+### State Space Representation
+
+**State Information Included:**
+- [ ] **Joint Positions** (all articulated joints)
+- [ ] **Joint Velocities**
+- [ ] **End-Effector Pose** (Cartesian position/orientation)
+- [ ] **Force/Torque Readings**
+- [ ] **Gripper State** (position, force, etc.)
+- [X] **Other** (Please specify: `See below`)
+
+**State Dimensions:**
+
+For all state dimensions, the symbol `<S>` (Side) is used to denote a state that has `_left` and `_right` variants.
+```
+observation.state.haptic_<S>_armengageable: [bool[2]] Whether the arm selected by the left or right controller is in an engageable state. An arm may be enagageable without being actively engaged. Arms may be unengageable due to being in the wrong mode (e.g Instrument Change).
+
+observation.state.armlinkedtohaptic_<S>: [int[2]] Index of the arm connected to each hand controller at any one time.
+
+observation.state.hapticengaged_<S>: [bool[2]] Whether the haptic is in an engaged state. An engaged haptic means that its movements are currently driving a robotic arm.
+
+observation.state.arm_<int>_color: [int[5]] Color of each arm connected to the system. The color of each arm is shown on the HUD icons.
+
+observation.state.arm_<int>_instrtype: [int[5]] Instrument type attached to each arm.
+
+observation.state.rotationscaling: [float] Current scaling applied to the relative hand controller orientation. Can change during surgery.
+
+observation.state.translationScaling: [float] Current scaling applied to the relative hand controller position. Can change during surgery.
+
+observation.state.electroSurgerymode_<S>: [int[2]]: Electrosurgery mode (CUT, COAG) selected on each hand controller.
+
+observation.state.haptic_engaged_<S>: [bool[2]] Whether the haptic is engaged or not. When this state is false, the arms are not being controlled by the surgeon and are therefore not performing surgical tasks. The arms may still be moved by bedside assistants to retract the instrument or reposition the arm.
+
+observation.state.icgmode: [int] ICG (Indocyanine Green) mode selected (grayscale, overlay, heatmap).
+
+observation.state.icgenabled: [bool] ICG state, toggled with the right `thumbstickBtn`.
+observation.state.instrtype_<S>: [int[2]] Instrument attached to each hand controller (and in turn attached to an arm as per state.linkedArmIndex)
+```
+
+---
+
+## ⏱️ Data Synchronization Approach
+
+Versius telemetry and video is timestamped using the clock in its network controller. The network controller timestamps at the point of recording telemetry at 50Hz and video at 60Hz.
+Although both telemetry and video are timestamped using the same clock, there is a small variation in timing between the telemetry and video. As such, the system's HUD (Heads-Up Display) was used to synchronise telemetry and video.
+The timestamps of events in the HUD (including but not limited to: changes in arm mode, activation of electro-surgery and changes in electro-surgery mode) are compared against the timestamps of the corresponding events in telemetry. This
+provides us with a reliable measure of lag and the rate of dropped frames. The result of this synchronisation process is used to post-process the telemetry such that it is as close to perfectly in sync as possible.
+
+
+---
+
+## 👥 Attribution & Contact
+This dataset is shared with the aim of advancing research and development that has the potential to benefit patients through improved outcomes and increased access to minimal access surgery. 
+| | |
+| :--- | :--- |
+| **Dataset Lead** | Patrick Thornycroft, Diego Marana, Jonathon Hawkins, Filip Binkiewicz, Joyce Zhang |
+| **Institution** | CMR Surgical Ltd. |
+| **Contact Email** | patrick.thornycroft@cmrsurgical.com, diego.marana@cmrsurgical.com |
+| **Citation (BibTeX)** | N/A |

Surgical/cmr_surgical/dry_box/README.md

@@ -0,0 +1,247 @@
+<!--
+Open-H Embodiment Dataset README Template (v1.0)
+Please fill out this template and include it in the ./metadata directory of your LeRobot dataset.
+This file helps others understand the context and details of your contribution.
+-->
+
+# CMR Surgical Versius Surgical Robot Dataset - README
+
+---
+
+## 📋 At a Glance
+
+*This dataset contains a variety of in-vivo surgeries performed with a Versius surgical system. The Versius system by CMR Surgical is a modular soft-tissue surgical robotic system.*
+*Versius can be used with up to 4 robotic arms simultaneously, and supports a variety of surgical instruments.* 
+<div align="center">
+    <img src="./images/CMR_Versius_System.png"  alt="Versius system used in surgery." width="600" />
+</div>
+
+*During surgery, each robotic arm is distinguishable by its LED band color. The arm with a white LED band is a Visualisation Arm and holds the endoscopic camera.*
+
+<div align="center">
+    <img src="./images/CMR_Versius_System_Console_Arms.png"  alt="Versius system used in surgery." width="600" />
+</div>
+
+*Versius is controlled from the Surgeon Console, where two 7 degrees of freedom hand controllers capture the surgeon's hand movements to enable precise control of the instruments attached to the robot arm.*
+
+<div align="center">
+<img src="./images/CMR_Surgeon_Control.png" alt="Versius console controlled by a surgeon." width="600"/>
+</div>
+
+
+
+*Each hand controller has an electro-surgery activation button (1), a clutch button (2), a thumbstick (3) and a trigger (also known as pince) (4).*
+
+
+<div align="center">
+    <img src="./images/CMR_Hand_Controller_labeled.png"  alt="Versius Hand Controller labeled." width="600" />
+</div>
+
+
+---
+
+## 📖 Dataset Overview
+
+This dataset contains video and telemetry for 399 surgeries, split into 99 Cholecystectomies, 100 Hysterectomies, 100 Hernia repairs, and 100 Prostatectomies. The video is taken from an endoscopic camera attached to one of the Versius arms. The video is only recorded when inside the patient to ensure it is anonymous.
+
+
+| | |
+| :--- | :--- |
+| **Total Trajectories** | `N/A` |
+| **Total Hours** | `485 hours` |
+| **Data Type** | `Clinical` |
+| **License** | CC BY 4.0 |
+| **Version** | `[1.0]` |
+
+---
+
+## 🎯 Tasks & Domain
+
+### Domain
+
+- [X] **Surgical Robotics**
+- [ ] **Ultrasound Robotics**
+- [ ] **Other Healthcare Robotics** (Please specify: `[Your Domain]`)
+
+### Demonstrated Skills
+
+The videos in this dataset demonstrate succesful completions of cholecystectomy, prostatectomy, hysterectomy, and inguinal hernia repair surgeries.
+
+---
+
+## 🔬 Data Collection Details
+
+### Collection Method
+
+- [X] **Human Teleoperation**
+- [ ] **Programmatic/State-Machine**
+- [ ] **AI Policy / Autonomous**
+- [ ] **Other** (Please specify: `[Your Method]`)
+
+
+### Operator Details
+
+| | Description |
+| :--- | :--- |
+| **Operator Count** | `N/A` |
+| **Operator Skill Level** | `[X] Expert Surgeon` <br> `[ ] Intermediate (e.g., Trained Researcher)` <br> `[ ] Novice (e.g., ML Researcher with minimal experience)` <br> `[ ] N/A` |
+| **Collection Period** | From `[2025-01-01]` to `[2025-11-05]` |
+
+### Recovery Demonstrations
+
+This dataset is provided unlabeled, but will include instances of failure and recovery when completing a surgical task.
+
+- [X] **Yes**
+- [ ] **No**
+
+---
+
+## 💡 Diversity Dimensions
+
+- [ ] **Camera Position / Angle**
+- [ ] **Lighting Conditions**
+- [X] **Target Object** (Different procedure types)
+- [X] **Spatial Layout** (e.g., placing the target suture needle in various locations)
+- [ ] **Robot Embodiment** (if multiple robots were used)
+- [X] **Task Execution** (e.g., different techniques for the same task)
+- [X] **Background / Scene**
+- [ ] **Other** (Please specify: `[Your Dimension]`)
+
+The target objects and tasks were varied by selecting surgeries of different procedure types. Camera angles and background naturally change depending on the patient and procedure. The background will mostly be patient anatomy, except when the 
+endoscope is close to its trocar - in these situations the view will be occluded by the trocar.
+
+The spatial layout is mostly constant, with at least one surgical instrument to either side of the camera. Most often two instruments will be in the field of view of the camera.
+
+
+---
+
+## 🛠️ Equipment & Setup
+
+### Robotic Platform(s)
+
+
+- **Robot:** `Versius Surgical System`
+
+Depending on the surgery, anywhere between 1-4 robotic arms are used simultaneously.
+
+
+### Sensors & Cameras
+
+| Type | Model/Details |
+| :--- | :--- |
+| **Primary Camera** | `Endoscopic Camera, 1920x1080 @ 60fps` |
+
+---
+
+## 🎯 Action & State Space Representation
+
+The actions in the dataset map directly to all the inputs available to a surgeon on a Versius system.
+The actions map to the buttons in two hand controllers (left and right), and include:
+1. Orientation quaternion (a,b,c,w)
+2. Position (x,y,z)
+2. Pince
+3. Clutch Button
+4. Energy Button
+5. Thumbstick Button
+6. Joystick X and Joystick Y
+
+The movement of the hand controllers maps to the camera frame.
+
+### Action Space Representation
+
+**Primary Action Representation:**
+- [X] **Absolute Cartesian** (position/orientation relative to robot base)
+- [ ] **Relative Cartesian** (delta position/orientation from current pose)
+- [ ] **Joint Space** (direct joint angle commands)
+- [ ] **Other** (Please specify: `[Your Representation]`)
+
+**Orientation Representation:**
+- [X] **Quaternions** (x, y, z, w)
+- [ ] **Euler Angles** (roll, pitch, yaw)
+- [ ] **Axis-Angle** (rotation vector)
+- [ ] **Rotation Matrix** (3x3 matrix)
+- [ ] **Other** (Please specify: `[Your Representation]`)
+
+**Reference Frame:**
+- [ ] **Robot Base Frame**
+- [ ] **Tool/End-Effector Frame**
+- [ ] **World/Global Frame**
+- [X] **Camera Frame**
+- [ ] **Other** (Please specify: `[Your Frame]`)
+
+**Action Dimensions:**
+```
+action: [x_left, y_left, z_left, quat_x_left,... quat_z_left, clutchBtn_left, energyBtn_left, thumbstickBtn_left, pince_left, thumbstick_x_left, thumbstick_y_left, ... (repeated for right)]
+
+- x,y,z: [float[3]] Absolute position of the hand controller (left or right) (metres)
+- quat_(x,y,z): [float[4]] Absolute orientation of the hand controller (left or right) (radians)
+- pince: [float from 0-1] value of the pince on the hand controller. Maps to the instrument jaws.
+- clutchBtn: [bool] Button input to clutch in/out. Engages and disengages a selected arm.
+- energyBtn: [bool] Button to activate electrosurgery.
+- thumbstickBtn: [bool] Button to activate ICG (Indocyanine green) visualisation.
+- thumbstick_x/y: [float[2]] Value of thumbstick - used to control the endoscope.
+```
+<div align="center">
+    <img src="./images/CMR_Versius_System_Dual_Hand_Controller.png"  alt="Versius Hand Controllers." width="600" />
+</div>
+
+*All actions come in `_left` and `_right` variants, which map to the left and right hand controllers.*
+
+### State Space Representation
+
+**State Information Included:**
+- [ ] **Joint Positions** (all articulated joints)
+- [ ] **Joint Velocities**
+- [ ] **End-Effector Pose** (Cartesian position/orientation)
+- [ ] **Force/Torque Readings**
+- [ ] **Gripper State** (position, force, etc.)
+- [X] **Other** (Please specify: `See below`)
+
+**State Dimensions:**
+
+For all state dimensions, the symbol `<S>` (Side) is used to denote a state that has `_left` and `_right` variants.
+```
+observation.state.haptic_<S>_armengageable: [bool[2]] Whether the arm selected by the left or right controller is in an engageable state. An arm may be enagageable without being actively engaged. Arms may be unengageable due to being in the wrong mode (e.g Instrument Change).
+
+observation.state.armlinkedtohaptic_<S>: [int[2]] Index of the arm connected to each hand controller at any one time.
+
+observation.state.hapticengaged_<S>: [bool[2]] Whether the haptic is in an engaged state. An engaged haptic means that its movements are currently driving a robotic arm.
+
+observation.state.arm_<int>_color: [int[5]] Color of each arm connected to the system. The color of each arm is shown on the HUD icons.
+
+observation.state.arm_<int>_instrtype: [int[5]] Instrument type attached to each arm.
+
+observation.state.rotationscaling: [float] Current scaling applied to the relative hand controller orientation. Can change during surgery.
+
+observation.state.translationScaling: [float] Current scaling applied to the relative hand controller position. Can change during surgery.
+
+observation.state.electroSurgerymode_<S>: [int[2]]: Electrosurgery mode (CUT, COAG) selected on each hand controller.
+
+observation.state.haptic_engaged_<S>: [bool[2]] Whether the haptic is engaged or not. When this state is false, the arms are not being controlled by the surgeon and are therefore not performing surgical tasks. The arms may still be moved by bedside assistants to retract the instrument or reposition the arm.
+
+observation.state.icgmode: [int] ICG (Indocyanine Green) mode selected (grayscale, overlay, heatmap).
+
+observation.state.icgenabled: [bool] ICG state, toggled with the right `thumbstickBtn`.
+observation.state.instrtype_<S>: [int[2]] Instrument attached to each hand controller (and in turn attached to an arm as per state.linkedArmIndex)
+```
+
+---
+
+## ⏱️ Data Synchronization Approach
+
+Versius telemetry and video is timestamped using the clock in its network controller. The network controller timestamps at the point of recording telemetry at 50Hz and video at 60Hz.
+Although both telemetry and video are timestamped using the same clock, there is a small variation in timing between the telemetry and video. As such, the system's HUD (Heads-Up Display) was used to synchronise telemetry and video.
+The timestamps of events in the HUD (including but not limited to: changes in arm mode, activation of electro-surgery and changes in electro-surgery mode) are compared against the timestamps of the corresponding events in telemetry. This
+provides us with a reliable measure of lag and the rate of dropped frames. The result of this synchronisation process is used to post-process the telemetry such that it is as close to perfectly in sync as possible.
+
+
+---
+
+## 👥 Attribution & Contact
+This dataset is shared with the aim of advancing research and development that has the potential to benefit patients through improved outcomes and increased access to minimal access surgery. 
+| | |
+| :--- | :--- |
+| **Dataset Lead** | Patrick Thornycroft, Diego Marana, Jonathon Hawkins, Filip Binkiewicz, Joyce Zhang |
+| **Institution** | CMR Surgical Ltd. |
+| **Contact Email** | patrick.thornycroft@cmrsurgical.com, diego.marana@cmrsurgical.com |
+| **Citation (BibTeX)** | N/A |

Surgical/cmr_surgical/hysterectomy/README.md

@@ -0,0 +1,247 @@
+<!--
+Open-H Embodiment Dataset README Template (v1.0)
+Please fill out this template and include it in the ./metadata directory of your LeRobot dataset.
+This file helps others understand the context and details of your contribution.
+-->
+
+# CMR Surgical Versius Surgical Robot Dataset - README
+
+---
+
+## 📋 At a Glance
+
+*This dataset contains a variety of in-vivo surgeries performed with a Versius surgical system. The Versius system by CMR Surgical is a modular soft-tissue surgical robotic system.*
+*Versius can be used with up to 4 robotic arms simultaneously, and supports a variety of surgical instruments.* 
+<div align="center">
+    <img src="./images/CMR_Versius_System.png"  alt="Versius system used in surgery." width="600" />
+</div>
+
+*During surgery, each robotic arm is distinguishable by its LED band color. The arm with a white LED band is a Visualisation Arm and holds the endoscopic camera.*
+
+<div align="center">
+    <img src="./images/CMR_Versius_System_Console_Arms.png"  alt="Versius system used in surgery." width="600" />
+</div>
+
+*Versius is controlled from the Surgeon Console, where two 7 degrees of freedom hand controllers capture the surgeon's hand movements to enable precise control of the instruments attached to the robot arm.*
+
+<div align="center">
+<img src="./images/CMR_Surgeon_Control.png" alt="Versius console controlled by a surgeon." width="600"/>
+</div>
+
+
+
+*Each hand controller has an electro-surgery activation button (1), a clutch button (2), a thumbstick (3) and a trigger (also known as pince) (4).*
+
+
+<div align="center">
+    <img src="./images/CMR_Hand_Controller_labeled.png"  alt="Versius Hand Controller labeled." width="600" />
+</div>
+
+
+---
+
+## 📖 Dataset Overview
+
+This dataset contains video and telemetry for 399 surgeries, split into 99 Cholecystectomies, 100 Hysterectomies, 100 Hernia repairs, and 100 Prostatectomies. The video is taken from an endoscopic camera attached to one of the Versius arms. The video is only recorded when inside the patient to ensure it is anonymous.
+
+
+| | |
+| :--- | :--- |
+| **Total Trajectories** | `N/A` |
+| **Total Hours** | `485 hours` |
+| **Data Type** | `Clinical` |
+| **License** | CC BY 4.0 |
+| **Version** | `[1.0]` |
+
+---
+
+## 🎯 Tasks & Domain
+
+### Domain
+
+- [X] **Surgical Robotics**
+- [ ] **Ultrasound Robotics**
+- [ ] **Other Healthcare Robotics** (Please specify: `[Your Domain]`)
+
+### Demonstrated Skills
+
+The videos in this dataset demonstrate succesful completions of cholecystectomy, prostatectomy, hysterectomy, and inguinal hernia repair surgeries.
+
+---
+
+## 🔬 Data Collection Details
+
+### Collection Method
+
+- [X] **Human Teleoperation**
+- [ ] **Programmatic/State-Machine**
+- [ ] **AI Policy / Autonomous**
+- [ ] **Other** (Please specify: `[Your Method]`)
+
+
+### Operator Details
+
+| | Description |
+| :--- | :--- |
+| **Operator Count** | `N/A` |
+| **Operator Skill Level** | `[X] Expert Surgeon` <br> `[ ] Intermediate (e.g., Trained Researcher)` <br> `[ ] Novice (e.g., ML Researcher with minimal experience)` <br> `[ ] N/A` |
+| **Collection Period** | From `[2025-01-01]` to `[2025-11-05]` |
+
+### Recovery Demonstrations
+
+This dataset is provided unlabeled, but will include instances of failure and recovery when completing a surgical task.
+
+- [X] **Yes**
+- [ ] **No**
+
+---
+
+## 💡 Diversity Dimensions
+
+- [ ] **Camera Position / Angle**
+- [ ] **Lighting Conditions**
+- [X] **Target Object** (Different procedure types)
+- [X] **Spatial Layout** (e.g., placing the target suture needle in various locations)
+- [ ] **Robot Embodiment** (if multiple robots were used)
+- [X] **Task Execution** (e.g., different techniques for the same task)
+- [X] **Background / Scene**
+- [ ] **Other** (Please specify: `[Your Dimension]`)
+
+The target objects and tasks were varied by selecting surgeries of different procedure types. Camera angles and background naturally change depending on the patient and procedure. The background will mostly be patient anatomy, except when the 
+endoscope is close to its trocar - in these situations the view will be occluded by the trocar.
+
+The spatial layout is mostly constant, with at least one surgical instrument to either side of the camera. Most often two instruments will be in the field of view of the camera.
+
+
+---
+
+## 🛠️ Equipment & Setup
+
+### Robotic Platform(s)
+
+
+- **Robot:** `Versius Surgical System`
+
+Depending on the surgery, anywhere between 1-4 robotic arms are used simultaneously.
+
+
+### Sensors & Cameras
+
+| Type | Model/Details |
+| :--- | :--- |
+| **Primary Camera** | `Endoscopic Camera, 1920x1080 @ 60fps` |
+
+---
+
+## 🎯 Action & State Space Representation
+
+The actions in the dataset map directly to all the inputs available to a surgeon on a Versius system.
+The actions map to the buttons in two hand controllers (left and right), and include:
+1. Orientation quaternion (a,b,c,w)
+2. Position (x,y,z)
+2. Pince
+3. Clutch Button
+4. Energy Button
+5. Thumbstick Button
+6. Joystick X and Joystick Y
+
+The movement of the hand controllers maps to the camera frame.
+
+### Action Space Representation
+
+**Primary Action Representation:**
+- [X] **Absolute Cartesian** (position/orientation relative to robot base)
+- [ ] **Relative Cartesian** (delta position/orientation from current pose)
+- [ ] **Joint Space** (direct joint angle commands)
+- [ ] **Other** (Please specify: `[Your Representation]`)
+
+**Orientation Representation:**
+- [X] **Quaternions** (x, y, z, w)
+- [ ] **Euler Angles** (roll, pitch, yaw)
+- [ ] **Axis-Angle** (rotation vector)
+- [ ] **Rotation Matrix** (3x3 matrix)
+- [ ] **Other** (Please specify: `[Your Representation]`)
+
+**Reference Frame:**
+- [ ] **Robot Base Frame**
+- [ ] **Tool/End-Effector Frame**
+- [ ] **World/Global Frame**
+- [X] **Camera Frame**
+- [ ] **Other** (Please specify: `[Your Frame]`)
+
+**Action Dimensions:**
+```
+action: [x_left, y_left, z_left, quat_x_left,... quat_z_left, clutchBtn_left, energyBtn_left, thumbstickBtn_left, pince_left, thumbstick_x_left, thumbstick_y_left, ... (repeated for right)]
+
+- x,y,z: [float[3]] Absolute position of the hand controller (left or right) (metres)
+- quat_(x,y,z): [float[4]] Absolute orientation of the hand controller (left or right) (radians)
+- pince: [float from 0-1] value of the pince on the hand controller. Maps to the instrument jaws.
+- clutchBtn: [bool] Button input to clutch in/out. Engages and disengages a selected arm.
+- energyBtn: [bool] Button to activate electrosurgery.
+- thumbstickBtn: [bool] Button to activate ICG (Indocyanine green) visualisation.
+- thumbstick_x/y: [float[2]] Value of thumbstick - used to control the endoscope.
+```
+<div align="center">
+    <img src="./images/CMR_Versius_System_Dual_Hand_Controller.png"  alt="Versius Hand Controllers." width="600" />
+</div>
+
+*All actions come in `_left` and `_right` variants, which map to the left and right hand controllers.*
+
+### State Space Representation
+
+**State Information Included:**
+- [ ] **Joint Positions** (all articulated joints)
+- [ ] **Joint Velocities**
+- [ ] **End-Effector Pose** (Cartesian position/orientation)
+- [ ] **Force/Torque Readings**
+- [ ] **Gripper State** (position, force, etc.)
+- [X] **Other** (Please specify: `See below`)
+
+**State Dimensions:**
+
+For all state dimensions, the symbol `<S>` (Side) is used to denote a state that has `_left` and `_right` variants.
+```
+observation.state.haptic_<S>_armengageable: [bool[2]] Whether the arm selected by the left or right controller is in an engageable state. An arm may be enagageable without being actively engaged. Arms may be unengageable due to being in the wrong mode (e.g Instrument Change).
+
+observation.state.armlinkedtohaptic_<S>: [int[2]] Index of the arm connected to each hand controller at any one time.
+
+observation.state.hapticengaged_<S>: [bool[2]] Whether the haptic is in an engaged state. An engaged haptic means that its movements are currently driving a robotic arm.
+
+observation.state.arm_<int>_color: [int[5]] Color of each arm connected to the system. The color of each arm is shown on the HUD icons.
+
+observation.state.arm_<int>_instrtype: [int[5]] Instrument type attached to each arm.
+
+observation.state.rotationscaling: [float] Current scaling applied to the relative hand controller orientation. Can change during surgery.
+
+observation.state.translationScaling: [float] Current scaling applied to the relative hand controller position. Can change during surgery.
+
+observation.state.electroSurgerymode_<S>: [int[2]]: Electrosurgery mode (CUT, COAG) selected on each hand controller.
+
+observation.state.haptic_engaged_<S>: [bool[2]] Whether the haptic is engaged or not. When this state is false, the arms are not being controlled by the surgeon and are therefore not performing surgical tasks. The arms may still be moved by bedside assistants to retract the instrument or reposition the arm.
+
+observation.state.icgmode: [int] ICG (Indocyanine Green) mode selected (grayscale, overlay, heatmap).
+
+observation.state.icgenabled: [bool] ICG state, toggled with the right `thumbstickBtn`.
+observation.state.instrtype_<S>: [int[2]] Instrument attached to each hand controller (and in turn attached to an arm as per state.linkedArmIndex)
+```
+
+---
+
+## ⏱️ Data Synchronization Approach
+
+Versius telemetry and video is timestamped using the clock in its network controller. The network controller timestamps at the point of recording telemetry at 50Hz and video at 60Hz.
+Although both telemetry and video are timestamped using the same clock, there is a small variation in timing between the telemetry and video. As such, the system's HUD (Heads-Up Display) was used to synchronise telemetry and video.
+The timestamps of events in the HUD (including but not limited to: changes in arm mode, activation of electro-surgery and changes in electro-surgery mode) are compared against the timestamps of the corresponding events in telemetry. This
+provides us with a reliable measure of lag and the rate of dropped frames. The result of this synchronisation process is used to post-process the telemetry such that it is as close to perfectly in sync as possible.
+
+
+---
+
+## 👥 Attribution & Contact
+This dataset is shared with the aim of advancing research and development that has the potential to benefit patients through improved outcomes and increased access to minimal access surgery. 
+| | |
+| :--- | :--- |
+| **Dataset Lead** | Patrick Thornycroft, Diego Marana, Jonathon Hawkins, Filip Binkiewicz, Joyce Zhang |
+| **Institution** | CMR Surgical Ltd. |
+| **Contact Email** | patrick.thornycroft@cmrsurgical.com, diego.marana@cmrsurgical.com |
+| **Citation (BibTeX)** | N/A |

Surgical/cmr_surgical/inguinal_hernia/README.md

@@ -0,0 +1,247 @@
+<!--
+Open-H Embodiment Dataset README Template (v1.0)
+Please fill out this template and include it in the ./metadata directory of your LeRobot dataset.
+This file helps others understand the context and details of your contribution.
+-->
+
+# CMR Surgical Versius Surgical Robot Dataset - README
+
+---
+
+## 📋 At a Glance
+
+*This dataset contains a variety of in-vivo surgeries performed with a Versius surgical system. The Versius system by CMR Surgical is a modular soft-tissue surgical robotic system.*
+*Versius can be used with up to 4 robotic arms simultaneously, and supports a variety of surgical instruments.* 
+<div align="center">
+    <img src="./images/CMR_Versius_System.png"  alt="Versius system used in surgery." width="600" />
+</div>
+
+*During surgery, each robotic arm is distinguishable by its LED band color. The arm with a white LED band is a Visualisation Arm and holds the endoscopic camera.*
+
+<div align="center">
+    <img src="./images/CMR_Versius_System_Console_Arms.png"  alt="Versius system used in surgery." width="600" />
+</div>
+
+*Versius is controlled from the Surgeon Console, where two 7 degrees of freedom hand controllers capture the surgeon's hand movements to enable precise control of the instruments attached to the robot arm.*
+
+<div align="center">
+<img src="./images/CMR_Surgeon_Control.png" alt="Versius console controlled by a surgeon." width="600"/>
+</div>
+
+
+
+*Each hand controller has an electro-surgery activation button (1), a clutch button (2), a thumbstick (3) and a trigger (also known as pince) (4).*
+
+
+<div align="center">
+    <img src="./images/CMR_Hand_Controller_labeled.png"  alt="Versius Hand Controller labeled." width="600" />
+</div>
+
+
+---
+
+## 📖 Dataset Overview
+
+This dataset contains video and telemetry for 399 surgeries, split into 99 Cholecystectomies, 100 Hysterectomies, 100 Hernia repairs, and 100 Prostatectomies. The video is taken from an endoscopic camera attached to one of the Versius arms. The video is only recorded when inside the patient to ensure it is anonymous.
+
+
+| | |
+| :--- | :--- |
+| **Total Trajectories** | `N/A` |
+| **Total Hours** | `485 hours` |
+| **Data Type** | `Clinical` |
+| **License** | CC BY 4.0 |
+| **Version** | `[1.0]` |
+
+---
+
+## 🎯 Tasks & Domain
+
+### Domain
+
+- [X] **Surgical Robotics**
+- [ ] **Ultrasound Robotics**
+- [ ] **Other Healthcare Robotics** (Please specify: `[Your Domain]`)
+
+### Demonstrated Skills
+
+The videos in this dataset demonstrate succesful completions of cholecystectomy, prostatectomy, hysterectomy, and inguinal hernia repair surgeries.
+
+---
+
+## 🔬 Data Collection Details
+
+### Collection Method
+
+- [X] **Human Teleoperation**
+- [ ] **Programmatic/State-Machine**
+- [ ] **AI Policy / Autonomous**
+- [ ] **Other** (Please specify: `[Your Method]`)
+
+
+### Operator Details
+
+| | Description |
+| :--- | :--- |
+| **Operator Count** | `N/A` |
+| **Operator Skill Level** | `[X] Expert Surgeon` <br> `[ ] Intermediate (e.g., Trained Researcher)` <br> `[ ] Novice (e.g., ML Researcher with minimal experience)` <br> `[ ] N/A` |
+| **Collection Period** | From `[2025-01-01]` to `[2025-11-05]` |
+
+### Recovery Demonstrations
+
+This dataset is provided unlabeled, but will include instances of failure and recovery when completing a surgical task.
+
+- [X] **Yes**
+- [ ] **No**
+
+---
+
+## 💡 Diversity Dimensions
+
+- [ ] **Camera Position / Angle**
+- [ ] **Lighting Conditions**
+- [X] **Target Object** (Different procedure types)
+- [X] **Spatial Layout** (e.g., placing the target suture needle in various locations)
+- [ ] **Robot Embodiment** (if multiple robots were used)
+- [X] **Task Execution** (e.g., different techniques for the same task)
+- [X] **Background / Scene**
+- [ ] **Other** (Please specify: `[Your Dimension]`)
+
+The target objects and tasks were varied by selecting surgeries of different procedure types. Camera angles and background naturally change depending on the patient and procedure. The background will mostly be patient anatomy, except when the 
+endoscope is close to its trocar - in these situations the view will be occluded by the trocar.
+
+The spatial layout is mostly constant, with at least one surgical instrument to either side of the camera. Most often two instruments will be in the field of view of the camera.
+
+
+---
+
+## 🛠️ Equipment & Setup
+
+### Robotic Platform(s)
+
+
+- **Robot:** `Versius Surgical System`
+
+Depending on the surgery, anywhere between 1-4 robotic arms are used simultaneously.
+
+
+### Sensors & Cameras
+
+| Type | Model/Details |
+| :--- | :--- |
+| **Primary Camera** | `Endoscopic Camera, 1920x1080 @ 60fps` |
+
+---
+
+## 🎯 Action & State Space Representation
+
+The actions in the dataset map directly to all the inputs available to a surgeon on a Versius system.
+The actions map to the buttons in two hand controllers (left and right), and include:
+1. Orientation quaternion (a,b,c,w)
+2. Position (x,y,z)
+2. Pince
+3. Clutch Button
+4. Energy Button
+5. Thumbstick Button
+6. Joystick X and Joystick Y
+
+The movement of the hand controllers maps to the camera frame.
+
+### Action Space Representation
+
+**Primary Action Representation:**
+- [X] **Absolute Cartesian** (position/orientation relative to robot base)
+- [ ] **Relative Cartesian** (delta position/orientation from current pose)
+- [ ] **Joint Space** (direct joint angle commands)
+- [ ] **Other** (Please specify: `[Your Representation]`)
+
+**Orientation Representation:**
+- [X] **Quaternions** (x, y, z, w)
+- [ ] **Euler Angles** (roll, pitch, yaw)
+- [ ] **Axis-Angle** (rotation vector)
+- [ ] **Rotation Matrix** (3x3 matrix)
+- [ ] **Other** (Please specify: `[Your Representation]`)
+
+**Reference Frame:**
+- [ ] **Robot Base Frame**
+- [ ] **Tool/End-Effector Frame**
+- [ ] **World/Global Frame**
+- [X] **Camera Frame**
+- [ ] **Other** (Please specify: `[Your Frame]`)
+
+**Action Dimensions:**
+```
+action: [x_left, y_left, z_left, quat_x_left,... quat_z_left, clutchBtn_left, energyBtn_left, thumbstickBtn_left, pince_left, thumbstick_x_left, thumbstick_y_left, ... (repeated for right)]
+
+- x,y,z: [float[3]] Absolute position of the hand controller (left or right) (metres)
+- quat_(x,y,z): [float[4]] Absolute orientation of the hand controller (left or right) (radians)
+- pince: [float from 0-1] value of the pince on the hand controller. Maps to the instrument jaws.
+- clutchBtn: [bool] Button input to clutch in/out. Engages and disengages a selected arm.
+- energyBtn: [bool] Button to activate electrosurgery.
+- thumbstickBtn: [bool] Button to activate ICG (Indocyanine green) visualisation.
+- thumbstick_x/y: [float[2]] Value of thumbstick - used to control the endoscope.
+```
+<div align="center">
+    <img src="./images/CMR_Versius_System_Dual_Hand_Controller.png"  alt="Versius Hand Controllers." width="600" />
+</div>
+
+*All actions come in `_left` and `_right` variants, which map to the left and right hand controllers.*
+
+### State Space Representation
+
+**State Information Included:**
+- [ ] **Joint Positions** (all articulated joints)
+- [ ] **Joint Velocities**
+- [ ] **End-Effector Pose** (Cartesian position/orientation)
+- [ ] **Force/Torque Readings**
+- [ ] **Gripper State** (position, force, etc.)
+- [X] **Other** (Please specify: `See below`)
+
+**State Dimensions:**
+
+For all state dimensions, the symbol `<S>` (Side) is used to denote a state that has `_left` and `_right` variants.
+```
+observation.state.haptic_<S>_armengageable: [bool[2]] Whether the arm selected by the left or right controller is in an engageable state. An arm may be enagageable without being actively engaged. Arms may be unengageable due to being in the wrong mode (e.g Instrument Change).
+
+observation.state.armlinkedtohaptic_<S>: [int[2]] Index of the arm connected to each hand controller at any one time.
+
+observation.state.hapticengaged_<S>: [bool[2]] Whether the haptic is in an engaged state. An engaged haptic means that its movements are currently driving a robotic arm.
+
+observation.state.arm_<int>_color: [int[5]] Color of each arm connected to the system. The color of each arm is shown on the HUD icons.
+
+observation.state.arm_<int>_instrtype: [int[5]] Instrument type attached to each arm.
+
+observation.state.rotationscaling: [float] Current scaling applied to the relative hand controller orientation. Can change during surgery.
+
+observation.state.translationScaling: [float] Current scaling applied to the relative hand controller position. Can change during surgery.
+
+observation.state.electroSurgerymode_<S>: [int[2]]: Electrosurgery mode (CUT, COAG) selected on each hand controller.
+
+observation.state.haptic_engaged_<S>: [bool[2]] Whether the haptic is engaged or not. When this state is false, the arms are not being controlled by the surgeon and are therefore not performing surgical tasks. The arms may still be moved by bedside assistants to retract the instrument or reposition the arm.
+
+observation.state.icgmode: [int] ICG (Indocyanine Green) mode selected (grayscale, overlay, heatmap).
+
+observation.state.icgenabled: [bool] ICG state, toggled with the right `thumbstickBtn`.
+observation.state.instrtype_<S>: [int[2]] Instrument attached to each hand controller (and in turn attached to an arm as per state.linkedArmIndex)
+```
+
+---
+
+## ⏱️ Data Synchronization Approach
+
+Versius telemetry and video is timestamped using the clock in its network controller. The network controller timestamps at the point of recording telemetry at 50Hz and video at 60Hz.
+Although both telemetry and video are timestamped using the same clock, there is a small variation in timing between the telemetry and video. As such, the system's HUD (Heads-Up Display) was used to synchronise telemetry and video.
+The timestamps of events in the HUD (including but not limited to: changes in arm mode, activation of electro-surgery and changes in electro-surgery mode) are compared against the timestamps of the corresponding events in telemetry. This
+provides us with a reliable measure of lag and the rate of dropped frames. The result of this synchronisation process is used to post-process the telemetry such that it is as close to perfectly in sync as possible.
+
+
+---
+
+## 👥 Attribution & Contact
+This dataset is shared with the aim of advancing research and development that has the potential to benefit patients through improved outcomes and increased access to minimal access surgery. 
+| | |
+| :--- | :--- |
+| **Dataset Lead** | Patrick Thornycroft, Diego Marana, Jonathon Hawkins, Filip Binkiewicz, Joyce Zhang |
+| **Institution** | CMR Surgical Ltd. |
+| **Contact Email** | patrick.thornycroft@cmrsurgical.com, diego.marana@cmrsurgical.com |
+| **Citation (BibTeX)** | N/A |

Surgical/cmr_surgical/peg_transfer/README.md

@@ -0,0 +1,247 @@
+<!--
+Open-H Embodiment Dataset README Template (v1.0)
+Please fill out this template and include it in the ./metadata directory of your LeRobot dataset.
+This file helps others understand the context and details of your contribution.
+-->
+
+# CMR Surgical Versius Surgical Robot Dataset - README
+
+---
+
+## 📋 At a Glance
+
+*This dataset contains a variety of in-vivo surgeries performed with a Versius surgical system. The Versius system by CMR Surgical is a modular soft-tissue surgical robotic system.*
+*Versius can be used with up to 4 robotic arms simultaneously, and supports a variety of surgical instruments.* 
+<div align="center">
+    <img src="./images/CMR_Versius_System.png"  alt="Versius system used in surgery." width="600" />
+</div>
+
+*During surgery, each robotic arm is distinguishable by its LED band color. The arm with a white LED band is a Visualisation Arm and holds the endoscopic camera.*
+
+<div align="center">
+    <img src="./images/CMR_Versius_System_Console_Arms.png"  alt="Versius system used in surgery." width="600" />
+</div>
+
+*Versius is controlled from the Surgeon Console, where two 7 degrees of freedom hand controllers capture the surgeon's hand movements to enable precise control of the instruments attached to the robot arm.*
+
+<div align="center">
+<img src="./images/CMR_Surgeon_Control.png" alt="Versius console controlled by a surgeon." width="600"/>
+</div>
+
+
+
+*Each hand controller has an electro-surgery activation button (1), a clutch button (2), a thumbstick (3) and a trigger (also known as pince) (4).*
+
+
+<div align="center">
+    <img src="./images/CMR_Hand_Controller_labeled.png"  alt="Versius Hand Controller labeled." width="600" />
+</div>
+
+
+---
+
+## 📖 Dataset Overview
+
+This dataset contains video and telemetry for 399 surgeries, split into 99 Cholecystectomies, 100 Hysterectomies, 100 Hernia repairs, and 100 Prostatectomies. The video is taken from an endoscopic camera attached to one of the Versius arms. The video is only recorded when inside the patient to ensure it is anonymous.
+
+
+| | |
+| :--- | :--- |
+| **Total Trajectories** | `N/A` |
+| **Total Hours** | `485 hours` |
+| **Data Type** | `Clinical` |
+| **License** | CC BY 4.0 |
+| **Version** | `[1.0]` |
+
+---
+
+## 🎯 Tasks & Domain
+
+### Domain
+
+- [X] **Surgical Robotics**
+- [ ] **Ultrasound Robotics**
+- [ ] **Other Healthcare Robotics** (Please specify: `[Your Domain]`)
+
+### Demonstrated Skills
+
+The videos in this dataset demonstrate succesful completions of cholecystectomy, prostatectomy, hysterectomy, and inguinal hernia repair surgeries.
+
+---
+
+## 🔬 Data Collection Details
+
+### Collection Method
+
+- [X] **Human Teleoperation**
+- [ ] **Programmatic/State-Machine**
+- [ ] **AI Policy / Autonomous**
+- [ ] **Other** (Please specify: `[Your Method]`)
+
+
+### Operator Details
+
+| | Description |
+| :--- | :--- |
+| **Operator Count** | `N/A` |
+| **Operator Skill Level** | `[X] Expert Surgeon` <br> `[ ] Intermediate (e.g., Trained Researcher)` <br> `[ ] Novice (e.g., ML Researcher with minimal experience)` <br> `[ ] N/A` |
+| **Collection Period** | From `[2025-01-01]` to `[2025-11-05]` |
+
+### Recovery Demonstrations
+
+This dataset is provided unlabeled, but will include instances of failure and recovery when completing a surgical task.
+
+- [X] **Yes**
+- [ ] **No**
+
+---
+
+## 💡 Diversity Dimensions
+
+- [ ] **Camera Position / Angle**
+- [ ] **Lighting Conditions**
+- [X] **Target Object** (Different procedure types)
+- [X] **Spatial Layout** (e.g., placing the target suture needle in various locations)
+- [ ] **Robot Embodiment** (if multiple robots were used)
+- [X] **Task Execution** (e.g., different techniques for the same task)
+- [X] **Background / Scene**
+- [ ] **Other** (Please specify: `[Your Dimension]`)
+
+The target objects and tasks were varied by selecting surgeries of different procedure types. Camera angles and background naturally change depending on the patient and procedure. The background will mostly be patient anatomy, except when the 
+endoscope is close to its trocar - in these situations the view will be occluded by the trocar.
+
+The spatial layout is mostly constant, with at least one surgical instrument to either side of the camera. Most often two instruments will be in the field of view of the camera.
+
+
+---
+
+## 🛠️ Equipment & Setup
+
+### Robotic Platform(s)
+
+
+- **Robot:** `Versius Surgical System`
+
+Depending on the surgery, anywhere between 1-4 robotic arms are used simultaneously.
+
+
+### Sensors & Cameras
+
+| Type | Model/Details |
+| :--- | :--- |
+| **Primary Camera** | `Endoscopic Camera, 1920x1080 @ 60fps` |
+
+---
+
+## 🎯 Action & State Space Representation
+
+The actions in the dataset map directly to all the inputs available to a surgeon on a Versius system.
+The actions map to the buttons in two hand controllers (left and right), and include:
+1. Orientation quaternion (a,b,c,w)
+2. Position (x,y,z)
+2. Pince
+3. Clutch Button
+4. Energy Button
+5. Thumbstick Button
+6. Joystick X and Joystick Y
+
+The movement of the hand controllers maps to the camera frame.
+
+### Action Space Representation
+
+**Primary Action Representation:**
+- [X] **Absolute Cartesian** (position/orientation relative to robot base)
+- [ ] **Relative Cartesian** (delta position/orientation from current pose)
+- [ ] **Joint Space** (direct joint angle commands)
+- [ ] **Other** (Please specify: `[Your Representation]`)
+
+**Orientation Representation:**
+- [X] **Quaternions** (x, y, z, w)
+- [ ] **Euler Angles** (roll, pitch, yaw)
+- [ ] **Axis-Angle** (rotation vector)
+- [ ] **Rotation Matrix** (3x3 matrix)
+- [ ] **Other** (Please specify: `[Your Representation]`)
+
+**Reference Frame:**
+- [ ] **Robot Base Frame**
+- [ ] **Tool/End-Effector Frame**
+- [ ] **World/Global Frame**
+- [X] **Camera Frame**
+- [ ] **Other** (Please specify: `[Your Frame]`)
+
+**Action Dimensions:**
+```
+action: [x_left, y_left, z_left, quat_x_left,... quat_z_left, clutchBtn_left, energyBtn_left, thumbstickBtn_left, pince_left, thumbstick_x_left, thumbstick_y_left, ... (repeated for right)]
+
+- x,y,z: [float[3]] Absolute position of the hand controller (left or right) (metres)
+- quat_(x,y,z): [float[4]] Absolute orientation of the hand controller (left or right) (radians)
+- pince: [float from 0-1] value of the pince on the hand controller. Maps to the instrument jaws.
+- clutchBtn: [bool] Button input to clutch in/out. Engages and disengages a selected arm.
+- energyBtn: [bool] Button to activate electrosurgery.
+- thumbstickBtn: [bool] Button to activate ICG (Indocyanine green) visualisation.
+- thumbstick_x/y: [float[2]] Value of thumbstick - used to control the endoscope.
+```
+<div align="center">
+    <img src="./images/CMR_Versius_System_Dual_Hand_Controller.png"  alt="Versius Hand Controllers." width="600" />
+</div>
+
+*All actions come in `_left` and `_right` variants, which map to the left and right hand controllers.*
+
+### State Space Representation
+
+**State Information Included:**
+- [ ] **Joint Positions** (all articulated joints)
+- [ ] **Joint Velocities**
+- [ ] **End-Effector Pose** (Cartesian position/orientation)
+- [ ] **Force/Torque Readings**
+- [ ] **Gripper State** (position, force, etc.)
+- [X] **Other** (Please specify: `See below`)
+
+**State Dimensions:**
+
+For all state dimensions, the symbol `<S>` (Side) is used to denote a state that has `_left` and `_right` variants.
+```
+observation.state.haptic_<S>_armengageable: [bool[2]] Whether the arm selected by the left or right controller is in an engageable state. An arm may be enagageable without being actively engaged. Arms may be unengageable due to being in the wrong mode (e.g Instrument Change).
+
+observation.state.armlinkedtohaptic_<S>: [int[2]] Index of the arm connected to each hand controller at any one time.
+
+observation.state.hapticengaged_<S>: [bool[2]] Whether the haptic is in an engaged state. An engaged haptic means that its movements are currently driving a robotic arm.
+
+observation.state.arm_<int>_color: [int[5]] Color of each arm connected to the system. The color of each arm is shown on the HUD icons.
+
+observation.state.arm_<int>_instrtype: [int[5]] Instrument type attached to each arm.
+
+observation.state.rotationscaling: [float] Current scaling applied to the relative hand controller orientation. Can change during surgery.
+
+observation.state.translationScaling: [float] Current scaling applied to the relative hand controller position. Can change during surgery.
+
+observation.state.electroSurgerymode_<S>: [int[2]]: Electrosurgery mode (CUT, COAG) selected on each hand controller.
+
+observation.state.haptic_engaged_<S>: [bool[2]] Whether the haptic is engaged or not. When this state is false, the arms are not being controlled by the surgeon and are therefore not performing surgical tasks. The arms may still be moved by bedside assistants to retract the instrument or reposition the arm.
+
+observation.state.icgmode: [int] ICG (Indocyanine Green) mode selected (grayscale, overlay, heatmap).
+
+observation.state.icgenabled: [bool] ICG state, toggled with the right `thumbstickBtn`.
+observation.state.instrtype_<S>: [int[2]] Instrument attached to each hand controller (and in turn attached to an arm as per state.linkedArmIndex)
+```
+
+---
+
+## ⏱️ Data Synchronization Approach
+
+Versius telemetry and video is timestamped using the clock in its network controller. The network controller timestamps at the point of recording telemetry at 50Hz and video at 60Hz.
+Although both telemetry and video are timestamped using the same clock, there is a small variation in timing between the telemetry and video. As such, the system's HUD (Heads-Up Display) was used to synchronise telemetry and video.
+The timestamps of events in the HUD (including but not limited to: changes in arm mode, activation of electro-surgery and changes in electro-surgery mode) are compared against the timestamps of the corresponding events in telemetry. This
+provides us with a reliable measure of lag and the rate of dropped frames. The result of this synchronisation process is used to post-process the telemetry such that it is as close to perfectly in sync as possible.
+
+
+---
+
+## 👥 Attribution & Contact
+This dataset is shared with the aim of advancing research and development that has the potential to benefit patients through improved outcomes and increased access to minimal access surgery. 
+| | |
+| :--- | :--- |
+| **Dataset Lead** | Patrick Thornycroft, Diego Marana, Jonathon Hawkins, Filip Binkiewicz, Joyce Zhang |
+| **Institution** | CMR Surgical Ltd. |
+| **Contact Email** | patrick.thornycroft@cmrsurgical.com, diego.marana@cmrsurgical.com |
+| **Citation (BibTeX)** | N/A |

Surgical/cmr_surgical/prostatectomy/README.md

@@ -0,0 +1,247 @@
+<!--
+Open-H Embodiment Dataset README Template (v1.0)
+Please fill out this template and include it in the ./metadata directory of your LeRobot dataset.
+This file helps others understand the context and details of your contribution.
+-->
+
+# CMR Surgical Versius Surgical Robot Dataset - README
+
+---
+
+## 📋 At a Glance
+
+*This dataset contains a variety of in-vivo surgeries performed with a Versius surgical system. The Versius system by CMR Surgical is a modular soft-tissue surgical robotic system.*
+*Versius can be used with up to 4 robotic arms simultaneously, and supports a variety of surgical instruments.* 
+<div align="center">
+    <img src="./images/CMR_Versius_System.png"  alt="Versius system used in surgery." width="600" />
+</div>
+
+*During surgery, each robotic arm is distinguishable by its LED band color. The arm with a white LED band is a Visualisation Arm and holds the endoscopic camera.*
+
+<div align="center">
+    <img src="./images/CMR_Versius_System_Console_Arms.png"  alt="Versius system used in surgery." width="600" />
+</div>
+
+*Versius is controlled from the Surgeon Console, where two 7 degrees of freedom hand controllers capture the surgeon's hand movements to enable precise control of the instruments attached to the robot arm.*
+
+<div align="center">
+<img src="./images/CMR_Surgeon_Control.png" alt="Versius console controlled by a surgeon." width="600"/>
+</div>
+
+
+
+*Each hand controller has an electro-surgery activation button (1), a clutch button (2), a thumbstick (3) and a trigger (also known as pince) (4).*
+
+
+<div align="center">
+    <img src="./images/CMR_Hand_Controller_labeled.png"  alt="Versius Hand Controller labeled." width="600" />
+</div>
+
+
+---
+
+## 📖 Dataset Overview
+
+This dataset contains video and telemetry for 399 surgeries, split into 99 Cholecystectomies, 100 Hysterectomies, 100 Hernia repairs, and 100 Prostatectomies. The video is taken from an endoscopic camera attached to one of the Versius arms. The video is only recorded when inside the patient to ensure it is anonymous.
+
+
+| | |
+| :--- | :--- |
+| **Total Trajectories** | `N/A` |
+| **Total Hours** | `485 hours` |
+| **Data Type** | `Clinical` |
+| **License** | CC BY 4.0 |
+| **Version** | `[1.0]` |
+
+---
+
+## 🎯 Tasks & Domain
+
+### Domain
+
+- [X] **Surgical Robotics**
+- [ ] **Ultrasound Robotics**
+- [ ] **Other Healthcare Robotics** (Please specify: `[Your Domain]`)
+
+### Demonstrated Skills
+
+The videos in this dataset demonstrate succesful completions of cholecystectomy, prostatectomy, hysterectomy, and inguinal hernia repair surgeries.
+
+---
+
+## 🔬 Data Collection Details
+
+### Collection Method
+
+- [X] **Human Teleoperation**
+- [ ] **Programmatic/State-Machine**
+- [ ] **AI Policy / Autonomous**
+- [ ] **Other** (Please specify: `[Your Method]`)
+
+
+### Operator Details
+
+| | Description |
+| :--- | :--- |
+| **Operator Count** | `N/A` |
+| **Operator Skill Level** | `[X] Expert Surgeon` <br> `[ ] Intermediate (e.g., Trained Researcher)` <br> `[ ] Novice (e.g., ML Researcher with minimal experience)` <br> `[ ] N/A` |
+| **Collection Period** | From `[2025-01-01]` to `[2025-11-05]` |
+
+### Recovery Demonstrations
+
+This dataset is provided unlabeled, but will include instances of failure and recovery when completing a surgical task.
+
+- [X] **Yes**
+- [ ] **No**
+
+---
+
+## 💡 Diversity Dimensions
+
+- [ ] **Camera Position / Angle**
+- [ ] **Lighting Conditions**
+- [X] **Target Object** (Different procedure types)
+- [X] **Spatial Layout** (e.g., placing the target suture needle in various locations)
+- [ ] **Robot Embodiment** (if multiple robots were used)
+- [X] **Task Execution** (e.g., different techniques for the same task)
+- [X] **Background / Scene**
+- [ ] **Other** (Please specify: `[Your Dimension]`)
+
+The target objects and tasks were varied by selecting surgeries of different procedure types. Camera angles and background naturally change depending on the patient and procedure. The background will mostly be patient anatomy, except when the 
+endoscope is close to its trocar - in these situations the view will be occluded by the trocar.
+
+The spatial layout is mostly constant, with at least one surgical instrument to either side of the camera. Most often two instruments will be in the field of view of the camera.
+
+
+---
+
+## 🛠️ Equipment & Setup
+
+### Robotic Platform(s)
+
+
+- **Robot:** `Versius Surgical System`
+
+Depending on the surgery, anywhere between 1-4 robotic arms are used simultaneously.
+
+
+### Sensors & Cameras
+
+| Type | Model/Details |
+| :--- | :--- |
+| **Primary Camera** | `Endoscopic Camera, 1920x1080 @ 60fps` |
+
+---
+
+## 🎯 Action & State Space Representation
+
+The actions in the dataset map directly to all the inputs available to a surgeon on a Versius system.
+The actions map to the buttons in two hand controllers (left and right), and include:
+1. Orientation quaternion (a,b,c,w)
+2. Position (x,y,z)
+2. Pince
+3. Clutch Button
+4. Energy Button
+5. Thumbstick Button
+6. Joystick X and Joystick Y
+
+The movement of the hand controllers maps to the camera frame.
+
+### Action Space Representation
+
+**Primary Action Representation:**
+- [X] **Absolute Cartesian** (position/orientation relative to robot base)
+- [ ] **Relative Cartesian** (delta position/orientation from current pose)
+- [ ] **Joint Space** (direct joint angle commands)
+- [ ] **Other** (Please specify: `[Your Representation]`)
+
+**Orientation Representation:**
+- [X] **Quaternions** (x, y, z, w)
+- [ ] **Euler Angles** (roll, pitch, yaw)
+- [ ] **Axis-Angle** (rotation vector)
+- [ ] **Rotation Matrix** (3x3 matrix)
+- [ ] **Other** (Please specify: `[Your Representation]`)
+
+**Reference Frame:**
+- [ ] **Robot Base Frame**
+- [ ] **Tool/End-Effector Frame**
+- [ ] **World/Global Frame**
+- [X] **Camera Frame**
+- [ ] **Other** (Please specify: `[Your Frame]`)
+
+**Action Dimensions:**
+```
+action: [x_left, y_left, z_left, quat_x_left,... quat_z_left, clutchBtn_left, energyBtn_left, thumbstickBtn_left, pince_left, thumbstick_x_left, thumbstick_y_left, ... (repeated for right)]
+
+- x,y,z: [float[3]] Absolute position of the hand controller (left or right) (metres)
+- quat_(x,y,z): [float[4]] Absolute orientation of the hand controller (left or right) (radians)
+- pince: [float from 0-1] value of the pince on the hand controller. Maps to the instrument jaws.
+- clutchBtn: [bool] Button input to clutch in/out. Engages and disengages a selected arm.
+- energyBtn: [bool] Button to activate electrosurgery.
+- thumbstickBtn: [bool] Button to activate ICG (Indocyanine green) visualisation.
+- thumbstick_x/y: [float[2]] Value of thumbstick - used to control the endoscope.
+```
+<div align="center">
+    <img src="./images/CMR_Versius_System_Dual_Hand_Controller.png"  alt="Versius Hand Controllers." width="600" />
+</div>
+
+*All actions come in `_left` and `_right` variants, which map to the left and right hand controllers.*
+
+### State Space Representation
+
+**State Information Included:**
+- [ ] **Joint Positions** (all articulated joints)
+- [ ] **Joint Velocities**
+- [ ] **End-Effector Pose** (Cartesian position/orientation)
+- [ ] **Force/Torque Readings**
+- [ ] **Gripper State** (position, force, etc.)
+- [X] **Other** (Please specify: `See below`)
+
+**State Dimensions:**
+
+For all state dimensions, the symbol `<S>` (Side) is used to denote a state that has `_left` and `_right` variants.
+```
+observation.state.haptic_<S>_armengageable: [bool[2]] Whether the arm selected by the left or right controller is in an engageable state. An arm may be enagageable without being actively engaged. Arms may be unengageable due to being in the wrong mode (e.g Instrument Change).
+
+observation.state.armlinkedtohaptic_<S>: [int[2]] Index of the arm connected to each hand controller at any one time.
+
+observation.state.hapticengaged_<S>: [bool[2]] Whether the haptic is in an engaged state. An engaged haptic means that its movements are currently driving a robotic arm.
+
+observation.state.arm_<int>_color: [int[5]] Color of each arm connected to the system. The color of each arm is shown on the HUD icons.
+
+observation.state.arm_<int>_instrtype: [int[5]] Instrument type attached to each arm.
+
+observation.state.rotationscaling: [float] Current scaling applied to the relative hand controller orientation. Can change during surgery.
+
+observation.state.translationScaling: [float] Current scaling applied to the relative hand controller position. Can change during surgery.
+
+observation.state.electroSurgerymode_<S>: [int[2]]: Electrosurgery mode (CUT, COAG) selected on each hand controller.
+
+observation.state.haptic_engaged_<S>: [bool[2]] Whether the haptic is engaged or not. When this state is false, the arms are not being controlled by the surgeon and are therefore not performing surgical tasks. The arms may still be moved by bedside assistants to retract the instrument or reposition the arm.
+
+observation.state.icgmode: [int] ICG (Indocyanine Green) mode selected (grayscale, overlay, heatmap).
+
+observation.state.icgenabled: [bool] ICG state, toggled with the right `thumbstickBtn`.
+observation.state.instrtype_<S>: [int[2]] Instrument attached to each hand controller (and in turn attached to an arm as per state.linkedArmIndex)
+```
+
+---
+
+## ⏱️ Data Synchronization Approach
+
+Versius telemetry and video is timestamped using the clock in its network controller. The network controller timestamps at the point of recording telemetry at 50Hz and video at 60Hz.
+Although both telemetry and video are timestamped using the same clock, there is a small variation in timing between the telemetry and video. As such, the system's HUD (Heads-Up Display) was used to synchronise telemetry and video.
+The timestamps of events in the HUD (including but not limited to: changes in arm mode, activation of electro-surgery and changes in electro-surgery mode) are compared against the timestamps of the corresponding events in telemetry. This
+provides us with a reliable measure of lag and the rate of dropped frames. The result of this synchronisation process is used to post-process the telemetry such that it is as close to perfectly in sync as possible.
+
+
+---
+
+## 👥 Attribution & Contact
+This dataset is shared with the aim of advancing research and development that has the potential to benefit patients through improved outcomes and increased access to minimal access surgery. 
+| | |
+| :--- | :--- |
+| **Dataset Lead** | Patrick Thornycroft, Diego Marana, Jonathon Hawkins, Filip Binkiewicz, Joyce Zhang |
+| **Institution** | CMR Surgical Ltd. |
+| **Contact Email** | patrick.thornycroft@cmrsurgical.com, diego.marana@cmrsurgical.com |
+| **Citation (BibTeX)** | N/A |

Surgical/hamlyn/knot_tying/README.md

@@ -0,0 +1,256 @@
+<!--
+Open-H Embodiment Dataset README Template (v1.0)
+Hamlyn Centre dVRK Dataset Submission
+-->
+
+# Hamlyn-dVRK: Ex-Vivo Knot Tying - README
+
+---
+
+## At a Glance
+
+High-fidelity teleoperated demonstrations of a bimanual **da Vinci Research Kit (dVRK)** robot performing **knot tying** on **ex-vivo porcine tissue**. The setup uses DeBakey forceps (left) and a needle driver (right) to manipulate suture strands and form stable surgical knots on deformable, wet tissue.
+
+
+---
+
+## Dataset Overview
+
+This dataset contains teleoperated trajectories of trained operators using the dVRK to perform knot tying on ex-vivo animal tissue. The goal is to reliably form a stable knot while maintaining appropriate strand tension and avoiding strand slip on wet tissue.
+
+**Task Logic & Execution:**
+The operator utilises the **Patient Side Manipulators (PSMs)** equipped with DeBakey forceps (left) and a needle driver (right).
+*   **Initial Condition:** The needle has already been passed through the tissue slit, and both suture ends are visible outside the tissue.
+*   **Knot Tying:** The robot grasps and manipulates the two strand ends bimanually to create the required crossings and throws, then tightens the knot while controlling tension and strand alignment.
+
+**Key Features:**
+*   **Pre-Threaded Start State:** The needle is already passed through the tissue slit and both suture ends are outside the tissue at the start of each trajectory, isolating knot tying from needle driving.
+*   **Bimanual Throw Formation and Tightening:** Demonstrations capture coordinated grasping, crossing, and controlled tightening required to form consistent knot throws on ex-vivo porcine tissue.
+
+| | |
+| :--- | :--- |
+| **Total Trajectories** | `77` |
+| **Total Hours** | `0.254` |
+| **Data Type** | `[ ] Clinical` `[x] Ex-Vivo` `[ ] Table-Top Phantom` `[ ] Digital Simulation` `[ ] Physical Simulation` `[ ] Other` |
+| **License** | CC BY 4.0 |
+| **Version** | `1.0` |
+
+
+## **Hamlyn dVRK Dataset Overview**
+
+This is part of the **Hamlyn dVRK Dataset**, which encompasses **six distinct surgical tasks**. All data were collected through teleoperation on bimanual tasks performed on ex-vivo animal samples and phantoms.
+
+### **Task Conditioning**
+The tasks are categorized into two types:
+- **Non-conditioned**: Single sentence task descriptions
+- **Language-conditioned**: Multiple variants of task descriptions across episodes
+
+### **Dataset Statistics**
+- **Total Episodes**: 972
+- **Total Frames**: 545k @ 30Hz
+- **Dataset Duration**: 5.04 hours
+
+### **Episode Outcomes**
+| Outcome Category | Episodes | Description |
+|------------------|----------|-------------|
+| **Success** | 780 | Task completed successfully with specified conditions |
+| **Recovery** | 110 | Task finished with recovery behavior |
+| **Failed** | 82 | Task failed to complete |
+
+*Success rates vary across tasks due to varying difficulty levels for operators during teleoperation.
+
+
+---
+
+## Tasks & Domain
+
+### Domain
+
+*Select the primary domain for this dataset.*
+
+- [x] **Surgical Robotics**
+- [ ] **Ultrasound Robotics**
+- [ ] **Other Healthcare Robotics**
+
+### Demonstrated Skills
+
+This specific dataset subset focuses on **Knot Tying**.
+
+*Note: This is part of the Hamlyn dVRK Data Collection which also includes:*
+- Tissue Retraction / Exposure
+- Knot Tying
+- Suturing (Single Loop)
+- Suturing (Dual Loop)
+- Peg Transfer
+- Needle Grasp and Handover
+
+---
+
+## Data Collection Details
+
+### Collection Method
+
+*How was the data collected?*
+
+- [x] **Human Teleoperation**
+- [ ] **Programmatic/State-Machine**
+- [ ] **AI Policy / Autonomous**
+- [ ] **Other**
+
+
+### Operator Details
+
+| | Description |
+| :--- | :--- |
+| **Operator Count** | 2 |
+| **Operator Skill Level** | `[ ] Expert (e.g., Surgeon, Sonographer)` <br> `[x] Intermediate (e.g., Trained Researcher)` <br> `[x] Novice (e.g., ML Researcher with minimal experience)` <br> `[ ] N/A`|
+| **Collection Period** | From `2025-12-01` to `2026-01-15` |
+| **Input Interface** | `Teleoperation interface with bimanual Force Dimension Sigma 7 haptic device`
+
+### Recovery Demonstrations
+
+*Does this dataset include examples of recovering from failure?*
+
+- [x] **Yes**
+- [ ] **No**
+
+**If yes, please briefly describe the recovery process:**
+
+In the knot tying task, failure modes typically involve losing strand tension, or an unstable throw formation. In these instances, the operator does not abort the episode but performs a **recovery manoeuvre**: re-grasping one or both strands, re-establishing tension, and repeating the throw and tightening steps to complete the knot.
+
+
+---
+
+## Diversity Dimensions
+
+*Check all dimensions that were intentionally varied during data collection.*
+
+- [ ] **Camera Position / Angle**
+- [ ] **Lighting Conditions**
+- [x] **Target Object** (Varied tissue geometry and suture configuration)
+- [x] **Spatial Layout** (Randomised tissue placement)
+- [ ] **Robot Embodiment**
+- [x] **Task Execution** (Left vs. Right hand dominance / Approach angle)
+- [ ] **Background / Scene**
+- [x] **Initial Robot Configuration** (Ex-Vivo domain randomisation)
+
+*Elaboration on Diversity:*
+
+*   **Target Object:** Fresh **porcine tissue** was used to capture wet-surface interaction. Diversity is achieved through variations in tissue stiffness/geometry and suture strand configuration (e.g., visible strand lengths and relative positioning).
+*   **Spatial Layout:** The tissue pose was manually randomised within the workspace between sets to reduce overfitting to absolute coordinates.
+*   **Task Execution:** Operators vary grasp points, crossing direction, and tightening strategy to maintain tension and produce stable throws.
+*   **Initial Robot Configuration:** End-effectors were reset to a randomised "home" position above the workspace before each trajectory.
+
+
+---
+
+## Equipment & Setup
+
+### Robotic Platform(s)
+
+*List the primary robot(s) used.*
+
+- **Robot 1:** `dVRK (da Vinci Research Kit) - Patient Side Manipulators (PSM1, PSM2)`
+
+
+### Sensors & Cameras
+
+*List the sensors and cameras used. Specify model names where possible.*
+
+| Type | Model/Details |
+| :--- | :--- |
+| **Primary Camera** | `Intel RealSense D405 (RGB + Depth), 848x480 @ 30fps` |
+| **Wrist Camera** | `INSKAM 5.5mm USB Endo-Cam (Right/Left Arm) with custom built mount, view angle 66 degree, focal length 4-10cm, 640x480 @ 24fps` |
+| **Kinematics** | `dVRK High-Frequency Joint Encoders (100Hz)` |
+
+
+---
+
+## 🎯 Action & State Space Representation
+
+*The dataset follows the standard LeRobot format for bimanual manipulation.*
+
+### Action Space Representation
+
+**Primary Action Representation:**
+- [x] **Absolute Cartesian** (position/orientation relative to robot base)
+- [ ] **Relative Cartesian**
+- [ ] **Joint Space**
+
+**Orientation Representation:**
+- [x] **Quaternions** (x, y, z, w)
+- [ ] **Euler Angles**
+- [ ] **Rotation Matrix**
+
+**Reference Frame:**
+- [x] **Robot Base Frame** (Base of each PSM arm)
+- [ ] **Camera Frame**
+
+**Action Dimensions:**
+
+action: [x, y, z, qx, qy, qz, qw, gripper_angle, x, y, z, qx, qy, qz, qw, gripper_angle]
+
+- The first 8 dimensions are for the *left* arm and last 8 dimensions are for the *right* arm
+- x, y, z: Absolute position in PSM base frame (meters)
+- qx, qy, qz, qw: Absolute orientation as quaternion
+- gripper_angle: Normalized jaw opening angle (1 = opened, 0 = closed)
+
+
+### State Space Representation
+
+**State Information Included:**
+- [x] **Joint Positions**
+- [x] **Joint Velocities**
+- [ ] **End-Effector Pose** (No extra end-effector used with dVRK)
+- [x] **Gripper State**
+
+**Primary State:**
+
+*observation.state*: [x, y, z, qx, qy, qz, qw, gripper_angle, x, y, z, qx, qy, qz, qw, gripper_angle]
+
+- The first 8 dimensions are for the *left* arm and last 8 dimensions are for the *right* arm
+- x, y, z: Absolute position in PSM base frame (meters)
+- qx, qy, qz, qw: Absolute orientation as quaternion
+- gripper_angle: Normalized jaw opening angle (1 = opened, 0 = closed)
+
+**State Dimensions:**
+
+`observation.state.left_arm_cartesian`: [x, y, z, qx, qy, qz, qw, gripper_angle]
+
+`observation.state.right_arm_cartesian`: [x, y, z, qx, qy, qz, qw, gripper_angle]
+
+`observation.state.left_arm_joint`: [j1, j2, j3, j4, j5, j6, gripper_angle]
+
+`observation.state.right_arm_joint`: [j1, j2, j3, j4, j5, j6, gripper_angle]
+
+- j1-j6: Joint positions for 6-DOF PSM arm (radians)
+- gripper_angle: Normalized jaw opening angle (1 = opened, 0 = closed)
+
+
+---
+
+## ⏱️ Data Synchronization Approach
+
+*Describe how you achieved proper data synchronization.*
+
+**Distributed Synchronization Architecture:**
+The data collection setup involves a distributed system with a robot control machine that teleop the dVRK and a sensor recording machine that stream kinematics, and videos from camers.
+1.  **Clock Sync:** We utilize **Chrony (NTP)** to enforce strict time alignment between the two workstations, minimizing clock skew to < 3.3 ms.
+2.  **Timestamping:** All data streams (camera RGB and depth images, wrist camer images, robot kinematics) are timestamped with **ROS wall-time** headers at the exact moment of capture.
+3.  **Delay Compensattion and Trimming:** The different cameras can introduce small relative delays. These delays are first estimated and compensated by applying a common temporal offset to each stream. Subsequently, all streams are temporally trimmed to their common interval, bounded by the latest start time and the earliest end time across all modalities.
+4.  **Clutch Masking:** As clutch is frequently used in teleoperation, sudden and meaningless breaks in the trajectory were removed to reduce stops during operation to improve trajectory smoothness.
+5.  **Alignment:** During the LeRobot conversion process, frames from all data streams are aligned to a reference timeline at 30 Hz30 Hz. This is achieved via nearest-neighbor temporal interpolation based on the synchronized ROS timestamps.
+
+---
+
+## Attribution & Contact
+
+*Please provide attribution for the dataset creators and a point of contact.*
+
+| | |
+| :--- | :--- |
+| **Dataset Lead** | `Jianhao Su, Kaizhong Deng, Zhaoyang Jacopo Hu` |
+| **Institution** | `The Hamlyn Centre for Robotic Surgery, Imperial College London` |
+| **Contact Email** | `k.deng21@imperial.ac.uk (for technical issues), stamatia.giannarou@imperial.ac.uk (for correspondence)` |
+| **Citation (BibTeX)** | <pre><code>@misc{hamlyn_dvrk_openh_2026,<br>  author = {Su, Jianhao and Deng, Kaizhong and Hu, Zhaoyang Jacopo and Rodriguez y Baena, Ferdinando and Giannarou, Stamatia},<br>  title = {Hamlyn dVRK Surgical Manipulation Dataset for Open-H},<br>  year = {2026},<br>  publisher = {Open-H-Embodiment},<br>  note = {Ex-Vivo Domain: Knot Tying}<br>}</code></pre> |
+

Surgical/hamlyn/needle_grasp_and_handover/README.md

@@ -0,0 +1,255 @@
+<!--
+Open-H Embodiment Dataset README Template (v1.0)
+Hamlyn Centre dVRK Dataset Submission
+-->
+
+# Hamlyn-dVRK: Table-Top Needle Grasp and Handover - README
+
+---
+
+## At a Glance
+
+High-fidelity teleoperated demonstrations of a bimanual **da Vinci Research Kit (dVRK)** robot performing **needle grasping and handover** on a **table-top setup**. The robot uses DeBakey forceps (left) and a needle driver (right) to pick up a surgical needle, optionally re-orient it in-hand using bimanual coordination, and hold it in a suturing-ready pose.
+
+
+---
+
+## Dataset Overview
+
+This dataset contains teleoperated trajectories of trained operators using the dVRK to acquire and prepare a needle for suturing on a table-top scene. The goal is to end with the needle held by the right hand in a stable, suturing-ready pose.
+
+**Task Logic & Execution:**
+The operator utilises the **Patient Side Manipulators (PSMs)** equipped with DeBakey forceps (left) and a needle driver (right).
+*   **Case A (Direct Right Grasp):** If the needle is in a suitable pose and location for the right hand, the right tool directly grasps the needle and moves it to a suturing-ready pose.
+*   **Case B (Left Assist + Re-Orientation):** If direct right grasp is unsuitable, the left tool first grasps the needle. The right tool then cooperates with the left in mid-air to adjust the needle pose, followed by a transfer to the right tool. The episode ends with the needle held by the right hand in a suturing-ready pose.
+
+**Key Features:**
+*   **Pose-Dependent Strategy Selection:** Demonstrations include both direct grasp and assisted re-orientation strategies depending on the initial needle pose.
+*   **Bimanual Mid-Air Manipulation:** Rich coordination signals for controlled needle handovers and in-hand pose adjustment.
+
+| | |
+| :--- | :--- |
+| **Total Trajectories** | `137` |
+| **Total Hours** | `0.245` |
+| **Data Type** | `[ ] Clinical` `[ ] Ex-Vivo` `[x] Table-Top Phantom` `[ ] Digital Simulation` `[ ] Physical Simulation` `[ ] Other` |
+| **License** | CC BY 4.0 |
+| **Version** | `1.0` |
+
+## **Hamlyn dVRK Dataset Overview**
+
+This is part of the **Hamlyn dVRK Dataset**, which encompasses **six distinct surgical tasks**. All data were collected through teleoperation on bimanual tasks performed on ex-vivo animal samples and phantoms.
+
+### **Task Conditioning**
+The tasks are categorized into two types:
+- **Non-conditioned**: Single sentence task descriptions
+- **Language-conditioned**: Multiple variants of task descriptions across episodes
+
+### **Dataset Statistics**
+- **Total Episodes**: 972
+- **Total Frames**: 545k @ 30Hz
+- **Dataset Duration**: 5.04 hours
+
+### **Episode Outcomes**
+| Outcome Category | Episodes | Description |
+|------------------|----------|-------------|
+| **Success** | 780 | Task completed successfully with specified conditions |
+| **Recovery** | 110 | Task finished with recovery behavior |
+| **Failed** | 82 | Task failed to complete |
+
+*Success rates vary across tasks due to varying difficulty levels for operators during teleoperation.
+
+
+---
+
+## Tasks & Domain
+
+### Domain
+
+*Select the primary domain for this dataset.*
+
+- [x] **Surgical Robotics**
+- [ ] **Ultrasound Robotics**
+- [ ] **Other Healthcare Robotics**
+
+### Demonstrated Skills
+
+This specific dataset subset focuses on **Needle Grasp and Handover**.
+
+*Note: This is part of the Hamlyn dVRK Data Collection which also includes:*
+- Tissue Retraction / Exposure
+- Knot Tying
+- Suturing (Single Loop)
+- Suturing (Dual Loop)
+- Peg Transfer
+- Needle Grasp and Handover
+
+---
+
+## Data Collection Details
+
+### Collection Method
+
+*How was the data collected?*
+
+- [x] **Human Teleoperation**
+- [ ] **Programmatic/State-Machine**
+- [ ] **AI Policy / Autonomous**
+- [ ] **Other**
+
+
+### Operator Details
+
+| | Description |
+| :--- | :--- |
+| **Operator Count** | 2 |
+| **Operator Skill Level** | `[ ] Expert (e.g., Surgeon, Sonographer)` <br> `[x] Intermediate (e.g., Trained Researcher)` <br> `[x] Novice (e.g., ML Researcher with minimal experience)` <br> `[ ] N/A`|
+| **Collection Period** | From `2025-12-01` to `2026-01-15` |
+| **Input Interface** | `Teleoperation interface with bimanual Force Dimension Sigma 7 haptic device`
+
+### Recovery Demonstrations
+
+*Does this dataset include examples of recovering from failure?*
+
+- [x] **Yes**
+- [ ] **No**
+
+**If yes, please briefly describe the recovery process:**
+
+In the needle grasp and handover task, failure modes typically involve an unsuccessful grasp, needle drop. In these instances, the operator does not abort the episode but performs a **recovery manoeuvre**: re-grasping the needle, adjusting the grasp point and tool orientation, and repeating the re-orientation and transfer steps to reach a suturing-ready pose.
+
+
+---
+
+## Diversity Dimensions
+
+*Check all dimensions that were intentionally varied during data collection.*
+
+- [ ] **Camera Position / Angle**
+- [ ] **Lighting Conditions**
+- [x] **Target Object** (Varied initial needle pose)
+- [x] **Spatial Layout** (Varied needle placement in workspace)
+- [ ] **Robot Embodiment**
+- [x] **Task Execution** (Left vs. Right hand dominance / Approach angle)
+- [ ] **Background / Scene**
+- [x] **Initial Robot Configuration** (Table-top domain randomisation)
+
+*Elaboration on Diversity:*
+
+*   **Target Object:** A surgical needle is used, with varied initial orientation and grasp affordances.
+*   **Spatial Layout:** The needle is placed at varied poses within the workspace to elicit either direct right-hand grasping or left-assisted re-orientation.
+*   **Task Execution:** Operators vary grasp points, mid-air re-orientation tactics, and transfer timing to achieve a stable final pose.
+*   **Initial Robot Configuration:** End-effectors were reset to a randomised "home" position above the workspace before each trajectory.
+
+
+---
+
+## 🛠️ Equipment & Setup
+
+### Robotic Platform(s)
+
+*List the primary robot(s) used.*
+
+- **Robot 1:** `dVRK (da Vinci Research Kit) - Patient Side Manipulators (PSM1, PSM2)`
+
+
+### Sensors & Cameras
+
+*List the sensors and cameras used. Specify model names where possible.*
+
+| Type | Model/Details |
+| :--- | :--- |
+| **Primary Camera** | `Intel RealSense D405 (RGB + Depth), 848x480 @ 30fps` |
+| **Wrist Camera** | `INSKAM 5.5mm USB Endo-Cam (Right/Left Arm) with custom built mount, view angle 66 degree, focal length 4-10cm, 640x480 @ 24fps` |
+| **Kinematics** | `dVRK High-Frequency Joint Encoders (100Hz)` |
+
+
+
+---
+
+## 🎯 Action & State Space Representation
+
+*The dataset follows the standard LeRobot format for bimanual manipulation.*
+
+### Action Space Representation
+
+**Primary Action Representation:**
+- [x] **Absolute Cartesian** (position/orientation relative to robot base)
+- [ ] **Relative Cartesian**
+- [ ] **Joint Space**
+
+**Orientation Representation:**
+- [x] **Quaternions** (x, y, z, w)
+- [ ] **Euler Angles**
+- [ ] **Rotation Matrix**
+
+**Reference Frame:**
+- [x] **Robot Base Frame** (Base of each PSM arm)
+- [ ] **Camera Frame**
+
+**Action Dimensions:**
+
+action: [x, y, z, qx, qy, qz, qw, gripper_angle, x, y, z, qx, qy, qz, qw, gripper_angle]
+
+- The first 8 dimensions are for the *left* arm and last 8 dimensions are for the *right* arm
+- x, y, z: Absolute position in PSM base frame (meters)
+- qx, qy, qz, qw: Absolute orientation as quaternion
+- gripper_angle: Normalized jaw opening angle (1 = opened, 0 = closed)
+
+
+### State Space Representation
+
+**State Information Included:**
+- [x] **Joint Positions**
+- [x] **Joint Velocities**
+- [ ] **End-Effector Pose** (No extra end-effector used with dVRK)
+- [x] **Gripper State**
+
+**Primary State:**
+
+*observation.state*: [x, y, z, qx, qy, qz, qw, gripper_angle, x, y, z, qx, qy, qz, qw, gripper_angle]
+
+- The first 8 dimensions are for the *left* arm and last 8 dimensions are for the *right* arm
+- x, y, z: Absolute position in PSM base frame (meters)
+- qx, qy, qz, qw: Absolute orientation as quaternion
+- gripper_angle: Normalized jaw opening angle (1 = opened, 0 = closed)
+
+**State Dimensions:**
+
+`observation.state.left_arm_cartesian`: [x, y, z, qx, qy, qz, qw, gripper_angle]
+
+`observation.state.right_arm_cartesian`: [x, y, z, qx, qy, qz, qw, gripper_angle]
+
+`observation.state.left_arm_joint`: [j1, j2, j3, j4, j5, j6, gripper_angle]
+
+`observation.state.right_arm_joint`: [j1, j2, j3, j4, j5, j6, gripper_angle]
+
+- j1-j6: Joint positions for 6-DOF PSM arm (radians)
+- gripper_angle: Normalized jaw opening angle (1 = opened, 0 = closed)
+
+
+---
+
+## ⏱️ Data Synchronization Approach
+
+*Describe how you achieved proper data synchronization.*
+
+**Distributed Synchronization Architecture:**
+The data collection setup involves a distributed system with a robot control machine that teleop the dVRK and a sensor recording machine that stream kinematics, and videos from camers.
+1.  **Clock Sync:** We utilize **Chrony (NTP)** to enforce strict time alignment between the two workstations, minimizing clock skew to < 3.3 ms.
+2.  **Timestamping:** All data streams (camera RGB and depth images, wrist camer images, robot kinematics) are timestamped with **ROS wall-time** headers at the exact moment of capture.
+3.  **Delay Compensattion and Trimming:** The different cameras can introduce small relative delays. These delays are first estimated and compensated by applying a common temporal offset to each stream. Subsequently, all streams are temporally trimmed to their common interval, bounded by the latest start time and the earliest end time across all modalities.
+4.  **Alignment:** During the LeRobot conversion process, frames from all data streams are aligned to a reference timeline at 30 Hz30 Hz. This is achieved via nearest-neighbor temporal interpolation based on the synchronized ROS timestamps.
+
+---
+
+## Attribution & Contact
+
+*Please provide attribution for the dataset creators and a point of contact.*
+
+| | |
+| :--- | :--- |
+| **Dataset Lead** | `Jianhao Su, Kaizhong Deng, Zhaoyang Jacopo Hu` |
+| **Institution** | `The Hamlyn Centre for Robotic Surgery, Imperial College London` |
+| **Contact Email** | `k.deng21@imperial.ac.uk (for technical issues), stamatia.giannarou@imperial.ac.uk (for correspondence)` |
+| **Citation (BibTeX)** | <pre><code>@misc{hamlyn_dvrk_openh_2026,<br>  author = {Su, Jianhao and Deng, Kaizhong and Hu, Zhaoyang Jacopo and Rodriguez y Baena, Ferdinando and Giannarou, Stamatia},<br>  title = {Hamlyn dVRK Surgical Manipulation Dataset for Open-H},<br>  year = {2026},<br>  publisher = {Open-H-Embodiment},<br>  note = {Table-Top Domain: Needle Grasp and Handover}<br>}</code></pre> |
+

Surgical/hamlyn/peg_transfer/README.md

@@ -0,0 +1,255 @@
+<!--
+Open-H Embodiment Dataset README Template (v1.0)
+Hamlyn Centre dVRK Dataset Submission
+-->
+
+# Hamlyn-dVRK: Table-Top Peg Transfer - README
+
+---
+
+## At a Glance
+
+High-fidelity teleoperated demonstrations of a bimanual **da Vinci Research Kit (dVRK)** robot performing **peg transfer** on a **table-top training rig**. The robot uses DeBakey forceps (left) and a needle driver (right) to pick up pegs, perform a handover between tools, and place pegs onto specified posts.
+
+
+---
+
+## Dataset Overview
+
+This dataset contains teleoperated trajectories of trained operators using the dVRK to perform peg transfer on a table-top training apparatus. The goal is to pick up a peg with the left hand, transfer it to the right hand, and place it accurately onto a target post according to a given instruction.
+
+**Task Logic & Execution:**
+The operator utilises the **Patient Side Manipulators (PSMs)** equipped with DeBakey forceps (left) and a needle driver (right).
+*   **Pick (Left Hand):** The left hand grasps the instructed peg.
+*   **Handover + Place (Right Hand):** The peg is transferred to the right hand and placed onto a target post on the board.
+
+**Key Features:**
+*   **Table-Top Dexterity Primitive:** Clean visual scene with rigid objects, supporting accurate benchmarking of bimanual handovers and placement.
+*   **Instruction-Conditioned Placement:** Demonstrations follow a given instruction specifying which peg to move and where to place it.
+
+| | |
+| :--- | :--- |
+| **Total Trajectories** | `317` |
+| **Total Hours** | `0.545` |
+| **Data Type** | `[ ] Clinical` `[ ] Ex-Vivo` `[x] Table-Top Phantom` `[ ] Digital Simulation` `[ ] Physical Simulation` `[ ] Other` |
+| **License** | CC BY 4.0 |
+| **Version** | `1.0` |
+
+## **Hamlyn dVRK Dataset Overview**
+
+This is part of the **Hamlyn dVRK Dataset**, which encompasses **six distinct surgical tasks**. All data were collected through teleoperation on bimanual tasks performed on ex-vivo animal samples and phantoms.
+
+### **Task Conditioning**
+The tasks are categorized into two types:
+- **Non-conditioned**: Single sentence task descriptions
+- **Language-conditioned**: Multiple variants of task descriptions across episodes
+
+### **Dataset Statistics**
+- **Total Episodes**: 972
+- **Total Frames**: 545k @ 30Hz
+- **Dataset Duration**: 5.04 hours
+
+### **Episode Outcomes**
+| Outcome Category | Episodes | Description |
+|------------------|----------|-------------|
+| **Success** | 780 | Task completed successfully with specified conditions |
+| **Recovery** | 110 | Task finished with recovery behavior |
+| **Failed** | 82 | Task failed to complete |
+
+*Success rates vary across tasks due to varying difficulty levels for operators during teleoperation.
+
+
+---
+
+## Tasks & Domain
+
+### Domain
+
+*Select the primary domain for this dataset.*
+
+- [x] **Surgical Robotics**
+- [ ] **Ultrasound Robotics**
+- [ ] **Other Healthcare Robotics**
+
+### Demonstrated Skills
+
+This specific dataset subset focuses on **Peg Transfer**.
+
+*Note: This is part of the Hamlyn dVRK Data Collection which also includes:*
+- Tissue Retraction / Exposure
+- Knot Tying
+- Suturing (Single Loop)
+- Suturing (Dual Loop)
+- Peg Transfer
+- Needle Grasp and Handover
+
+---
+
+## Data Collection Details
+
+### Collection Method
+
+*How was the data collected?*
+
+- [x] **Human Teleoperation**
+- [ ] **Programmatic/State-Machine**
+- [ ] **AI Policy / Autonomous**
+- [ ] **Other**
+
+
+### Operator Details
+
+| | Description |
+| :--- | :--- |
+| **Operator Count** | 2 |
+| **Operator Skill Level** | `[ ] Expert (e.g., Surgeon, Sonographer)` <br> `[x] Intermediate (e.g., Trained Researcher)` <br> `[x] Novice (e.g., ML Researcher with minimal experience)` <br> `[ ] N/A`|
+| **Collection Period** | From `2025-12-01` to `2026-01-15` |
+| **Input Interface** | `Teleoperation interface with bimanual Force Dimension Sigma 7 haptic device`
+
+### Recovery Demonstrations
+
+*Does this dataset include examples of recovering from failure?*
+
+- [x] **Yes**
+- [ ] **No**
+
+**If yes, please briefly describe the recovery process:**
+
+In the peg transfer task, failure modes typically involve dropping the peg during transport, an unstable handover, or misplacement onto the post. In these instances, the operator does not abort the episode but performs a **recovery manoeuvre**: re-grasping the peg, re-aligning the approach, and repeating the handover and placement steps.
+
+
+---
+
+## Diversity Dimensions
+
+*Check all dimensions that were intentionally varied during data collection.*
+
+- [ ] **Camera Position / Angle**
+- [ ] **Lighting Conditions**
+- [x] **Target Object** (Varied peg and target post)
+- [x] **Spatial Layout** (Varied board pose / peg arrangement)
+- [ ] **Robot Embodiment**
+- [x] **Task Execution** (Left vs. Right hand dominance / Approach angle)
+- [ ] **Background / Scene**
+- [x] **Initial Robot Configuration** (Table-top domain randomisation)
+
+*Elaboration on Diversity:*
+
+*   **Target Object:** A table-top peg board was used. Diversity is achieved through varying which peg is selected and which post is the placement target.
+*   **Spatial Layout:** The board pose and/or peg arrangement are varied between sets to reduce overfitting to absolute coordinates.
+*   **Task Execution:** Operators vary approach angle, grasp point, and handover strategy to maintain stable transport and accurate placement.
+*   **Initial Robot Configuration:** End-effectors were reset to a randomised "home" position above the workspace before each trajectory.
+
+
+---
+
+## 🛠️ Equipment & Setup
+
+### Robotic Platform(s)
+
+*List the primary robot(s) used.*
+
+- **Robot 1:** `dVRK (da Vinci Research Kit) - Patient Side Manipulators (PSM1, PSM2)`
+
+
+### Sensors & Cameras
+
+*List the sensors and cameras used. Specify model names where possible.*
+
+| Type | Model/Details |
+| :--- | :--- |
+| **Primary Camera** | `Intel RealSense D405 (RGB + Depth), 848x480 @ 30fps` |
+| **Wrist Camera** | `INSKAM 5.5mm USB Endo-Cam (Right/Left Arm) with custom built mount, view angle 66 degree, focal length 4-10cm, 640x480 @ 24fps` |
+| **Kinematics** | `dVRK High-Frequency Joint Encoders (100Hz)` |
+
+
+
+---
+
+## 🎯 Action & State Space Representation
+
+*The dataset follows the standard LeRobot format for bimanual manipulation.*
+
+### Action Space Representation
+
+**Primary Action Representation:**
+- [x] **Absolute Cartesian** (position/orientation relative to robot base)
+- [ ] **Relative Cartesian**
+- [ ] **Joint Space**
+
+**Orientation Representation:**
+- [x] **Quaternions** (x, y, z, w)
+- [ ] **Euler Angles**
+- [ ] **Rotation Matrix**
+
+**Reference Frame:**
+- [x] **Robot Base Frame** (Base of each PSM arm)
+- [ ] **Camera Frame**
+
+**Action Dimensions:**
+
+action: [x, y, z, qx, qy, qz, qw, gripper_angle, x, y, z, qx, qy, qz, qw, gripper_angle]
+
+- The first 8 dimensions are for the *left* arm and last 8 dimensions are for the *right* arm
+- x, y, z: Absolute position in PSM base frame (meters)
+- qx, qy, qz, qw: Absolute orientation as quaternion
+- gripper_angle: Normalized jaw opening angle (1 = opened, 0 = closed)
+
+
+### State Space Representation
+
+**State Information Included:**
+- [x] **Joint Positions**
+- [x] **Joint Velocities**
+- [ ] **End-Effector Pose** (No extra end-effector used with dVRK)
+- [x] **Gripper State**
+
+**Primary State:**
+
+*observation.state*: [x, y, z, qx, qy, qz, qw, gripper_angle, x, y, z, qx, qy, qz, qw, gripper_angle]
+
+- The first 8 dimensions are for the *left* arm and last 8 dimensions are for the *right* arm
+- x, y, z: Absolute position in PSM base frame (meters)
+- qx, qy, qz, qw: Absolute orientation as quaternion
+- gripper_angle: Normalized jaw opening angle (1 = opened, 0 = closed)
+
+**State Dimensions:**
+
+`observation.state.left_arm_cartesian`: [x, y, z, qx, qy, qz, qw, gripper_angle]
+
+`observation.state.right_arm_cartesian`: [x, y, z, qx, qy, qz, qw, gripper_angle]
+
+`observation.state.left_arm_joint`: [j1, j2, j3, j4, j5, j6, gripper_angle]
+
+`observation.state.right_arm_joint`: [j1, j2, j3, j4, j5, j6, gripper_angle]
+
+- j1-j6: Joint positions for 6-DOF PSM arm (radians)
+- gripper_angle: Normalized jaw opening angle (1 = opened, 0 = closed)
+
+
+---
+
+## ⏱️ Data Synchronization Approach
+
+*Describe how you achieved proper data synchronization.*
+
+**Distributed Synchronization Architecture:**
+The data collection setup involves a distributed system with a robot control machine that teleop the dVRK and a sensor recording machine that stream kinematics, and videos from camers.
+1.  **Clock Sync:** We utilize **Chrony (NTP)** to enforce strict time alignment between the two workstations, minimizing clock skew to < 3.3 ms.
+2.  **Timestamping:** All data streams (camera RGB and depth images, wrist camer images, robot kinematics) are timestamped with **ROS wall-time** headers at the exact moment of capture.
+3.  **Delay Compensattion and Trimming:** The different cameras can introduce small relative delays. These delays are first estimated and compensated by applying a common temporal offset to each stream. Subsequently, all streams are temporally trimmed to their common interval, bounded by the latest start time and the earliest end time across all modalities.
+4.  **Alignment:** During the LeRobot conversion process, frames from all data streams are aligned to a reference timeline at 30 Hz30 Hz. This is achieved via nearest-neighbor temporal interpolation based on the synchronized ROS timestamps.
+
+---
+
+## Attribution & Contact
+
+*Please provide attribution for the dataset creators and a point of contact.*
+
+| | |
+| :--- | :--- |
+| **Dataset Lead** | `Jianhao Su, Kaizhong Deng, Zhaoyang Jacopo Hu` |
+| **Institution** | `The Hamlyn Centre for Robotic Surgery, Imperial College London` |
+| **Contact Email** | `k.deng21@imperial.ac.uk (for technical issues), stamatia.giannarou@imperial.ac.uk (for correspondence)` |
+| **Citation (BibTeX)** | <pre><code>@misc{hamlyn_dvrk_openh_2026,<br>  author = {Su, Jianhao and Deng, Kaizhong and Hu, Zhaoyang Jacopo and Rodriguez y Baena, Ferdinando and Giannarou, Stamatia},<br>  title = {Hamlyn dVRK Surgical Manipulation Dataset for Open-H},<br>  year = {2026},<br>  publisher = {Open-H-Embodiment},<br>  note = {Table-Top Domain: Peg Transfer}<br>}</code></pre> |
+

Surgical/hamlyn/suturing_1/README.md

@@ -0,0 +1,255 @@
+<!--
+Open-H Embodiment Dataset README Template (v1.0)
+Hamlyn Centre dVRK Dataset Submission
+-->
+
+# Hamlyn-dVRK: Ex-Vivo Suturing (Single Loop) - README
+
+---
+
+## At a Glance
+
+High-fidelity teleoperated demonstrations of a bimanual **da Vinci Research Kit (dVRK)** robot performing **suturing (single loop)** on **ex-vivo porcine tissue**. The setup uses DeBakey forceps (left) and a needle driver (right) to execute a single needle pass through tissue, including a bimanual needle extraction handover.
+
+
+---
+
+## Dataset Overview
+
+This dataset contains teleoperated trajectories of trained operators using the dVRK to perform a single-loop suturing action on ex-vivo animal tissue. The goal is to perform one needle pass through tissue with correct needle orientation and controlled interaction with deformable, wet tissue.
+
+**Task Logic & Execution:**
+The operator utilises the **Patient Side Manipulators (PSMs)** equipped with DeBakey forceps (left) and a needle driver (right).
+*   **Needle Hold (Right Hand):** The right tool grasps the needle in a suturing-ready pose.
+*   **Single Pass + Extraction (Left Hand):** The right hand drives the needle through the tissue, then the left hand grasps and extracts the needle from the far side.
+
+**Key Features:**
+*   **Real Ex-Vivo Needle-Tissue Interaction:** Uses **porcine tissue** to capture compliance and deformation during needle driving.
+*   **Bimanual Needle Handover:** Demonstrations include the right-to-left extraction handover required to complete a single pass.
+
+| | |
+| :--- | :--- |
+| **Total Trajectories** | `180` |
+| **Total Hours** | `0.476` |
+| **Data Type** | `[ ] Clinical` `[x] Ex-Vivo` `[ ] Table-Top Phantom` `[ ] Digital Simulation` `[ ] Physical Simulation` `[ ] Other` |
+| **License** | CC BY 4.0 |
+| **Version** | `1.0` |
+
+## **Hamlyn dVRK Dataset Overview**
+
+This is part of the **Hamlyn dVRK Dataset**, which encompasses **six distinct surgical tasks**. All data were collected through teleoperation on bimanual tasks performed on ex-vivo animal samples and phantoms.
+
+### **Task Conditioning**
+The tasks are categorized into two types:
+- **Non-conditioned**: Single sentence task descriptions
+- **Language-conditioned**: Multiple variants of task descriptions across episodes
+
+### **Dataset Statistics**
+- **Total Episodes**: 972
+- **Total Frames**: 545k @ 30Hz
+- **Dataset Duration**: 5.04 hours
+
+### **Episode Outcomes**
+| Outcome Category | Episodes | Description |
+|------------------|----------|-------------|
+| **Success** | 780 | Task completed successfully with specified conditions |
+| **Recovery** | 110 | Task finished with recovery behavior |
+| **Failed** | 82 | Task failed to complete |
+
+*Success rates vary across tasks due to varying difficulty levels for operators during teleoperation.
+
+
+---
+
+## Tasks & Domain
+
+### Domain
+
+*Select the primary domain for this dataset.*
+
+- [x] **Surgical Robotics**
+- [ ] **Ultrasound Robotics**
+- [ ] **Other Healthcare Robotics**
+
+### Demonstrated Skills
+
+This specific dataset subset focuses on **Suturing (Single Loop)**.
+
+*Note: This is part of the Hamlyn dVRK Data Collection which also includes:*
+- Tissue Retraction / Exposure
+- Knot Tying
+- Suturing (Single Loop)
+- Suturing (Dual Loop)
+- Peg Transfer
+- Needle Grasp and Handover
+
+---
+
+## Data Collection Details
+
+### Collection Method
+
+*How was the data collected?*
+
+- [x] **Human Teleoperation**
+- [ ] **Programmatic/State-Machine**
+- [ ] **AI Policy / Autonomous**
+- [ ] **Other**
+
+
+### Operator Details
+
+| | Description |
+| :--- | :--- |
+| **Operator Count** | 2 |
+| **Operator Skill Level** | `[ ] Expert (e.g., Surgeon, Sonographer)` <br> `[x] Intermediate (e.g., Trained Researcher)` <br> `[x] Novice (e.g., ML Researcher with minimal experience)` <br> `[ ] N/A`|
+| **Collection Period** | From `2025-12-01` to `2026-01-15` |
+| **Input Interface** | `Teleoperation interface with bimanual Force Dimension Sigma 7 haptic device`
+
+### Recovery Demonstrations
+
+*Does this dataset include examples of recovering from failure?*
+
+- [x] **Yes**
+- [ ] **No**
+
+**If yes, please briefly describe the recovery process:**
+
+In the single-loop suturing task, failure modes typically involve an incorrect bite, needle drop, loss of needle control during extraction, or unstable tissue interaction. In these instances, the operator does not abort the episode but performs a **recovery manoeuvre**: re-grasping the needle, re-aligning the approach, and repeating the pass and extraction steps to complete the stitch.
+
+
+---
+
+## Diversity Dimensions
+
+*Check all dimensions that were intentionally varied during data collection.*
+
+- [ ] **Camera Position / Angle**
+- [ ] **Lighting Conditions**
+- [x] **Target Object** (Varied tissue geometry and stitch location)
+- [x] **Spatial Layout** (Randomised tissue placement)
+- [ ] **Robot Embodiment**
+- [x] **Task Execution** (Left vs. Right hand dominance / Approach angle)
+- [ ] **Background / Scene**
+- [x] **Initial Robot Configuration** (Ex-Vivo domain randomisation)
+
+*Elaboration on Diversity:*
+
+*   **Target Object:** Fresh **porcine tissue** was used to capture deformation during needle driving. Diversity is achieved through variations in tissue stiffness/geometry and stitch location.
+*   **Spatial Layout:** The tissue pose was manually randomised within the workspace between sets to reduce overfitting to absolute coordinates.
+*   **Task Execution:** Operators vary entry angle, wrist orientation, and extraction strategy to maintain stable needle control.
+*   **Initial Robot Configuration:** End-effectors were reset to a randomised "home" position above the workspace before each trajectory.
+
+
+---
+
+## 🛠️ Equipment & Setup
+
+### Robotic Platform(s)
+
+*List the primary robot(s) used.*
+
+- **Robot 1:** `dVRK (da Vinci Research Kit) - Patient Side Manipulators (PSM1, PSM2)`
+
+
+### Sensors & Cameras
+
+*List the sensors and cameras used. Specify model names where possible.*
+
+| Type | Model/Details |
+| :--- | :--- |
+| **Primary Camera** | `Intel RealSense D405 (RGB + Depth), 848x480 @ 30fps` |
+| **Wrist Camera** | `INSKAM 5.5mm USB Endo-Cam (Right/Left Arm) with custom built mount, view angle 66 degree, focal length 4-10cm, 640x480 @ 24fps` |
+| **Kinematics** | `dVRK High-Frequency Joint Encoders (100Hz)` |
+
+
+
+---
+
+## 🎯 Action & State Space Representation
+
+*The dataset follows the standard LeRobot format for bimanual manipulation.*
+
+### Action Space Representation
+
+**Primary Action Representation:**
+- [x] **Absolute Cartesian** (position/orientation relative to robot base)
+- [ ] **Relative Cartesian**
+- [ ] **Joint Space**
+
+**Orientation Representation:**
+- [x] **Quaternions** (x, y, z, w)
+- [ ] **Euler Angles**
+- [ ] **Rotation Matrix**
+
+**Reference Frame:**
+- [x] **Robot Base Frame** (Base of each PSM arm)
+- [ ] **Camera Frame**
+
+**Action Dimensions:**
+
+action: [x, y, z, qx, qy, qz, qw, gripper_angle, x, y, z, qx, qy, qz, qw, gripper_angle]
+
+- The first 8 dimensions are for the *left* arm and last 8 dimensions are for the *right* arm
+- x, y, z: Absolute position in PSM base frame (meters)
+- qx, qy, qz, qw: Absolute orientation as quaternion
+- gripper_angle: Normalized jaw opening angle (1 = opened, 0 = closed)
+
+
+### State Space Representation
+
+**State Information Included:**
+- [x] **Joint Positions**
+- [x] **Joint Velocities**
+- [ ] **End-Effector Pose** (No extra end-effector used with dVRK)
+- [x] **Gripper State**
+
+**Primary State:**
+
+*observation.state*: [x, y, z, qx, qy, qz, qw, gripper_angle, x, y, z, qx, qy, qz, qw, gripper_angle]
+
+- The first 8 dimensions are for the *left* arm and last 8 dimensions are for the *right* arm
+- x, y, z: Absolute position in PSM base frame (meters)
+- qx, qy, qz, qw: Absolute orientation as quaternion
+- gripper_angle: Normalized jaw opening angle (1 = opened, 0 = closed)
+
+**State Dimensions:**
+
+`observation.state.left_arm_cartesian`: [x, y, z, qx, qy, qz, qw, gripper_angle]
+
+`observation.state.right_arm_cartesian`: [x, y, z, qx, qy, qz, qw, gripper_angle]
+
+`observation.state.left_arm_joint`: [j1, j2, j3, j4, j5, j6, gripper_angle]
+
+`observation.state.right_arm_joint`: [j1, j2, j3, j4, j5, j6, gripper_angle]
+
+- j1-j6: Joint positions for 6-DOF PSM arm (radians)
+- gripper_angle: Normalized jaw opening angle (1 = opened, 0 = closed)
+
+
+---
+
+## ⏱️ Data Synchronization Approach
+
+*Describe how you achieved proper data synchronization.*
+
+**Distributed Synchronization Architecture:**
+The data collection setup involves a distributed system with a robot control machine that teleop the dVRK and a sensor recording machine that stream kinematics, and videos from camers.
+1.  **Clock Sync:** We utilize **Chrony (NTP)** to enforce strict time alignment between the two workstations, minimizing clock skew to < 3.3 ms.
+2.  **Timestamping:** All data streams (camera RGB and depth images, wrist camer images, robot kinematics) are timestamped with **ROS wall-time** headers at the exact moment of capture.
+3.  **Delay Compensattion and Trimming:** The different cameras can introduce small relative delays. These delays are first estimated and compensated by applying a common temporal offset to each stream. Subsequently, all streams are temporally trimmed to their common interval, bounded by the latest start time and the earliest end time across all modalities.
+4.  **Alignment:** During the LeRobot conversion process, frames from all data streams are aligned to a reference timeline at 30 Hz30 Hz. This is achieved via nearest-neighbor temporal interpolation based on the synchronized ROS timestamps.
+
+---
+
+## Attribution & Contact
+
+*Please provide attribution for the dataset creators and a point of contact.*
+
+| | |
+| :--- | :--- |
+| **Dataset Lead** | `Jianhao Su, Kaizhong Deng, Zhaoyang Jacopo Hu` |
+| **Institution** | `The Hamlyn Centre for Robotic Surgery, Imperial College London` |
+| **Contact Email** | `k.deng21@imperial.ac.uk (for technical issues), stamatia.giannarou@imperial.ac.uk (for correspondence)` |
+| **Citation (BibTeX)** | <pre><code>@misc{hamlyn_dvrk_openh_2026,<br>  author = {Su, Jianhao and Deng, Kaizhong and Hu, Zhaoyang Jacopo and Rodriguez y Baena, Ferdinando and Giannarou, Stamatia},<br>  title = {Hamlyn dVRK Surgical Manipulation Dataset for Open-H},<br>  year = {2026},<br>  publisher = {Open-H-Embodiment},<br>  note = {Ex-Vivo Domain: Suturing (Single Loop)}<br>}</code></pre> |
+

Surgical/hamlyn/suturing_2/README.md

@@ -0,0 +1,255 @@
+<!--
+Open-H Embodiment Dataset README Template (v1.0)
+Hamlyn Centre dVRK Dataset Submission
+-->
+
+# Hamlyn-dVRK: Ex-Vivo Suturing (Dual Loop) - README
+
+---
+
+## At a Glance
+
+High-fidelity teleoperated demonstrations of a bimanual **da Vinci Research Kit (dVRK)** robot performing **suturing (dual loop)** on **ex-vivo porcine tissue**. This task captures a continuous, multi-step suturing sequence with repeated needle passes and bimanual handovers using DeBakey forceps (left) and a needle driver (right).
+
+
+---
+
+## Dataset Overview
+
+This dataset contains teleoperated trajectories of trained operators using the dVRK to perform a dual-loop (continuous) suturing sequence on ex-vivo animal tissue. The goal is to complete two consecutive needle passes through tissue, including all intermediate steps needed to keep the needle in a suturing-ready pose.
+
+**Task Logic & Execution:**
+The operator utilises the **Patient Side Manipulators (PSMs)** equipped with DeBakey forceps (left) and a needle driver (right).
+*   **First Stitch:** The right hand drives the needle through tissue; the left hand grasps and extracts the needle.
+*   **Needle Return + Second Stitch:** The left hand transfers the needle back to the right hand in a suturing-ready pose, then the right hand performs a second pass through tissue followed by left-hand extraction.
+
+**Key Features:**
+*   **Real Ex-Vivo Interaction:** Uses **porcine tissue** to capture compliance and deformation during repeated needle passes.
+*   **Continuous Multi-Step Sequence:** Demonstrations include bimanual extraction, mid-air re-orientation, handover back to the right, and a second stitch without resetting the episode.
+
+| | |
+| :--- | :--- |
+| **Total Trajectories** | `186` |
+| **Total Hours** | `1.189` |
+| **Data Type** | `[ ] Clinical` `[x] Ex-Vivo` `[ ] Table-Top Phantom` `[ ] Digital Simulation` `[ ] Physical Simulation` `[ ] Other` |
+| **License** | CC BY 4.0 |
+| **Version** | `1.0` |
+
+## **Hamlyn dVRK Dataset Overview**
+
+This is part of the **Hamlyn dVRK Dataset**, which encompasses **six distinct surgical tasks**. All data were collected through teleoperation on bimanual tasks performed on ex-vivo animal samples and phantoms.
+
+### **Task Conditioning**
+The tasks are categorized into two types:
+- **Non-conditioned**: Single sentence task descriptions
+- **Language-conditioned**: Multiple variants of task descriptions across episodes
+
+### **Dataset Statistics**
+- **Total Episodes**: 972
+- **Total Frames**: 545k @ 30Hz
+- **Dataset Duration**: 5.04 hours
+
+### **Episode Outcomes**
+| Outcome Category | Episodes | Description |
+|------------------|----------|-------------|
+| **Success** | 780 | Task completed successfully with specified conditions |
+| **Recovery** | 110 | Task finished with recovery behavior |
+| **Failed** | 82 | Task failed to complete |
+
+*Success rates vary across tasks due to varying difficulty levels for operators during teleoperation.
+
+
+---
+
+## Tasks & Domain
+
+### Domain
+
+*Select the primary domain for this dataset.*
+
+- [x] **Surgical Robotics**
+- [ ] **Ultrasound Robotics**
+- [ ] **Other Healthcare Robotics**
+
+### Demonstrated Skills
+
+This specific dataset subset focuses on **Suturing (Dual Loop)**.
+
+*Note: This is part of the Hamlyn dVRK Data Collection which also includes:*
+- Tissue Retraction / Exposure
+- Knot Tying
+- Suturing (Single Loop)
+- Suturing (Dual Loop)
+- Peg Transfer
+- Needle Grasp and Handover
+
+---
+
+## Data Collection Details
+
+### Collection Method
+
+*How was the data collected?*
+
+- [x] **Human Teleoperation**
+- [ ] **Programmatic/State-Machine**
+- [ ] **AI Policy / Autonomous**
+- [ ] **Other**
+
+
+### Operator Details
+
+| | Description |
+| :--- | :--- |
+| **Operator Count** | 2 |
+| **Operator Skill Level** | `[ ] Expert (e.g., Surgeon, Sonographer)` <br> `[x] Intermediate (e.g., Trained Researcher)` <br> `[x] Novice (e.g., ML Researcher with minimal experience)` <br> `[ ] N/A`|
+| **Collection Period** | From `2025-12-01` to `2026-01-15` |
+| **Input Interface** | `Teleoperation interface with bimanual Force Dimension Sigma 7 haptic device`
+
+### Recovery Demonstrations
+
+*Does this dataset include examples of recovering from failure?*
+
+- [x] **Yes**
+- [ ] **No**
+
+**If yes, please briefly describe the recovery process:**
+
+In the dual-loop suturing task, failure modes typically involve an incorrect bite, needle drop, loss of needle control during transfer, or unstable tissue interaction. In these instances, the operator does not abort the episode but performs a **recovery manoeuvre**: re-grasping and re-orienting the needle, re-aligning the approach, and repeating the necessary intermediate steps before continuing the sequence.
+
+
+---
+
+## Diversity Dimensions
+
+*Check all dimensions that were intentionally varied during data collection.*
+
+- [ ] **Camera Position / Angle**
+- [ ] **Lighting Conditions**
+- [x] **Target Object** (Varied tissue geometry and stitch locations)
+- [x] **Spatial Layout** (Randomised tissue placement)
+- [ ] **Robot Embodiment**
+- [x] **Task Execution** (Left vs. Right hand dominance / Approach angle)
+- [ ] **Background / Scene**
+- [x] **Initial Robot Configuration** (Ex-Vivo domain randomisation)
+
+*Elaboration on Diversity:*
+
+*   **Target Object:** Fresh **porcine tissue** was used to capture deformation during repeated needle driving. Diversity is achieved through variations in tissue stiffness/geometry and stitch locations for the two consecutive passes.
+*   **Spatial Layout:** The tissue pose was manually randomised within the workspace between sets to reduce overfitting to absolute coordinates.
+*   **Task Execution:** Operators vary entry angle, transfer strategy, and re-orientation in mid-air to keep the needle in a stable pose across the sequence.
+*   **Initial Robot Configuration:** End-effectors were reset to a randomised "home" position above the workspace before each trajectory.
+
+
+---
+
+## 🛠️ Equipment & Setup
+
+### Robotic Platform(s)
+
+*List the primary robot(s) used.*
+
+- **Robot 1:** `dVRK (da Vinci Research Kit) - Patient Side Manipulators (PSM1, PSM2)`
+
+
+### Sensors & Cameras
+
+*List the sensors and cameras used. Specify model names where possible.*
+
+| Type | Model/Details |
+| :--- | :--- |
+| **Primary Camera** | `Intel RealSense D405 (RGB + Depth), 848x480 @ 30fps` |
+| **Wrist Camera** | `INSKAM 5.5mm USB Endo-Cam (Right/Left Arm) with custom built mount, view angle 66 degree, focal length 4-10cm, 640x480 @ 24fps` |
+| **Kinematics** | `dVRK High-Frequency Joint Encoders (100Hz)` |
+
+
+
+---
+
+## 🎯 Action & State Space Representation
+
+*The dataset follows the standard LeRobot format for bimanual manipulation.*
+
+### Action Space Representation
+
+**Primary Action Representation:**
+- [x] **Absolute Cartesian** (position/orientation relative to robot base)
+- [ ] **Relative Cartesian**
+- [ ] **Joint Space**
+
+**Orientation Representation:**
+- [x] **Quaternions** (x, y, z, w)
+- [ ] **Euler Angles**
+- [ ] **Rotation Matrix**
+
+**Reference Frame:**
+- [x] **Robot Base Frame** (Base of each PSM arm)
+- [ ] **Camera Frame**
+
+**Action Dimensions:**
+
+action: [x, y, z, qx, qy, qz, qw, gripper_angle, x, y, z, qx, qy, qz, qw, gripper_angle]
+
+- The first 8 dimensions are for the *left* arm and last 8 dimensions are for the *right* arm
+- x, y, z: Absolute position in PSM base frame (meters)
+- qx, qy, qz, qw: Absolute orientation as quaternion
+- gripper_angle: Normalized jaw opening angle (1 = opened, 0 = closed)
+
+
+### State Space Representation
+
+**State Information Included:**
+- [x] **Joint Positions**
+- [x] **Joint Velocities**
+- [ ] **End-Effector Pose** (No extra end-effector used with dVRK)
+- [x] **Gripper State**
+
+**Primary State:**
+
+*observation.state*: [x, y, z, qx, qy, qz, qw, gripper_angle, x, y, z, qx, qy, qz, qw, gripper_angle]
+
+- The first 8 dimensions are for the *left* arm and last 8 dimensions are for the *right* arm
+- x, y, z: Absolute position in PSM base frame (meters)
+- qx, qy, qz, qw: Absolute orientation as quaternion
+- gripper_angle: Normalized jaw opening angle (1 = opened, 0 = closed)
+
+**State Dimensions:**
+
+`observation.state.left_arm_cartesian`: [x, y, z, qx, qy, qz, qw, gripper_angle]
+
+`observation.state.right_arm_cartesian`: [x, y, z, qx, qy, qz, qw, gripper_angle]
+
+`observation.state.left_arm_joint`: [j1, j2, j3, j4, j5, j6, gripper_angle]
+
+`observation.state.right_arm_joint`: [j1, j2, j3, j4, j5, j6, gripper_angle]
+
+- j1-j6: Joint positions for 6-DOF PSM arm (radians)
+- gripper_angle: Normalized jaw opening angle (1 = opened, 0 = closed)
+
+
+---
+
+## ⏱️ Data Synchronization Approach
+
+*Describe how you achieved proper data synchronization.*
+
+**Distributed Synchronization Architecture:**
+The data collection setup involves a distributed system with a robot control machine that teleop the dVRK and a sensor recording machine that stream kinematics, and videos from camers.
+1.  **Clock Sync:** We utilize **Chrony (NTP)** to enforce strict time alignment between the two workstations, minimizing clock skew to < 3.3 ms.
+2.  **Timestamping:** All data streams (camera RGB and depth images, wrist camer images, robot kinematics) are timestamped with **ROS wall-time** headers at the exact moment of capture.
+3.  **Delay Compensattion and Trimming:** The different cameras can introduce small relative delays. These delays are first estimated and compensated by applying a common temporal offset to each stream. Subsequently, all streams are temporally trimmed to their common interval, bounded by the latest start time and the earliest end time across all modalities.
+4.  **Alignment:** During the LeRobot conversion process, frames from all data streams are aligned to a reference timeline at 30 Hz30 Hz. This is achieved via nearest-neighbor temporal interpolation based on the synchronized ROS timestamps.
+
+---
+
+## Attribution & Contact
+
+*Please provide attribution for the dataset creators and a point of contact.*
+
+| | |
+| :--- | :--- |
+| **Dataset Lead** | `Jianhao Su, Kaizhong Deng, Zhaoyang Jacopo Hu` |
+| **Institution** | `The Hamlyn Centre for Robotic Surgery, Imperial College London` |
+| **Contact Email** | `k.deng21@imperial.ac.uk (for technical issues), stamatia.giannarou@imperial.ac.uk (for correspondence)` |
+| **Citation (BibTeX)** | <pre><code>@misc{hamlyn_dvrk_openh_2026,<br>  author = {Su, Jianhao and Deng, Kaizhong and Hu, Zhaoyang Jacopo and Rodriguez y Baena, Ferdinando and Giannarou, Stamatia},<br>  title = {Hamlyn dVRK Surgical Manipulation Dataset for Open-H},<br>  year = {2026},<br>  publisher = {Open-H-Embodiment},<br>  note = {Ex-Vivo Domain: Suturing (Dual Loop)}<br>}</code></pre> |
+

Surgical/hamlyn/tissue_retraction/README.md

@@ -0,0 +1,257 @@
+<!--
+Open-H Embodiment Dataset README Template (v1.0)
+Hamlyn Centre dVRK Dataset Submission
+-->
+
+# Hamlyn-dVRK: Ex-Vivo Tissue Retraction - README
+
+---
+
+## At a Glance
+
+High-fidelity teleoperated demonstrations of a bimanual **da Vinci Research Kit (dVRK)** robot performing tissue retraction and exposure tasks on **ex-vivo porcine tissue**. This dataset captures the nuanced physical interaction between rigid surgical instruments (DeBakey forceps and Needle Drivers) and deformable, wet biological environments. It specifically highlights the challenge of maintaining stable grasps on slippery tissue surfaces while executing precise lifting maneuvers to expose underlying anatomical planes.
+
+
+---
+
+## Dataset Overview
+
+This dataset contains **75 trajectories** of trained operators using the dVRK to perform tissue retraction tasks on ex-vivo animal tissue. The goal is to expose underlying structures or distinct tissue planes, a fundamental primitive in surgery.
+
+**Task Logic & Execution:**
+The operator utilizes the **Patient Side Manipulators (PSMs)** equipped with DeBakey Forceps (Left) and a Needle Driver (Right).
+*   **Partially Dissected Tissue:** Mimics the retraction workflow during the dissection of porcine tissue loops. The operator must grasp the separated tissue flap and retract the upper layer towards a specific direction to assist with visual exposure.
+*   **Intact Tissue:** Involves unseparated tissue blocks. The operator selects one of three predefined retraction grasp locations (**left / middle / right**), establishes a grasp at the specified location, and applies upward traction to lift the tissue mass.
+
+**Key Features:**
+*   **Real Ex-Vivo Texture:** Utilizes **porcine meat** to capture authentic physical properties, specifically **tissue wetness**, specular reflections, and complex **soft-body deformations** that are difficult to simulate.
+*   **Robustness Data:** Contains explicitly labeled successful trials, failures (e.g., tissue slip), and **recovery attempts** (re-grasping), providing rich data for training error-correction policies.
+
+| | |
+| :--- | :--- |
+| **Total Trajectories** | `75` |
+| **Total Hours** | `0.076` |
+| **Data Type** | `[ ] Clinical` `[x] Ex-Vivo` `[ ] Table-Top Phantom` `[ ] Digital Simulation` `[ ] Physical Simulation` `[ ] Other` |
+| **License** | CC BY 4.0 |
+| **Version** | `1.0` |
+
+## **Hamlyn dVRK Dataset Overview**
+
+This is part of the **Hamlyn dVRK Dataset**, which encompasses **six distinct surgical tasks**. All data were collected through teleoperation on bimanual tasks performed on ex-vivo animal samples and phantoms.
+
+### **Task Conditioning**
+The tasks are categorized into two types:
+- **Non-conditioned**: Single sentence task descriptions
+- **Language-conditioned**: Multiple variants of task descriptions across episodes
+
+### **Dataset Statistics**
+- **Total Episodes**: 972
+- **Total Frames**: 545k @ 30Hz
+- **Dataset Duration**: 5.04 hours
+
+### **Episode Outcomes**
+| Outcome Category | Episodes | Description |
+|------------------|----------|-------------|
+| **Success** | 780 | Task completed successfully with specified conditions |
+| **Recovery** | 110 | Task finished with recovery behavior |
+| **Failed** | 82 | Task failed to complete |
+
+*Success rates vary across tasks due to varying difficulty levels for operators during teleoperation.
+
+
+---
+
+## Tasks & Domain
+
+### Domain
+
+*Select the primary domain for this dataset.*
+
+- [x] **Surgical Robotics**
+- [ ] **Ultrasound Robotics**
+- [ ] **Other Healthcare Robotics**
+
+### Demonstrated Skills
+
+This specific dataset subset focuses on **Tissue Retraction / Exposure**.
+
+*Note: This is part of the Hamlyn dVRK Data Collection which also includes:*
+- Tissue Retraction / Exposure
+- Knot Tying
+- Suturing (Single Loop)
+- Suturing (Dual Loop)
+- Peg Transfer
+- Needle Grasp and Handover
+
+---
+
+## Data Collection Details
+
+### Collection Method
+
+*How was the data collected?*
+
+- [x] **Human Teleoperation**
+- [ ] **Programmatic/State-Machine**
+- [ ] **AI Policy / Autonomous**
+- [ ] **Other**
+
+
+### Operator Details
+
+| | Description |
+| :--- | :--- |
+| **Operator Count** | 2 |
+| **Operator Skill Level** | `[ ] Expert (e.g., Surgeon, Sonographer)` <br> `[x] Intermediate (e.g., Trained Researcher)` <br> `[x] Novice (e.g., ML Researcher with minimal experience)` <br> `[ ] N/A`|
+| **Collection Period** | From `2025-12-01` to `2026-01-15` |
+| **Input Interface** | `Teleoperation interface with bimanual Force Dimension Sigma 7 haptic device`
+
+### Recovery Demonstrations
+
+*Does this dataset include examples of recovering from failure?*
+
+- [x] **Yes**
+- [ ] **No**
+
+**If yes, please briefly describe the recovery process:**
+
+In the tissue retraction task, failure modes typically involve the forceps slipping off the moist tissue surface or grasping insufficient material to sustain the lift. In these instances, the operator does not abort the episode but performs a **recovery maneuver**: the tool is repositioned, re-oriented to maximize contact area, and a second grasp attempt is made to successfully complete the retraction. These sequences are preserved to aid in learning robust grasping policies.
+
+
+---
+
+## Diversity Dimensions
+
+*Check all dimensions that were intentionally varied during data collection.*
+
+- [ ] **Camera Position / Angle**
+- [ ] **Lighting Conditions**
+- [x] **Target Object** (Varied tissue geometry and composition)
+- [x] **Spatial Layout** (Randomized tissue placement)
+- [ ] **Robot Embodiment**
+- [x] **Task Execution** (Left vs. Right hand dominance / Approach angle)
+- [ ] **Background / Scene**
+- [x] **Initial Robot Configuration** (Ex-Vivo Domain Randomization)
+
+*Elaboration on Diversity:*
+
+*   **Target Object:** We utilized fresh **Porcine Meat** samples exclusively to capture authentic tissue wetness and deformation. Diversity is achieved through:
+    *   **Tissue State:** alternating between **partially dissected tissue flaps** (requiring delamination logic) and **intact tissue blocks** (requiring surface grasping logic).
+    *   **Material Properties:** targeting areas with varying stiffness, such as fibrous muscle tissue versus softer fatty layers.
+*   **Spatial Layout:** The relative pose of the porcine tissue sample was manually randomized within the workspace boundaries between sets to prevent overfitting to absolute coordinates.
+*   **Task Execution:** The task requires adapting the approach vector and grasp strategy based on the direction of the tissue flap (for partially dissected tissue) or the specified retraction grasp location (**left / middle / right**) on the intact block.
+*   **Initial Robot Configuration:** The robot end-effectors were reset to a randomized "home" position (~10cm variance) above the workspace before every trajectory.
+
+
+---
+
+## 🛠️ Equipment & Setup
+
+### Robotic Platform(s)
+
+*List the primary robot(s) used.*
+
+- **Robot 1:** `dVRK (da Vinci Research Kit) - Patient Side Manipulators (PSM1, PSM2)`
+
+
+### Sensors & Cameras
+
+*List the sensors and cameras used. Specify model names where possible.*
+
+| Type | Model/Details |
+| :--- | :--- |
+| **Primary Camera** | `Intel RealSense D405 (RGB + Depth), 848x480 @ 30fps` |
+| **Wrist Camera** | `INSKAM 5.5mm USB Endo-Cam (Right/Left Arm) with custom built mount, view angle 66 degree, focal length 4-10cm, 640x480 @ 24fps` |
+| **Kinematics** | `dVRK High-Frequency Joint Encoders (100Hz)` |
+
+
+
+---
+
+## 🎯 Action & State Space Representation
+
+*The dataset follows the standard LeRobot format for bimanual manipulation.*
+
+### Action Space Representation
+
+**Primary Action Representation:**
+- [x] **Absolute Cartesian** (position/orientation relative to robot base)
+- [ ] **Relative Cartesian**
+- [ ] **Joint Space**
+
+**Orientation Representation:**
+- [x] **Quaternions** (x, y, z, w)
+- [ ] **Euler Angles**
+- [ ] **Rotation Matrix**
+
+**Reference Frame:**
+- [x] **Robot Base Frame** (Base of each PSM arm)
+- [ ] **Camera Frame**
+
+**Action Dimensions:**
+
+action: [x, y, z, qx, qy, qz, qw, gripper_angle, x, y, z, qx, qy, qz, qw, gripper_angle]
+
+- The first 8 dimensions are for the *left* arm and last 8 dimensions are for the *right* arm
+- x, y, z: Absolute position in PSM base frame (meters)
+- qx, qy, qz, qw: Absolute orientation as quaternion
+- gripper_angle: Normalized jaw opening angle (1 = opened, 0 = closed)
+
+
+### State Space Representation
+
+**State Information Included:**
+- [x] **Joint Positions**
+- [x] **Joint Velocities**
+- [ ] **End-Effector Pose** (No extra end-effector used with dVRK)
+- [x] **Gripper State**
+
+**Primary State:**
+
+*observation.state*: [x, y, z, qx, qy, qz, qw, gripper_angle, x, y, z, qx, qy, qz, qw, gripper_angle]
+
+- The first 8 dimensions are for the *left* arm and last 8 dimensions are for the *right* arm
+- x, y, z: Absolute position in PSM base frame (meters)
+- qx, qy, qz, qw: Absolute orientation as quaternion
+- gripper_angle: Normalized jaw opening angle (1 = opened, 0 = closed)
+
+**State Dimensions:**
+
+`observation.state.left_arm_cartesian`: [x, y, z, qx, qy, qz, qw, gripper_angle]
+
+`observation.state.right_arm_cartesian`: [x, y, z, qx, qy, qz, qw, gripper_angle]
+
+`observation.state.left_arm_joint`: [j1, j2, j3, j4, j5, j6, gripper_angle]
+
+`observation.state.right_arm_joint`: [j1, j2, j3, j4, j5, j6, gripper_angle]
+
+- j1-j6: Joint positions for 6-DOF PSM arm (radians)
+- gripper_angle: Normalized jaw opening angle (1 = opened, 0 = closed)
+
+
+---
+
+## ⏱️ Data Synchronization Approach
+
+*Describe how you achieved proper data synchronization.*
+
+**Distributed Synchronization Architecture:**
+The data collection setup involves a distributed system with a robot control machine that teleop the dVRK and a sensor recording machine that stream kinematics, and videos from camers.
+1.  **Clock Sync:** We utilize **Chrony (NTP)** to enforce strict time alignment between the two workstations, minimizing clock skew to < 3.3 ms.
+2.  **Timestamping:** All data streams (camera RGB and depth images, wrist camer images, robot kinematics) are timestamped with **ROS wall-time** headers at the exact moment of capture.
+3.  **Delay Compensattion and Trimming:** The different cameras can introduce small relative delays. These delays are first estimated and compensated by applying a common temporal offset to each stream. Subsequently, all streams are temporally trimmed to their common interval, bounded by the latest start time and the earliest end time across all modalities.
+4.  **Alignment:** During the LeRobot conversion process, frames from all data streams are aligned to a reference timeline at 30 Hz30 Hz. This is achieved via nearest-neighbor temporal interpolation based on the synchronized ROS timestamps.
+
+---
+
+## Attribution & Contact
+
+*Please provide attribution for the dataset creators and a point of contact.*
+
+| | |
+| :--- | :--- |
+| **Dataset Lead** | `Jianhao Su, Kaizhong Deng, Zhaoyang Jacopo Hu` |
+| **Institution** | `The Hamlyn Centre for Robotic Surgery, Imperial College London` |
+| **Contact Email** | `k.deng21@imperial.ac.uk (for technical issues), stamatia.giannarou@imperial.ac.uk (for correspondence)` |
+| **Citation (BibTeX)** | <pre><code>@misc{hamlyn_dvrk_openh_2026,<br>  author = {Su, Jianhao and Deng, Kaizhong and Hu, Zhaoyang Jacopo and Rodriguez y Baena, Ferdinando and Giannarou, Stamatia},<br>  title = {Hamlyn dVRK Surgical Manipulation Dataset for Open-H},<br>  year = {2026},<br>  publisher = {Open-H-Embodiment},<br>  note = {Ex-Vivo Domain: Tissue Retraction / Exposure}<br>}</code></pre> |
+

Surgical/jhu/imerse/wound_closure/point_labeled/fausto_0_1_jesse_0_1_2_labeled/README.md

@@ -0,0 +1,220 @@
+<!--
+Open-H Embodiment Dataset README Template (v1.0)
+Please fill out this template and include it in the ./metadata directory of your LeRobot dataset.
+This file helps others understand the context and details of your contribution.
+-->
+
+# SutureBot Wound Closure - README
+
+---
+
+## 📋 At a Glance
+
+*Provide a one-sentence summary of your dataset.*
+
+**Example:** *Teleoperated demonstrations of a da Vinci robot performing needle passing on a silicone phantom.*
+
+---
+
+## 📖 Dataset Overview
+
+*Briefly describe the purpose and content of this dataset. What key skills or scenarios does it demonstrate?*
+
+**Example:** *This dataset contains 2,500 trajectories of expert surgeons using the dVRK to perform surgical suturing tasks. It includes successful trials, failures, and recovery attempts to provide a robust dataset for training imitation learning policies.*
+
+| | |
+| :--- | :--- |
+| **Total Trajectories** | `[Number]` |
+| **Total Hours** | `12` |
+| **Data Type** | `[ ] Clinical` `[X] Ex-Vivo` `[ ] Table-Top Phantom` `[ ] Digital Simulation` `[ ] Physical Simulation` `[ ] Other (If checked, update "Other")` |
+| **License** | CC BY 4.0 |
+| **Version** | `1.0` |
+
+---
+
+## 🎯 Tasks & Domain
+
+### Domain
+
+*Select the primary domain for this dataset.*
+
+- [X] **Surgical Robotics**
+- [ ] **Ultrasound Robotics**
+- [ ] **Other Healthcare Robotics** (Please specify: `[Your Domain]`)
+
+### Demonstrated Skills
+
+*List the primary skills or procedures demonstrated in this dataset.*
+
+***Example:***
+- Needle-passing
+- Knot tying (2-1-1)
+- Needle Pickup
+
+---
+
+## 🔬 Data Collection Details
+
+### Collection Method
+
+*How was the data collected?*
+
+- [X] **Human Teleoperation**
+- [ ] **Programmatic/State-Machine**
+- [ ] **AI Policy / Autonomous**
+- [ ] **Other** (Please specify: `[Your Method]`)
+
+### Operator Details
+
+| | Description |
+| :--- | :--- |
+| **Operator Count** | `2` |
+| **Operator Skill Level** | `[ ] Expert (e.g., Surgeon, Sonographer)` <br> `[X] Intermediate (e.g., Trained Researcher)` <br> `[ ] Novice (e.g., ML Researcher with minimal experience)` <br> `[ ] N/A` |
+| **Collection Period** | From `[2026-01-09]` to `[2026-02-04]` |
+
+### Recovery Demonstrations
+
+*Does this dataset include examples of recovering from failure?*
+
+- [X] **Yes**
+- [ ] **No**
+
+**If yes, please briefly describe the recovery process:**
+
+*Example: For 250 demonstrations, demonstrations are initialized from a failed needle grasp position, the operator re-orients the robotic grippers and attempts to grasp the needle again from a different angle.*
+Data is collected as full procedure episodes. Is a mistake is make during data collection, the demonstrator immediately recovers from that mistake. 
+
+---
+
+## 💡 Diversity Dimensions
+
+*Check all dimensions that were intentionally varied during data collection.*
+
+- [X] **Camera Position / Angle**
+- [ ] **Lighting Conditions**
+- [X] **Target Object** (e.g., different phantom models, suture types)
+- [X] **Spatial Layout** (e.g., placing the target suture needle in various locations)
+- [ ] **Robot Embodiment** (if multiple robots were used)
+- [ ] **Task Execution** (e.g., different techniques for the same task)
+- [ ] **Background / Scene**
+- [X] **Other** (Please specify: `Suture spacing, bite depth`)
+
+*If you checked any of the above please briefly elaborate below.*
+
+**Example:** We adjusted the room camera perspective every 100 demonstrations. The camera angle was varied by panning up and down by +/- 10 degrees, as well as manually adjusting the height of the camera mount by +/- 2 cm. Additionally, we varied the needle used by swapping out various curvatures, including 1/4, 3/8, 1/2, and 5/8.
+We adjusted the tissue and robot position, orientation, wrist camera placement, bite depth, and suture spacing before every episode. Suture and needle type were occasionally changed throughout. 
+
+---
+
+## 🛠️ Equipment & Setup
+
+### Robotic Platform(s)
+
+*List the primary robot(s) used.*
+
+- **Robot 1:** `dVRK Si`
+
+### Sensors & Cameras
+
+*List the sensors and cameras used. Specify model names where possible. (Add and remove rows as needed)*
+
+| Type | Model/Details |
+| :--- | :--- |
+| **Primary Camera** | `[e.g., Endoscopic Camera, 960x540 @ 30fps]` |
+| **Wrist Cameras** | `[e.g., USB Endoscope Borescope Snake Camera, 640x480 @ 30fps]` |
+
+---
+
+## 🎯 Action & State Space Representation
+
+*Describe how actions and robot states are represented in your dataset. This is crucial for understanding data compatibility and enabling effective policy learning.*
+
+### Action Space Representation
+
+**Primary Action Representation:**
+- [X] **Absolute Cartesian** (position/orientation relative to robot base)
+- [ ] **Relative Cartesian** (delta position/orientation from current pose)
+- [ ] **Joint Space** (direct joint angle commands)
+- [ ] **Other** (Please specify: `[Your Representation]`)
+
+**Orientation Representation:**
+- [X] **Quaternions** (x, y, z, w)
+- [ ] **Euler Angles** (roll, pitch, yaw)
+- [ ] **Axis-Angle** (rotation vector)
+- [ ] **Rotation Matrix** (3x3 matrix)
+- [ ] **Other** (Please specify: `[Your Representation]`)
+
+**Reference Frame:**
+- [X] **Robot Base Frame**
+- [ ] **Tool/End-Effector Frame**
+- [ ] **World/Global Frame**
+- [ ] **Camera Frame**
+- [ ] **Other** (Please specify: `[Your Frame]`)
+
+**Action Dimensions:**
+*List the action space dimensions and their meanings.*
+
+```
+action: [x, y, z, qx, qy, qz, qw, gripper]
+- x, y, z: Absolute position in robot base frame (meters)
+- qx, qy, qz, qw: Absolute orientation as quaternion
+- gripper: Gripper opening angle (radians)
+```
+
+### State Space Representation
+
+**State Information Included:**
+- [X] **Joint Positions** (all articulated joints)
+- [ ] **Joint Velocities**
+- [X] **End-Effector Pose** (Cartesian position/orientation)
+- [ ] **Force/Torque Readings**
+- [X] **Gripper State** (position, force, etc.)
+- [ ] **Other** (Please specify: `[Your State Info]`)
+
+**State Dimensions:**
+*List the state space dimensions and their meanings.*
+
+**Example:**
+```
+observation.state: [j1, j2, j3, j4, j5, j6, j7, gripper_pos]
+- j1-j7: Absolute joint positions for 7-DOF arm (radians)
+- gripper_pos: Current gripper opening (meters)
+```
+
+### 📋 Recommended Additional Representations
+
+*Even if not your primary action/state representation, we strongly encourage including these standardized formats for maximum compatibility:*
+
+**Recommended Action Fields:**
+- **`action.cartesian_absolute`**: Absolute Cartesian pose with absolute quaternions
+  ```
+  [x, y, z, qx, qy, qz, qw, gripper_angle]
+  ```
+
+**Recommended State Fields:**
+- **`observation.state.joint_positions`**: Absolute positions for all articulated joints
+  ```
+  [joint_1, joint_2, ..., joint_n]
+  ```
+
+
+---
+
+## ⏱️ Data Synchronization Approach
+
+*Describe how you achieved proper data synchronization across different sensors, cameras, and robotic systems during data collection. This is crucial for ensuring temporal alignment of all modalities in your dataset.*
+
+**Example:** *We collect joint kinematics from our Franka Research 3 and RGB-D frames from Intel RealSense D435 cameras, all running in ROS 2 Galactic on the same workstation clocked with ROS Time. Both drivers stamp their outgoing messages’ header.stamp fields with the shared system clock, and we record /joint_states, /camera/*/image_raw, and /camera/*/camera_info in a single rosbag2 session. During export to LeRobot, each data point’s ROS header.stamp is written verbatim into the timestamp attribute. Offline checks show inter-sensor skew stays below ±2 ms across a 2-minute capture.*
+
+---
+
+## 👥 Attribution & Contact
+
+*Please provide attribution for the dataset creators and a point of contact.*
+
+| | |
+| :--- | :--- |
+| **Dataset Lead** | `[Name1, Name2, ...]` |
+| **Institution** | `[Your Institution]` |
+| **Contact Email** | `[email1@example.com, email2@example.com, ...]` |
+| **Citation (BibTeX)** | <pre><code>@misc{[your_dataset_name_2025],<br>  author = {[Your Name(s)]},<br>  title = {[Your Dataset Title]},<br>  year = {2025},<br>  publisher = {Open-H-Embodiment},<br>  note = {https://hrpp.research.virginia.edu/teams/irb-sbs/researcher-guide-irb-sbs/identifiers}<br>}</code></pre> |

Surgical/sanoscience/sanoscience_v1_2_merged/expert_demonstrations/README.md

@@ -0,0 +1,259 @@
+# SANO_VR_Cholecystectomy - README
+
+---
+
+## 📋 At a Glance
+
+*Provide a one-sentence summary of your dataset.*
+
+Cholecystectomy procedure performed in a Virtual Reality simulator.
+
+---
+
+## 📖 Dataset Overview
+
+*Briefly describe the purpose and content of this dataset. What key skills or scenarios does it demonstrate?*
+
+This dataset contains recordings of both experts and non-experts performing robotic cholecystectomy in a Virtual Reality simulated environment. It includes successful attempts as well as failures and recovery attempts.
+The data includes multiple camera angles, different lighting scenarios, different quality levels of simulated anatomy, and each camera view includes additional depth data, segmentation masks, approximate surface normals and approximated optical flow.
+
+Data breakdown:
+15 372 total trajectories, of which:
+- 5 454 are performed by experts
+- 9 918 are performed by non-experts
+
+- 15 120 are successes
+- 126 + 126 are failures + recoveries
+
+Episodes vary in length from 2 seconds to 10 seconds.
+The data contains both per-episode and per-frame labels. Both are generated automatically from simulation data based on current instrument state, including activation state and physics contacts. Per-frame labels are generated directly in this way, while per-episode labels are generated by finding the closest event in the near future. If task can't be determined, "Unknown" is set as the task. As these labels are generated automatically, they may not always accurately represent the intention of the operator.
+
+The observations and actions for dVRK PSM manipulator joints are obtained from nVidia Newton inverse kinematic simulation and may not be precise.
+IMPORTANT - PATCH: In order to use the joints data, the videos and images folders from the v1.1 directory should be copied here.
+
+| | |
+| :--- | :--- |
+| **Total Trajectories** | `[15 372]` |
+| **Total Hours** | `[~0.94h augmented x18, total ~17h]` |
+| **Data Type** | `[ ] Clinical` `[ ] Ex-Vivo` `[ ] Table-Top Phantom` `[X] Digital Simulation` `[ ] Physical Simulation` `[ ] Other` |
+| **License** | CC BY 4.0 |
+| **Version** | `[1.1]` |
+
+---
+
+## 🎯 Tasks & Domain
+
+### Domain
+
+*Select the primary domain for this dataset.*
+
+- [X] **Surgical Robotics**
+- [ ] **Ultrasound Robotics**
+- [ ] **Other Healthcare Robotics** 
+
+### Demonstrated Skills
+
+*List the primary skills or procedures demonstrated in this dataset.*
+
+- Robotic cholecystectomy
+- Tissue grasping
+- Tissue prepping
+- Vessels clipping and cutting
+- Gallbladder dissection
+- Tissue diathermy / electrocautery
+
+---
+
+## 🔬 Data Collection Details
+
+### Collection Method
+
+*How was the data collected?*
+
+- [ ] **Human Teleoperation**
+- [ ] **Programmatic/State-Machine**
+- [ ] **AI Policy / Autonomous**
+- [X] **Other (Virtual Reality Simulator)**
+
+### Operator Details
+
+| | Description |
+| :--- | :--- |
+| **Operator Count** | `[4]` |
+| **Operator Skill Level** | `[X] Expert (e.g., Surgeon, Sonographer)` <br> `[ ] Intermediate (e.g., Trained Researcher)` <br> `[X] Novice (e.g., ML Researcher with minimal experience)` <br> `[ ] N/A` |
+| **Collection Period** | From `[2026-01-01]` to `[2026-01-16]` |
+
+### Recovery Demonstrations
+
+*Does this dataset include examples of recovering from failure?*
+
+- [X] **Yes**
+- [ ] **No**
+
+**If yes, please briefly describe the recovery process:**
+
+126 failure case trajectories and a matching number of recovery trajectories are included. The failures relate to the process of clipping vessels. A failure occurs if the user cuts a vessel or damages it excessively by improperly using diathermy. The recovery process consists of clipping the leaking vessel.
+
+---
+
+## 💡 Diversity Dimensions
+
+*Check all dimensions that were intentionally varied during data collection.*
+
+- [X] **Camera Position / Angle**
+- [X] **Lighting Conditions**
+- [ ] **Target Object** (e.g., different phantom models, suture types)
+- [ ] **Spatial Layout** (e.g., placing the target suture needle in various locations)
+- [ ] **Robot Embodiment** (if multiple robots were used)
+- [ ] **Task Execution** (e.g., different techniques for the same task)
+- [ ] **Background / Scene**
+- [ ] **Other** (Please specify: `[Your Dimension]`)
+
+*If you checked any of the above please briefly elaborate below.*
+
+Each attempt was rendered with 6 virtual cameras, each with 3 lighting scenarios:
+- 1 camera follows the user's view while performing the procedure, with FOV matched to the real FOV of the VR HMD.
+- 1 camera follows the user's view, with FOV set at a fixed values, to better represent the important areas during the procedure.
+- 4 remaining cameras are placed in static spots around the area of interest.
+
+The lighting scenarios include a spot light following the user's view, a spot light positioned at the currently rendering camera, and a fixed point light setup around the area of interest.
+Additionally, instrument starting positions and equipped instruments naturally vary between operators and between runs.
+
+
+---
+
+## 🛠️ Equipment & Setup
+
+### Robotic Platform(s)
+
+N/A (simulated environment)
+dVRK (simulated joints)
+
+### Sensors & Cameras
+
+*List the sensors and cameras used. Specify model names where possible. (Add and remove rows as needed)*
+
+| Type | Model/Details |
+| :--- | :--- |
+| **Virtual Mono Camera** | `[Simulated 3D camera, 1920x1080 @ 25-30fps]` |
+| **Virtual Stereo Camera** | `[Simulated 3D Stereo camera, 2x1080x1080 @ 30fps]` |
+
+
+In addition to shaded color views, additional camera views are provided:
+- Depth (reversed, 1 = closest, 0 = furthest)
+- View-space normals (scaled from [-1, 1] range into [0, 1] on each axis for video encoding purposes)
+- Segmentation mask
+- Optical flow generated from motion vectors (format based on https://www.mdpi.com/1424-8220/10/4/2975, hue represents angle while value represents the motion strength; value is in UV-space (UV coordinates, not pixel offsets) scaled 10x to reduce the impact of compression artifacts)
+
+Most trajectories (13 770) are provided using the mono camera. 
+1 602 trajectories are provided in stereo format, lenses are spaced 8 mm apart. The two camera views are provided side-by-side in one video and are not distorted. Due to technical constraints, stereo trajectories do not include optical flow views.
+
+---
+
+## 🎯 Action & State Space Representation
+
+*Describe how actions and robot states are represented in your dataset. This is crucial for understanding data compatibility and enabling effective policy learning.*
+
+### Action Space Representation
+
+**Primary Action Representation:**
+- [X] **Absolute Cartesian** (position/orientation relative to robot base)
+- [ ] **Relative Cartesian** (delta position/orientation from current pose)
+- [X] **Joint Space** (direct joint angle commands)
+- [ ] **Other** (Please specify: `[Your Representation]`)
+
+**Orientation Representation:**
+- [X] **Quaternions** (x, y, z, w)
+- [ ] **Euler Angles** (roll, pitch, yaw)
+- [ ] **Axis-Angle** (rotation vector)
+- [ ] **Rotation Matrix** (3x3 matrix)
+- [ ] **Other** (Please specify: `[Your Representation]`)
+
+**Reference Frame:**
+- [ ] **Robot Base Frame**
+- [ ] **Tool/End-Effector Frame**
+- [X] **World/Global Frame**
+- [X] **Camera Frame**
+- [ ] **Other** (Please specify: `[Your Frame]`)
+
+Both world and camera frames are provided as separate fields.
+
+**Action Dimensions:**
+*List the action space dimensions and their meanings.*
+
+**Example:**
+```
+action: [tip_pos_x, tip_pos_y, tip_pos_z, tip_rot_x, tip_rot_y, tip_rot_z, tip_rot_w,
+tip_cam_x, tip_cam_y, tip_cam_z, tip_cam_x, tip_cam_y, tip_cam_z, tip_cam_w,
+handle, gripper]
+- tip_pos_x, tip_pos_y, tip_pos_z: Absolute position of the end effector tip in world frame (meters, approximate)
+- tip_rot_x, tip_rot_y, tip_rot_z, tip_rot_w: Absolute rotation (quaternion) of the end effector tip in world frame
+- tip_cam_x, tip_cam_y, tip_cam_z: Relative position of the end effector tip in camera frame (meters, approximate)
+- tip_cam_x, tip_cam_y, tip_cam_z, tip_rot_w: Relative rotation (quaternion) of the end effector tip in camera frame
+- handle: user handle value, controls graspers, scissors, clipper activation. (0-1, with 0 being fully released and 1 being fully pressed)
+- gripper: Gripper opening angle (radians), this is valid only for instruments which have a gripper
+
+action.joint_positions: ["psm_yaw_joint", "psm_pitch_end_joint", "psm_main_insertion_joint", "psm_tool_roll_joint", "psm_tool_pitch_joint", "psm_tool_yaw_joint", "psm_tool_gripper1_joint", "psm_tool_gripper2_joint", "psm_pitch_back_joint", "psm_pitch_bottom_joint", "psm_pitch_top_joint", "psm_pitch_front_joint"]
+
+These actions are repeated 4 times, as there are up to 4 instruments available in the simulated procedure.
+
+```
+
+### State Space Representation
+
+**State Information Included:**
+- [X] **Joint Positions** (all articulated joints)
+- [ ] **Joint Velocities**
+- [X] **End-Effector Pose** (Cartesian position/orientation)
+- [ ] **Force/Torque Readings**
+- [X] **Gripper State** (position, force, etc.)
+- [ ] **Other** (Please specify: `[Your State Info]`)
+
+**State Dimensions:**
+State space is identical to action space with additional joint velocity data
+
+```
+observation.state_joints_pos: ["psm_yaw_joint", "psm_pitch_end_joint", "psm_main_insertion_joint", "psm_tool_roll_joint", "psm_tool_pitch_joint", "psm_tool_yaw_joint", "psm_tool_gripper1_joint", "psm_tool_gripper2_joint", "psm_pitch_back_joint", "psm_pitch_bottom_joint", "psm_pitch_top_joint", "psm_pitch_front_joint"]
+observation.state_joints_vel: ["psm_yaw_joint", "psm_pitch_end_joint", "psm_main_insertion_joint", "psm_tool_roll_joint", "psm_tool_pitch_joint", "psm_tool_yaw_joint", "psm_tool_gripper1_joint", "psm_tool_gripper2_joint", "psm_pitch_back_joint", "psm_pitch_bottom_joint", "psm_pitch_top_joint", "psm_pitch_front_joint"]
+```
+### 📋 Recommended Additional Representations
+
+*Even if not your primary action/state representation, we strongly encourage including these standardized formats for maximum compatibility:*
+
+**Recommended Action Fields:**
+- **`action.cartesian_absolute`**: Absolute Cartesian pose with absolute quaternions
+  ```
+  [x, y, z, qx, qy, qz, qw, gripper_angle]
+  ```
+
+**Available**, these are the same as action values: [tip_pos_x, tip_pos_y, tip_pos_z, tip_rot_x, tip_rot_y, tip_rot_z, tip_rot_w, gripper]
+
+**Recommended State Fields:**
+- **`observation.state.joint_positions`**: Absolute positions for all articulated joints
+  ```
+  [joint_1, joint_2, ..., joint_n]
+  ```
+
+  **Available** as the simulation does not include computed joints.
+
+
+---
+
+## ⏱️ Data Synchronization Approach
+
+*Describe how you achieved proper data synchronization across different sensors, cameras, and robotic systems during data collection. This is crucial for ensuring temporal alignment of all modalities in your dataset.*
+
+All sensors and cameras are inherently perfectly synchronized due to the simulated nature of the dataset. At the start of the frame, instruments are updated from the most recent available input data (from connected controllers). Camera views are rendered later, but the instrument positions will not be updated until the next frame starts processing. Therefore the sensor/action data is guaranteed to be synchronized with the camera data. 
+
+---
+
+## 👥 Attribution & Contact
+
+*Please provide attribution for the dataset creators and a point of contact.*
+
+| | |
+| :--- | :--- |
+| **Dataset Lead** | `Mateusz Wójcikowski, Przemysław Korzeniowski, Sabina Martyniak, Michał Naskręt, Prof Filippo Filicori, Dr Aditya Amit Godbole, Dr Maria Clara Morais, Dr Mattia Ballo` |
+| **Institution** | `Sano - Centre for Computational Medicine` |
+| **Contact Email** | `m.wojcikowski@sanoscience.org, p.korzeniowski@sanoscience.org` |
+| **Citation (BibTeX)** | <pre><code>@misc{SANO_VR_Cholecystectomy_2025,<br>  author = {Mateusz Wójcikowski, Przemysław Korzeniowski, Sabina Martyniak, Michał Naskręt, Prof Filippo Filicori, Dr Aditya Amit Godbole, Dr Maria Clara Morais, Dr Mattia Ballo},<br>  title = {SANO_VR_Cholecystectomy},<br>  year = {2025},<br>  publisher = {Open-H-Embodiment},<br>  note = {Simulated Domain: Robotic Cholecystectomy}<br>}</code></pre> |

Surgical/sanoscience/sanoscience_v1_2_merged/nonexpert_failure/README.md

@@ -0,0 +1,259 @@
+# SANO_VR_Cholecystectomy - README
+
+---
+
+## 📋 At a Glance
+
+*Provide a one-sentence summary of your dataset.*
+
+Cholecystectomy procedure performed in a Virtual Reality simulator.
+
+---
+
+## 📖 Dataset Overview
+
+*Briefly describe the purpose and content of this dataset. What key skills or scenarios does it demonstrate?*
+
+This dataset contains recordings of both experts and non-experts performing robotic cholecystectomy in a Virtual Reality simulated environment. It includes successful attempts as well as failures and recovery attempts.
+The data includes multiple camera angles, different lighting scenarios, different quality levels of simulated anatomy, and each camera view includes additional depth data, segmentation masks, approximate surface normals and approximated optical flow.
+
+Data breakdown:
+15 372 total trajectories, of which:
+- 5 454 are performed by experts
+- 9 918 are performed by non-experts
+
+- 15 120 are successes
+- 126 + 126 are failures + recoveries
+
+Episodes vary in length from 2 seconds to 10 seconds.
+The data contains both per-episode and per-frame labels. Both are generated automatically from simulation data based on current instrument state, including activation state and physics contacts. Per-frame labels are generated directly in this way, while per-episode labels are generated by finding the closest event in the near future. If task can't be determined, "Unknown" is set as the task. As these labels are generated automatically, they may not always accurately represent the intention of the operator.
+
+The observations and actions for dVRK PSM manipulator joints are obtained from nVidia Newton inverse kinematic simulation and may not be precise.
+IMPORTANT - PATCH: In order to use the joints data, the videos and images folders from the v1.1 directory should be copied here.
+
+| | |
+| :--- | :--- |
+| **Total Trajectories** | `[15 372]` |
+| **Total Hours** | `[~0.94h augmented x18, total ~17h]` |
+| **Data Type** | `[ ] Clinical` `[ ] Ex-Vivo` `[ ] Table-Top Phantom` `[X] Digital Simulation` `[ ] Physical Simulation` `[ ] Other` |
+| **License** | CC BY 4.0 |
+| **Version** | `[1.1]` |
+
+---
+
+## 🎯 Tasks & Domain
+
+### Domain
+
+*Select the primary domain for this dataset.*
+
+- [X] **Surgical Robotics**
+- [ ] **Ultrasound Robotics**
+- [ ] **Other Healthcare Robotics** 
+
+### Demonstrated Skills
+
+*List the primary skills or procedures demonstrated in this dataset.*
+
+- Robotic cholecystectomy
+- Tissue grasping
+- Tissue prepping
+- Vessels clipping and cutting
+- Gallbladder dissection
+- Tissue diathermy / electrocautery
+
+---
+
+## 🔬 Data Collection Details
+
+### Collection Method
+
+*How was the data collected?*
+
+- [ ] **Human Teleoperation**
+- [ ] **Programmatic/State-Machine**
+- [ ] **AI Policy / Autonomous**
+- [X] **Other (Virtual Reality Simulator)**
+
+### Operator Details
+
+| | Description |
+| :--- | :--- |
+| **Operator Count** | `[4]` |
+| **Operator Skill Level** | `[X] Expert (e.g., Surgeon, Sonographer)` <br> `[ ] Intermediate (e.g., Trained Researcher)` <br> `[X] Novice (e.g., ML Researcher with minimal experience)` <br> `[ ] N/A` |
+| **Collection Period** | From `[2026-01-01]` to `[2026-01-16]` |
+
+### Recovery Demonstrations
+
+*Does this dataset include examples of recovering from failure?*
+
+- [X] **Yes**
+- [ ] **No**
+
+**If yes, please briefly describe the recovery process:**
+
+126 failure case trajectories and a matching number of recovery trajectories are included. The failures relate to the process of clipping vessels. A failure occurs if the user cuts a vessel or damages it excessively by improperly using diathermy. The recovery process consists of clipping the leaking vessel.
+
+---
+
+## 💡 Diversity Dimensions
+
+*Check all dimensions that were intentionally varied during data collection.*
+
+- [X] **Camera Position / Angle**
+- [X] **Lighting Conditions**
+- [ ] **Target Object** (e.g., different phantom models, suture types)
+- [ ] **Spatial Layout** (e.g., placing the target suture needle in various locations)
+- [ ] **Robot Embodiment** (if multiple robots were used)
+- [ ] **Task Execution** (e.g., different techniques for the same task)
+- [ ] **Background / Scene**
+- [ ] **Other** (Please specify: `[Your Dimension]`)
+
+*If you checked any of the above please briefly elaborate below.*
+
+Each attempt was rendered with 6 virtual cameras, each with 3 lighting scenarios:
+- 1 camera follows the user's view while performing the procedure, with FOV matched to the real FOV of the VR HMD.
+- 1 camera follows the user's view, with FOV set at a fixed values, to better represent the important areas during the procedure.
+- 4 remaining cameras are placed in static spots around the area of interest.
+
+The lighting scenarios include a spot light following the user's view, a spot light positioned at the currently rendering camera, and a fixed point light setup around the area of interest.
+Additionally, instrument starting positions and equipped instruments naturally vary between operators and between runs.
+
+
+---
+
+## 🛠️ Equipment & Setup
+
+### Robotic Platform(s)
+
+N/A (simulated environment)
+dVRK (simulated joints)
+
+### Sensors & Cameras
+
+*List the sensors and cameras used. Specify model names where possible. (Add and remove rows as needed)*
+
+| Type | Model/Details |
+| :--- | :--- |
+| **Virtual Mono Camera** | `[Simulated 3D camera, 1920x1080 @ 25-30fps]` |
+| **Virtual Stereo Camera** | `[Simulated 3D Stereo camera, 2x1080x1080 @ 30fps]` |
+
+
+In addition to shaded color views, additional camera views are provided:
+- Depth (reversed, 1 = closest, 0 = furthest)
+- View-space normals (scaled from [-1, 1] range into [0, 1] on each axis for video encoding purposes)
+- Segmentation mask
+- Optical flow generated from motion vectors (format based on https://www.mdpi.com/1424-8220/10/4/2975, hue represents angle while value represents the motion strength; value is in UV-space (UV coordinates, not pixel offsets) scaled 10x to reduce the impact of compression artifacts)
+
+Most trajectories (13 770) are provided using the mono camera. 
+1 602 trajectories are provided in stereo format, lenses are spaced 8 mm apart. The two camera views are provided side-by-side in one video and are not distorted. Due to technical constraints, stereo trajectories do not include optical flow views.
+
+---
+
+## 🎯 Action & State Space Representation
+
+*Describe how actions and robot states are represented in your dataset. This is crucial for understanding data compatibility and enabling effective policy learning.*
+
+### Action Space Representation
+
+**Primary Action Representation:**
+- [X] **Absolute Cartesian** (position/orientation relative to robot base)
+- [ ] **Relative Cartesian** (delta position/orientation from current pose)
+- [X] **Joint Space** (direct joint angle commands)
+- [ ] **Other** (Please specify: `[Your Representation]`)
+
+**Orientation Representation:**
+- [X] **Quaternions** (x, y, z, w)
+- [ ] **Euler Angles** (roll, pitch, yaw)
+- [ ] **Axis-Angle** (rotation vector)
+- [ ] **Rotation Matrix** (3x3 matrix)
+- [ ] **Other** (Please specify: `[Your Representation]`)
+
+**Reference Frame:**
+- [ ] **Robot Base Frame**
+- [ ] **Tool/End-Effector Frame**
+- [X] **World/Global Frame**
+- [X] **Camera Frame**
+- [ ] **Other** (Please specify: `[Your Frame]`)
+
+Both world and camera frames are provided as separate fields.
+
+**Action Dimensions:**
+*List the action space dimensions and their meanings.*
+
+**Example:**
+```
+action: [tip_pos_x, tip_pos_y, tip_pos_z, tip_rot_x, tip_rot_y, tip_rot_z, tip_rot_w,
+tip_cam_x, tip_cam_y, tip_cam_z, tip_cam_x, tip_cam_y, tip_cam_z, tip_cam_w,
+handle, gripper]
+- tip_pos_x, tip_pos_y, tip_pos_z: Absolute position of the end effector tip in world frame (meters, approximate)
+- tip_rot_x, tip_rot_y, tip_rot_z, tip_rot_w: Absolute rotation (quaternion) of the end effector tip in world frame
+- tip_cam_x, tip_cam_y, tip_cam_z: Relative position of the end effector tip in camera frame (meters, approximate)
+- tip_cam_x, tip_cam_y, tip_cam_z, tip_rot_w: Relative rotation (quaternion) of the end effector tip in camera frame
+- handle: user handle value, controls graspers, scissors, clipper activation. (0-1, with 0 being fully released and 1 being fully pressed)
+- gripper: Gripper opening angle (radians), this is valid only for instruments which have a gripper
+
+action.joint_positions: ["psm_yaw_joint", "psm_pitch_end_joint", "psm_main_insertion_joint", "psm_tool_roll_joint", "psm_tool_pitch_joint", "psm_tool_yaw_joint", "psm_tool_gripper1_joint", "psm_tool_gripper2_joint", "psm_pitch_back_joint", "psm_pitch_bottom_joint", "psm_pitch_top_joint", "psm_pitch_front_joint"]
+
+These actions are repeated 4 times, as there are up to 4 instruments available in the simulated procedure.
+
+```
+
+### State Space Representation
+
+**State Information Included:**
+- [X] **Joint Positions** (all articulated joints)
+- [ ] **Joint Velocities**
+- [X] **End-Effector Pose** (Cartesian position/orientation)
+- [ ] **Force/Torque Readings**
+- [X] **Gripper State** (position, force, etc.)
+- [ ] **Other** (Please specify: `[Your State Info]`)
+
+**State Dimensions:**
+State space is identical to action space with additional joint velocity data
+
+```
+observation.state_joints_pos: ["psm_yaw_joint", "psm_pitch_end_joint", "psm_main_insertion_joint", "psm_tool_roll_joint", "psm_tool_pitch_joint", "psm_tool_yaw_joint", "psm_tool_gripper1_joint", "psm_tool_gripper2_joint", "psm_pitch_back_joint", "psm_pitch_bottom_joint", "psm_pitch_top_joint", "psm_pitch_front_joint"]
+observation.state_joints_vel: ["psm_yaw_joint", "psm_pitch_end_joint", "psm_main_insertion_joint", "psm_tool_roll_joint", "psm_tool_pitch_joint", "psm_tool_yaw_joint", "psm_tool_gripper1_joint", "psm_tool_gripper2_joint", "psm_pitch_back_joint", "psm_pitch_bottom_joint", "psm_pitch_top_joint", "psm_pitch_front_joint"]
+```
+### 📋 Recommended Additional Representations
+
+*Even if not your primary action/state representation, we strongly encourage including these standardized formats for maximum compatibility:*
+
+**Recommended Action Fields:**
+- **`action.cartesian_absolute`**: Absolute Cartesian pose with absolute quaternions
+  ```
+  [x, y, z, qx, qy, qz, qw, gripper_angle]
+  ```
+
+**Available**, these are the same as action values: [tip_pos_x, tip_pos_y, tip_pos_z, tip_rot_x, tip_rot_y, tip_rot_z, tip_rot_w, gripper]
+
+**Recommended State Fields:**
+- **`observation.state.joint_positions`**: Absolute positions for all articulated joints
+  ```
+  [joint_1, joint_2, ..., joint_n]
+  ```
+
+  **Available** as the simulation does not include computed joints.
+
+
+---
+
+## ⏱️ Data Synchronization Approach
+
+*Describe how you achieved proper data synchronization across different sensors, cameras, and robotic systems during data collection. This is crucial for ensuring temporal alignment of all modalities in your dataset.*
+
+All sensors and cameras are inherently perfectly synchronized due to the simulated nature of the dataset. At the start of the frame, instruments are updated from the most recent available input data (from connected controllers). Camera views are rendered later, but the instrument positions will not be updated until the next frame starts processing. Therefore the sensor/action data is guaranteed to be synchronized with the camera data. 
+
+---
+
+## 👥 Attribution & Contact
+
+*Please provide attribution for the dataset creators and a point of contact.*
+
+| | |
+| :--- | :--- |
+| **Dataset Lead** | `Mateusz Wójcikowski, Przemysław Korzeniowski, Sabina Martyniak, Michał Naskręt, Prof Filippo Filicori, Dr Aditya Amit Godbole, Dr Maria Clara Morais, Dr Mattia Ballo` |
+| **Institution** | `Sano - Centre for Computational Medicine` |
+| **Contact Email** | `m.wojcikowski@sanoscience.org, p.korzeniowski@sanoscience.org` |
+| **Citation (BibTeX)** | <pre><code>@misc{SANO_VR_Cholecystectomy_2025,<br>  author = {Mateusz Wójcikowski, Przemysław Korzeniowski, Sabina Martyniak, Michał Naskręt, Prof Filippo Filicori, Dr Aditya Amit Godbole, Dr Maria Clara Morais, Dr Mattia Ballo},<br>  title = {SANO_VR_Cholecystectomy},<br>  year = {2025},<br>  publisher = {Open-H-Embodiment},<br>  note = {Simulated Domain: Robotic Cholecystectomy}<br>}</code></pre> |

Surgical/sanoscience/sanoscience_v1_2_merged/nonexpert_full_modalities/README.md

@@ -0,0 +1,259 @@
+# SANO_VR_Cholecystectomy - README
+
+---
+
+## 📋 At a Glance
+
+*Provide a one-sentence summary of your dataset.*
+
+Cholecystectomy procedure performed in a Virtual Reality simulator.
+
+---
+
+## 📖 Dataset Overview
+
+*Briefly describe the purpose and content of this dataset. What key skills or scenarios does it demonstrate?*
+
+This dataset contains recordings of both experts and non-experts performing robotic cholecystectomy in a Virtual Reality simulated environment. It includes successful attempts as well as failures and recovery attempts.
+The data includes multiple camera angles, different lighting scenarios, different quality levels of simulated anatomy, and each camera view includes additional depth data, segmentation masks, approximate surface normals and approximated optical flow.
+
+Data breakdown:
+15 372 total trajectories, of which:
+- 5 454 are performed by experts
+- 9 918 are performed by non-experts
+
+- 15 120 are successes
+- 126 + 126 are failures + recoveries
+
+Episodes vary in length from 2 seconds to 10 seconds.
+The data contains both per-episode and per-frame labels. Both are generated automatically from simulation data based on current instrument state, including activation state and physics contacts. Per-frame labels are generated directly in this way, while per-episode labels are generated by finding the closest event in the near future. If task can't be determined, "Unknown" is set as the task. As these labels are generated automatically, they may not always accurately represent the intention of the operator.
+
+The observations and actions for dVRK PSM manipulator joints are obtained from nVidia Newton inverse kinematic simulation and may not be precise.
+IMPORTANT - PATCH: In order to use the joints data, the videos and images folders from the v1.1 directory should be copied here.
+
+| | |
+| :--- | :--- |
+| **Total Trajectories** | `[15 372]` |
+| **Total Hours** | `[~0.94h augmented x18, total ~17h]` |
+| **Data Type** | `[ ] Clinical` `[ ] Ex-Vivo` `[ ] Table-Top Phantom` `[X] Digital Simulation` `[ ] Physical Simulation` `[ ] Other` |
+| **License** | CC BY 4.0 |
+| **Version** | `[1.1]` |
+
+---
+
+## 🎯 Tasks & Domain
+
+### Domain
+
+*Select the primary domain for this dataset.*
+
+- [X] **Surgical Robotics**
+- [ ] **Ultrasound Robotics**
+- [ ] **Other Healthcare Robotics** 
+
+### Demonstrated Skills
+
+*List the primary skills or procedures demonstrated in this dataset.*
+
+- Robotic cholecystectomy
+- Tissue grasping
+- Tissue prepping
+- Vessels clipping and cutting
+- Gallbladder dissection
+- Tissue diathermy / electrocautery
+
+---
+
+## 🔬 Data Collection Details
+
+### Collection Method
+
+*How was the data collected?*
+
+- [ ] **Human Teleoperation**
+- [ ] **Programmatic/State-Machine**
+- [ ] **AI Policy / Autonomous**
+- [X] **Other (Virtual Reality Simulator)**
+
+### Operator Details
+
+| | Description |
+| :--- | :--- |
+| **Operator Count** | `[4]` |
+| **Operator Skill Level** | `[X] Expert (e.g., Surgeon, Sonographer)` <br> `[ ] Intermediate (e.g., Trained Researcher)` <br> `[X] Novice (e.g., ML Researcher with minimal experience)` <br> `[ ] N/A` |
+| **Collection Period** | From `[2026-01-01]` to `[2026-01-16]` |
+
+### Recovery Demonstrations
+
+*Does this dataset include examples of recovering from failure?*
+
+- [X] **Yes**
+- [ ] **No**
+
+**If yes, please briefly describe the recovery process:**
+
+126 failure case trajectories and a matching number of recovery trajectories are included. The failures relate to the process of clipping vessels. A failure occurs if the user cuts a vessel or damages it excessively by improperly using diathermy. The recovery process consists of clipping the leaking vessel.
+
+---
+
+## 💡 Diversity Dimensions
+
+*Check all dimensions that were intentionally varied during data collection.*
+
+- [X] **Camera Position / Angle**
+- [X] **Lighting Conditions**
+- [ ] **Target Object** (e.g., different phantom models, suture types)
+- [ ] **Spatial Layout** (e.g., placing the target suture needle in various locations)
+- [ ] **Robot Embodiment** (if multiple robots were used)
+- [ ] **Task Execution** (e.g., different techniques for the same task)
+- [ ] **Background / Scene**
+- [ ] **Other** (Please specify: `[Your Dimension]`)
+
+*If you checked any of the above please briefly elaborate below.*
+
+Each attempt was rendered with 6 virtual cameras, each with 3 lighting scenarios:
+- 1 camera follows the user's view while performing the procedure, with FOV matched to the real FOV of the VR HMD.
+- 1 camera follows the user's view, with FOV set at a fixed values, to better represent the important areas during the procedure.
+- 4 remaining cameras are placed in static spots around the area of interest.
+
+The lighting scenarios include a spot light following the user's view, a spot light positioned at the currently rendering camera, and a fixed point light setup around the area of interest.
+Additionally, instrument starting positions and equipped instruments naturally vary between operators and between runs.
+
+
+---
+
+## 🛠️ Equipment & Setup
+
+### Robotic Platform(s)
+
+N/A (simulated environment)
+dVRK (simulated joints)
+
+### Sensors & Cameras
+
+*List the sensors and cameras used. Specify model names where possible. (Add and remove rows as needed)*
+
+| Type | Model/Details |
+| :--- | :--- |
+| **Virtual Mono Camera** | `[Simulated 3D camera, 1920x1080 @ 25-30fps]` |
+| **Virtual Stereo Camera** | `[Simulated 3D Stereo camera, 2x1080x1080 @ 30fps]` |
+
+
+In addition to shaded color views, additional camera views are provided:
+- Depth (reversed, 1 = closest, 0 = furthest)
+- View-space normals (scaled from [-1, 1] range into [0, 1] on each axis for video encoding purposes)
+- Segmentation mask
+- Optical flow generated from motion vectors (format based on https://www.mdpi.com/1424-8220/10/4/2975, hue represents angle while value represents the motion strength; value is in UV-space (UV coordinates, not pixel offsets) scaled 10x to reduce the impact of compression artifacts)
+
+Most trajectories (13 770) are provided using the mono camera. 
+1 602 trajectories are provided in stereo format, lenses are spaced 8 mm apart. The two camera views are provided side-by-side in one video and are not distorted. Due to technical constraints, stereo trajectories do not include optical flow views.
+
+---
+
+## 🎯 Action & State Space Representation
+
+*Describe how actions and robot states are represented in your dataset. This is crucial for understanding data compatibility and enabling effective policy learning.*
+
+### Action Space Representation
+
+**Primary Action Representation:**
+- [X] **Absolute Cartesian** (position/orientation relative to robot base)
+- [ ] **Relative Cartesian** (delta position/orientation from current pose)
+- [X] **Joint Space** (direct joint angle commands)
+- [ ] **Other** (Please specify: `[Your Representation]`)
+
+**Orientation Representation:**
+- [X] **Quaternions** (x, y, z, w)
+- [ ] **Euler Angles** (roll, pitch, yaw)
+- [ ] **Axis-Angle** (rotation vector)
+- [ ] **Rotation Matrix** (3x3 matrix)
+- [ ] **Other** (Please specify: `[Your Representation]`)
+
+**Reference Frame:**
+- [ ] **Robot Base Frame**
+- [ ] **Tool/End-Effector Frame**
+- [X] **World/Global Frame**
+- [X] **Camera Frame**
+- [ ] **Other** (Please specify: `[Your Frame]`)
+
+Both world and camera frames are provided as separate fields.
+
+**Action Dimensions:**
+*List the action space dimensions and their meanings.*
+
+**Example:**
+```
+action: [tip_pos_x, tip_pos_y, tip_pos_z, tip_rot_x, tip_rot_y, tip_rot_z, tip_rot_w,
+tip_cam_x, tip_cam_y, tip_cam_z, tip_cam_x, tip_cam_y, tip_cam_z, tip_cam_w,
+handle, gripper]
+- tip_pos_x, tip_pos_y, tip_pos_z: Absolute position of the end effector tip in world frame (meters, approximate)
+- tip_rot_x, tip_rot_y, tip_rot_z, tip_rot_w: Absolute rotation (quaternion) of the end effector tip in world frame
+- tip_cam_x, tip_cam_y, tip_cam_z: Relative position of the end effector tip in camera frame (meters, approximate)
+- tip_cam_x, tip_cam_y, tip_cam_z, tip_rot_w: Relative rotation (quaternion) of the end effector tip in camera frame
+- handle: user handle value, controls graspers, scissors, clipper activation. (0-1, with 0 being fully released and 1 being fully pressed)
+- gripper: Gripper opening angle (radians), this is valid only for instruments which have a gripper
+
+action.joint_positions: ["psm_yaw_joint", "psm_pitch_end_joint", "psm_main_insertion_joint", "psm_tool_roll_joint", "psm_tool_pitch_joint", "psm_tool_yaw_joint", "psm_tool_gripper1_joint", "psm_tool_gripper2_joint", "psm_pitch_back_joint", "psm_pitch_bottom_joint", "psm_pitch_top_joint", "psm_pitch_front_joint"]
+
+These actions are repeated 4 times, as there are up to 4 instruments available in the simulated procedure.
+
+```
+
+### State Space Representation
+
+**State Information Included:**
+- [X] **Joint Positions** (all articulated joints)
+- [ ] **Joint Velocities**
+- [X] **End-Effector Pose** (Cartesian position/orientation)
+- [ ] **Force/Torque Readings**
+- [X] **Gripper State** (position, force, etc.)
+- [ ] **Other** (Please specify: `[Your State Info]`)
+
+**State Dimensions:**
+State space is identical to action space with additional joint velocity data
+
+```
+observation.state_joints_pos: ["psm_yaw_joint", "psm_pitch_end_joint", "psm_main_insertion_joint", "psm_tool_roll_joint", "psm_tool_pitch_joint", "psm_tool_yaw_joint", "psm_tool_gripper1_joint", "psm_tool_gripper2_joint", "psm_pitch_back_joint", "psm_pitch_bottom_joint", "psm_pitch_top_joint", "psm_pitch_front_joint"]
+observation.state_joints_vel: ["psm_yaw_joint", "psm_pitch_end_joint", "psm_main_insertion_joint", "psm_tool_roll_joint", "psm_tool_pitch_joint", "psm_tool_yaw_joint", "psm_tool_gripper1_joint", "psm_tool_gripper2_joint", "psm_pitch_back_joint", "psm_pitch_bottom_joint", "psm_pitch_top_joint", "psm_pitch_front_joint"]
+```
+### 📋 Recommended Additional Representations
+
+*Even if not your primary action/state representation, we strongly encourage including these standardized formats for maximum compatibility:*
+
+**Recommended Action Fields:**
+- **`action.cartesian_absolute`**: Absolute Cartesian pose with absolute quaternions
+  ```
+  [x, y, z, qx, qy, qz, qw, gripper_angle]
+  ```
+
+**Available**, these are the same as action values: [tip_pos_x, tip_pos_y, tip_pos_z, tip_rot_x, tip_rot_y, tip_rot_z, tip_rot_w, gripper]
+
+**Recommended State Fields:**
+- **`observation.state.joint_positions`**: Absolute positions for all articulated joints
+  ```
+  [joint_1, joint_2, ..., joint_n]
+  ```
+
+  **Available** as the simulation does not include computed joints.
+
+
+---
+
+## ⏱️ Data Synchronization Approach
+
+*Describe how you achieved proper data synchronization across different sensors, cameras, and robotic systems during data collection. This is crucial for ensuring temporal alignment of all modalities in your dataset.*
+
+All sensors and cameras are inherently perfectly synchronized due to the simulated nature of the dataset. At the start of the frame, instruments are updated from the most recent available input data (from connected controllers). Camera views are rendered later, but the instrument positions will not be updated until the next frame starts processing. Therefore the sensor/action data is guaranteed to be synchronized with the camera data. 
+
+---
+
+## 👥 Attribution & Contact
+
+*Please provide attribution for the dataset creators and a point of contact.*
+
+| | |
+| :--- | :--- |
+| **Dataset Lead** | `Mateusz Wójcikowski, Przemysław Korzeniowski, Sabina Martyniak, Michał Naskręt, Prof Filippo Filicori, Dr Aditya Amit Godbole, Dr Maria Clara Morais, Dr Mattia Ballo` |
+| **Institution** | `Sano - Centre for Computational Medicine` |
+| **Contact Email** | `m.wojcikowski@sanoscience.org, p.korzeniowski@sanoscience.org` |
+| **Citation (BibTeX)** | <pre><code>@misc{SANO_VR_Cholecystectomy_2025,<br>  author = {Mateusz Wójcikowski, Przemysław Korzeniowski, Sabina Martyniak, Michał Naskręt, Prof Filippo Filicori, Dr Aditya Amit Godbole, Dr Maria Clara Morais, Dr Mattia Ballo},<br>  title = {SANO_VR_Cholecystectomy},<br>  year = {2025},<br>  publisher = {Open-H-Embodiment},<br>  note = {Simulated Domain: Robotic Cholecystectomy}<br>}</code></pre> |

Surgical/sanoscience/sanoscience_v1_2_merged/nonexpert_partial_modalities/README.md

@@ -0,0 +1,259 @@
+# SANO_VR_Cholecystectomy - README
+
+---
+
+## 📋 At a Glance
+
+*Provide a one-sentence summary of your dataset.*
+
+Cholecystectomy procedure performed in a Virtual Reality simulator.
+
+---
+
+## 📖 Dataset Overview
+
+*Briefly describe the purpose and content of this dataset. What key skills or scenarios does it demonstrate?*
+
+This dataset contains recordings of both experts and non-experts performing robotic cholecystectomy in a Virtual Reality simulated environment. It includes successful attempts as well as failures and recovery attempts.
+The data includes multiple camera angles, different lighting scenarios, different quality levels of simulated anatomy, and each camera view includes additional depth data, segmentation masks, approximate surface normals and approximated optical flow.
+
+Data breakdown:
+15 372 total trajectories, of which:
+- 5 454 are performed by experts
+- 9 918 are performed by non-experts
+
+- 15 120 are successes
+- 126 + 126 are failures + recoveries
+
+Episodes vary in length from 2 seconds to 10 seconds.
+The data contains both per-episode and per-frame labels. Both are generated automatically from simulation data based on current instrument state, including activation state and physics contacts. Per-frame labels are generated directly in this way, while per-episode labels are generated by finding the closest event in the near future. If task can't be determined, "Unknown" is set as the task. As these labels are generated automatically, they may not always accurately represent the intention of the operator.
+
+The observations and actions for dVRK PSM manipulator joints are obtained from nVidia Newton inverse kinematic simulation and may not be precise.
+IMPORTANT - PATCH: In order to use the joints data, the videos and images folders from the v1.1 directory should be copied here.
+
+| | |
+| :--- | :--- |
+| **Total Trajectories** | `[15 372]` |
+| **Total Hours** | `[~0.94h augmented x18, total ~17h]` |
+| **Data Type** | `[ ] Clinical` `[ ] Ex-Vivo` `[ ] Table-Top Phantom` `[X] Digital Simulation` `[ ] Physical Simulation` `[ ] Other` |
+| **License** | CC BY 4.0 |
+| **Version** | `[1.1]` |
+
+---
+
+## 🎯 Tasks & Domain
+
+### Domain
+
+*Select the primary domain for this dataset.*
+
+- [X] **Surgical Robotics**
+- [ ] **Ultrasound Robotics**
+- [ ] **Other Healthcare Robotics** 
+
+### Demonstrated Skills
+
+*List the primary skills or procedures demonstrated in this dataset.*
+
+- Robotic cholecystectomy
+- Tissue grasping
+- Tissue prepping
+- Vessels clipping and cutting
+- Gallbladder dissection
+- Tissue diathermy / electrocautery
+
+---
+
+## 🔬 Data Collection Details
+
+### Collection Method
+
+*How was the data collected?*
+
+- [ ] **Human Teleoperation**
+- [ ] **Programmatic/State-Machine**
+- [ ] **AI Policy / Autonomous**
+- [X] **Other (Virtual Reality Simulator)**
+
+### Operator Details
+
+| | Description |
+| :--- | :--- |
+| **Operator Count** | `[4]` |
+| **Operator Skill Level** | `[X] Expert (e.g., Surgeon, Sonographer)` <br> `[ ] Intermediate (e.g., Trained Researcher)` <br> `[X] Novice (e.g., ML Researcher with minimal experience)` <br> `[ ] N/A` |
+| **Collection Period** | From `[2026-01-01]` to `[2026-01-16]` |
+
+### Recovery Demonstrations
+
+*Does this dataset include examples of recovering from failure?*
+
+- [X] **Yes**
+- [ ] **No**
+
+**If yes, please briefly describe the recovery process:**
+
+126 failure case trajectories and a matching number of recovery trajectories are included. The failures relate to the process of clipping vessels. A failure occurs if the user cuts a vessel or damages it excessively by improperly using diathermy. The recovery process consists of clipping the leaking vessel.
+
+---
+
+## 💡 Diversity Dimensions
+
+*Check all dimensions that were intentionally varied during data collection.*
+
+- [X] **Camera Position / Angle**
+- [X] **Lighting Conditions**
+- [ ] **Target Object** (e.g., different phantom models, suture types)
+- [ ] **Spatial Layout** (e.g., placing the target suture needle in various locations)
+- [ ] **Robot Embodiment** (if multiple robots were used)
+- [ ] **Task Execution** (e.g., different techniques for the same task)
+- [ ] **Background / Scene**
+- [ ] **Other** (Please specify: `[Your Dimension]`)
+
+*If you checked any of the above please briefly elaborate below.*
+
+Each attempt was rendered with 6 virtual cameras, each with 3 lighting scenarios:
+- 1 camera follows the user's view while performing the procedure, with FOV matched to the real FOV of the VR HMD.
+- 1 camera follows the user's view, with FOV set at a fixed values, to better represent the important areas during the procedure.
+- 4 remaining cameras are placed in static spots around the area of interest.
+
+The lighting scenarios include a spot light following the user's view, a spot light positioned at the currently rendering camera, and a fixed point light setup around the area of interest.
+Additionally, instrument starting positions and equipped instruments naturally vary between operators and between runs.
+
+
+---
+
+## 🛠️ Equipment & Setup
+
+### Robotic Platform(s)
+
+N/A (simulated environment)
+dVRK (simulated joints)
+
+### Sensors & Cameras
+
+*List the sensors and cameras used. Specify model names where possible. (Add and remove rows as needed)*
+
+| Type | Model/Details |
+| :--- | :--- |
+| **Virtual Mono Camera** | `[Simulated 3D camera, 1920x1080 @ 25-30fps]` |
+| **Virtual Stereo Camera** | `[Simulated 3D Stereo camera, 2x1080x1080 @ 30fps]` |
+
+
+In addition to shaded color views, additional camera views are provided:
+- Depth (reversed, 1 = closest, 0 = furthest)
+- View-space normals (scaled from [-1, 1] range into [0, 1] on each axis for video encoding purposes)
+- Segmentation mask
+- Optical flow generated from motion vectors (format based on https://www.mdpi.com/1424-8220/10/4/2975, hue represents angle while value represents the motion strength; value is in UV-space (UV coordinates, not pixel offsets) scaled 10x to reduce the impact of compression artifacts)
+
+Most trajectories (13 770) are provided using the mono camera. 
+1 602 trajectories are provided in stereo format, lenses are spaced 8 mm apart. The two camera views are provided side-by-side in one video and are not distorted. Due to technical constraints, stereo trajectories do not include optical flow views.
+
+---
+
+## 🎯 Action & State Space Representation
+
+*Describe how actions and robot states are represented in your dataset. This is crucial for understanding data compatibility and enabling effective policy learning.*
+
+### Action Space Representation
+
+**Primary Action Representation:**
+- [X] **Absolute Cartesian** (position/orientation relative to robot base)
+- [ ] **Relative Cartesian** (delta position/orientation from current pose)
+- [X] **Joint Space** (direct joint angle commands)
+- [ ] **Other** (Please specify: `[Your Representation]`)
+
+**Orientation Representation:**
+- [X] **Quaternions** (x, y, z, w)
+- [ ] **Euler Angles** (roll, pitch, yaw)
+- [ ] **Axis-Angle** (rotation vector)
+- [ ] **Rotation Matrix** (3x3 matrix)
+- [ ] **Other** (Please specify: `[Your Representation]`)
+
+**Reference Frame:**
+- [ ] **Robot Base Frame**
+- [ ] **Tool/End-Effector Frame**
+- [X] **World/Global Frame**
+- [X] **Camera Frame**
+- [ ] **Other** (Please specify: `[Your Frame]`)
+
+Both world and camera frames are provided as separate fields.
+
+**Action Dimensions:**
+*List the action space dimensions and their meanings.*
+
+**Example:**
+```
+action: [tip_pos_x, tip_pos_y, tip_pos_z, tip_rot_x, tip_rot_y, tip_rot_z, tip_rot_w,
+tip_cam_x, tip_cam_y, tip_cam_z, tip_cam_x, tip_cam_y, tip_cam_z, tip_cam_w,
+handle, gripper]
+- tip_pos_x, tip_pos_y, tip_pos_z: Absolute position of the end effector tip in world frame (meters, approximate)
+- tip_rot_x, tip_rot_y, tip_rot_z, tip_rot_w: Absolute rotation (quaternion) of the end effector tip in world frame
+- tip_cam_x, tip_cam_y, tip_cam_z: Relative position of the end effector tip in camera frame (meters, approximate)
+- tip_cam_x, tip_cam_y, tip_cam_z, tip_rot_w: Relative rotation (quaternion) of the end effector tip in camera frame
+- handle: user handle value, controls graspers, scissors, clipper activation. (0-1, with 0 being fully released and 1 being fully pressed)
+- gripper: Gripper opening angle (radians), this is valid only for instruments which have a gripper
+
+action.joint_positions: ["psm_yaw_joint", "psm_pitch_end_joint", "psm_main_insertion_joint", "psm_tool_roll_joint", "psm_tool_pitch_joint", "psm_tool_yaw_joint", "psm_tool_gripper1_joint", "psm_tool_gripper2_joint", "psm_pitch_back_joint", "psm_pitch_bottom_joint", "psm_pitch_top_joint", "psm_pitch_front_joint"]
+
+These actions are repeated 4 times, as there are up to 4 instruments available in the simulated procedure.
+
+```
+
+### State Space Representation
+
+**State Information Included:**
+- [X] **Joint Positions** (all articulated joints)
+- [ ] **Joint Velocities**
+- [X] **End-Effector Pose** (Cartesian position/orientation)
+- [ ] **Force/Torque Readings**
+- [X] **Gripper State** (position, force, etc.)
+- [ ] **Other** (Please specify: `[Your State Info]`)
+
+**State Dimensions:**
+State space is identical to action space with additional joint velocity data
+
+```
+observation.state_joints_pos: ["psm_yaw_joint", "psm_pitch_end_joint", "psm_main_insertion_joint", "psm_tool_roll_joint", "psm_tool_pitch_joint", "psm_tool_yaw_joint", "psm_tool_gripper1_joint", "psm_tool_gripper2_joint", "psm_pitch_back_joint", "psm_pitch_bottom_joint", "psm_pitch_top_joint", "psm_pitch_front_joint"]
+observation.state_joints_vel: ["psm_yaw_joint", "psm_pitch_end_joint", "psm_main_insertion_joint", "psm_tool_roll_joint", "psm_tool_pitch_joint", "psm_tool_yaw_joint", "psm_tool_gripper1_joint", "psm_tool_gripper2_joint", "psm_pitch_back_joint", "psm_pitch_bottom_joint", "psm_pitch_top_joint", "psm_pitch_front_joint"]
+```
+### 📋 Recommended Additional Representations
+
+*Even if not your primary action/state representation, we strongly encourage including these standardized formats for maximum compatibility:*
+
+**Recommended Action Fields:**
+- **`action.cartesian_absolute`**: Absolute Cartesian pose with absolute quaternions
+  ```
+  [x, y, z, qx, qy, qz, qw, gripper_angle]
+  ```
+
+**Available**, these are the same as action values: [tip_pos_x, tip_pos_y, tip_pos_z, tip_rot_x, tip_rot_y, tip_rot_z, tip_rot_w, gripper]
+
+**Recommended State Fields:**
+- **`observation.state.joint_positions`**: Absolute positions for all articulated joints
+  ```
+  [joint_1, joint_2, ..., joint_n]
+  ```
+
+  **Available** as the simulation does not include computed joints.
+
+
+---
+
+## ⏱️ Data Synchronization Approach
+
+*Describe how you achieved proper data synchronization across different sensors, cameras, and robotic systems during data collection. This is crucial for ensuring temporal alignment of all modalities in your dataset.*
+
+All sensors and cameras are inherently perfectly synchronized due to the simulated nature of the dataset. At the start of the frame, instruments are updated from the most recent available input data (from connected controllers). Camera views are rendered later, but the instrument positions will not be updated until the next frame starts processing. Therefore the sensor/action data is guaranteed to be synchronized with the camera data. 
+
+---
+
+## 👥 Attribution & Contact
+
+*Please provide attribution for the dataset creators and a point of contact.*
+
+| | |
+| :--- | :--- |
+| **Dataset Lead** | `Mateusz Wójcikowski, Przemysław Korzeniowski, Sabina Martyniak, Michał Naskręt, Prof Filippo Filicori, Dr Aditya Amit Godbole, Dr Maria Clara Morais, Dr Mattia Ballo` |
+| **Institution** | `Sano - Centre for Computational Medicine` |
+| **Contact Email** | `m.wojcikowski@sanoscience.org, p.korzeniowski@sanoscience.org` |
+| **Citation (BibTeX)** | <pre><code>@misc{SANO_VR_Cholecystectomy_2025,<br>  author = {Mateusz Wójcikowski, Przemysław Korzeniowski, Sabina Martyniak, Michał Naskręt, Prof Filippo Filicori, Dr Aditya Amit Godbole, Dr Maria Clara Morais, Dr Mattia Ballo},<br>  title = {SANO_VR_Cholecystectomy},<br>  year = {2025},<br>  publisher = {Open-H-Embodiment},<br>  note = {Simulated Domain: Robotic Cholecystectomy}<br>}</code></pre> |

Surgical/sanoscience/sanoscience_v1_2_merged/nonexpert_recovery/README.md

@@ -0,0 +1,259 @@
+# SANO_VR_Cholecystectomy - README
+
+---
+
+## 📋 At a Glance
+
+*Provide a one-sentence summary of your dataset.*
+
+Cholecystectomy procedure performed in a Virtual Reality simulator.
+
+---
+
+## 📖 Dataset Overview
+
+*Briefly describe the purpose and content of this dataset. What key skills or scenarios does it demonstrate?*
+
+This dataset contains recordings of both experts and non-experts performing robotic cholecystectomy in a Virtual Reality simulated environment. It includes successful attempts as well as failures and recovery attempts.
+The data includes multiple camera angles, different lighting scenarios, different quality levels of simulated anatomy, and each camera view includes additional depth data, segmentation masks, approximate surface normals and approximated optical flow.
+
+Data breakdown:
+15 372 total trajectories, of which:
+- 5 454 are performed by experts
+- 9 918 are performed by non-experts
+
+- 15 120 are successes
+- 126 + 126 are failures + recoveries
+
+Episodes vary in length from 2 seconds to 10 seconds.
+The data contains both per-episode and per-frame labels. Both are generated automatically from simulation data based on current instrument state, including activation state and physics contacts. Per-frame labels are generated directly in this way, while per-episode labels are generated by finding the closest event in the near future. If task can't be determined, "Unknown" is set as the task. As these labels are generated automatically, they may not always accurately represent the intention of the operator.
+
+The observations and actions for dVRK PSM manipulator joints are obtained from nVidia Newton inverse kinematic simulation and may not be precise.
+IMPORTANT - PATCH: In order to use the joints data, the videos and images folders from the v1.1 directory should be copied here.
+
+| | |
+| :--- | :--- |
+| **Total Trajectories** | `[15 372]` |
+| **Total Hours** | `[~0.94h augmented x18, total ~17h]` |
+| **Data Type** | `[ ] Clinical` `[ ] Ex-Vivo` `[ ] Table-Top Phantom` `[X] Digital Simulation` `[ ] Physical Simulation` `[ ] Other` |
+| **License** | CC BY 4.0 |
+| **Version** | `[1.1]` |
+
+---
+
+## 🎯 Tasks & Domain
+
+### Domain
+
+*Select the primary domain for this dataset.*
+
+- [X] **Surgical Robotics**
+- [ ] **Ultrasound Robotics**
+- [ ] **Other Healthcare Robotics** 
+
+### Demonstrated Skills
+
+*List the primary skills or procedures demonstrated in this dataset.*
+
+- Robotic cholecystectomy
+- Tissue grasping
+- Tissue prepping
+- Vessels clipping and cutting
+- Gallbladder dissection
+- Tissue diathermy / electrocautery
+
+---
+
+## 🔬 Data Collection Details
+
+### Collection Method
+
+*How was the data collected?*
+
+- [ ] **Human Teleoperation**
+- [ ] **Programmatic/State-Machine**
+- [ ] **AI Policy / Autonomous**
+- [X] **Other (Virtual Reality Simulator)**
+
+### Operator Details
+
+| | Description |
+| :--- | :--- |
+| **Operator Count** | `[4]` |
+| **Operator Skill Level** | `[X] Expert (e.g., Surgeon, Sonographer)` <br> `[ ] Intermediate (e.g., Trained Researcher)` <br> `[X] Novice (e.g., ML Researcher with minimal experience)` <br> `[ ] N/A` |
+| **Collection Period** | From `[2026-01-01]` to `[2026-01-16]` |
+
+### Recovery Demonstrations
+
+*Does this dataset include examples of recovering from failure?*
+
+- [X] **Yes**
+- [ ] **No**
+
+**If yes, please briefly describe the recovery process:**
+
+126 failure case trajectories and a matching number of recovery trajectories are included. The failures relate to the process of clipping vessels. A failure occurs if the user cuts a vessel or damages it excessively by improperly using diathermy. The recovery process consists of clipping the leaking vessel.
+
+---
+
+## 💡 Diversity Dimensions
+
+*Check all dimensions that were intentionally varied during data collection.*
+
+- [X] **Camera Position / Angle**
+- [X] **Lighting Conditions**
+- [ ] **Target Object** (e.g., different phantom models, suture types)
+- [ ] **Spatial Layout** (e.g., placing the target suture needle in various locations)
+- [ ] **Robot Embodiment** (if multiple robots were used)
+- [ ] **Task Execution** (e.g., different techniques for the same task)
+- [ ] **Background / Scene**
+- [ ] **Other** (Please specify: `[Your Dimension]`)
+
+*If you checked any of the above please briefly elaborate below.*
+
+Each attempt was rendered with 6 virtual cameras, each with 3 lighting scenarios:
+- 1 camera follows the user's view while performing the procedure, with FOV matched to the real FOV of the VR HMD.
+- 1 camera follows the user's view, with FOV set at a fixed values, to better represent the important areas during the procedure.
+- 4 remaining cameras are placed in static spots around the area of interest.
+
+The lighting scenarios include a spot light following the user's view, a spot light positioned at the currently rendering camera, and a fixed point light setup around the area of interest.
+Additionally, instrument starting positions and equipped instruments naturally vary between operators and between runs.
+
+
+---
+
+## 🛠️ Equipment & Setup
+
+### Robotic Platform(s)
+
+N/A (simulated environment)
+dVRK (simulated joints)
+
+### Sensors & Cameras
+
+*List the sensors and cameras used. Specify model names where possible. (Add and remove rows as needed)*
+
+| Type | Model/Details |
+| :--- | :--- |
+| **Virtual Mono Camera** | `[Simulated 3D camera, 1920x1080 @ 25-30fps]` |
+| **Virtual Stereo Camera** | `[Simulated 3D Stereo camera, 2x1080x1080 @ 30fps]` |
+
+
+In addition to shaded color views, additional camera views are provided:
+- Depth (reversed, 1 = closest, 0 = furthest)
+- View-space normals (scaled from [-1, 1] range into [0, 1] on each axis for video encoding purposes)
+- Segmentation mask
+- Optical flow generated from motion vectors (format based on https://www.mdpi.com/1424-8220/10/4/2975, hue represents angle while value represents the motion strength; value is in UV-space (UV coordinates, not pixel offsets) scaled 10x to reduce the impact of compression artifacts)
+
+Most trajectories (13 770) are provided using the mono camera. 
+1 602 trajectories are provided in stereo format, lenses are spaced 8 mm apart. The two camera views are provided side-by-side in one video and are not distorted. Due to technical constraints, stereo trajectories do not include optical flow views.
+
+---
+
+## 🎯 Action & State Space Representation
+
+*Describe how actions and robot states are represented in your dataset. This is crucial for understanding data compatibility and enabling effective policy learning.*
+
+### Action Space Representation
+
+**Primary Action Representation:**
+- [X] **Absolute Cartesian** (position/orientation relative to robot base)
+- [ ] **Relative Cartesian** (delta position/orientation from current pose)
+- [X] **Joint Space** (direct joint angle commands)
+- [ ] **Other** (Please specify: `[Your Representation]`)
+
+**Orientation Representation:**
+- [X] **Quaternions** (x, y, z, w)
+- [ ] **Euler Angles** (roll, pitch, yaw)
+- [ ] **Axis-Angle** (rotation vector)
+- [ ] **Rotation Matrix** (3x3 matrix)
+- [ ] **Other** (Please specify: `[Your Representation]`)
+
+**Reference Frame:**
+- [ ] **Robot Base Frame**
+- [ ] **Tool/End-Effector Frame**
+- [X] **World/Global Frame**
+- [X] **Camera Frame**
+- [ ] **Other** (Please specify: `[Your Frame]`)
+
+Both world and camera frames are provided as separate fields.
+
+**Action Dimensions:**
+*List the action space dimensions and their meanings.*
+
+**Example:**
+```
+action: [tip_pos_x, tip_pos_y, tip_pos_z, tip_rot_x, tip_rot_y, tip_rot_z, tip_rot_w,
+tip_cam_x, tip_cam_y, tip_cam_z, tip_cam_x, tip_cam_y, tip_cam_z, tip_cam_w,
+handle, gripper]
+- tip_pos_x, tip_pos_y, tip_pos_z: Absolute position of the end effector tip in world frame (meters, approximate)
+- tip_rot_x, tip_rot_y, tip_rot_z, tip_rot_w: Absolute rotation (quaternion) of the end effector tip in world frame
+- tip_cam_x, tip_cam_y, tip_cam_z: Relative position of the end effector tip in camera frame (meters, approximate)
+- tip_cam_x, tip_cam_y, tip_cam_z, tip_rot_w: Relative rotation (quaternion) of the end effector tip in camera frame
+- handle: user handle value, controls graspers, scissors, clipper activation. (0-1, with 0 being fully released and 1 being fully pressed)
+- gripper: Gripper opening angle (radians), this is valid only for instruments which have a gripper
+
+action.joint_positions: ["psm_yaw_joint", "psm_pitch_end_joint", "psm_main_insertion_joint", "psm_tool_roll_joint", "psm_tool_pitch_joint", "psm_tool_yaw_joint", "psm_tool_gripper1_joint", "psm_tool_gripper2_joint", "psm_pitch_back_joint", "psm_pitch_bottom_joint", "psm_pitch_top_joint", "psm_pitch_front_joint"]
+
+These actions are repeated 4 times, as there are up to 4 instruments available in the simulated procedure.
+
+```
+
+### State Space Representation
+
+**State Information Included:**
+- [X] **Joint Positions** (all articulated joints)
+- [ ] **Joint Velocities**
+- [X] **End-Effector Pose** (Cartesian position/orientation)
+- [ ] **Force/Torque Readings**
+- [X] **Gripper State** (position, force, etc.)
+- [ ] **Other** (Please specify: `[Your State Info]`)
+
+**State Dimensions:**
+State space is identical to action space with additional joint velocity data
+
+```
+observation.state_joints_pos: ["psm_yaw_joint", "psm_pitch_end_joint", "psm_main_insertion_joint", "psm_tool_roll_joint", "psm_tool_pitch_joint", "psm_tool_yaw_joint", "psm_tool_gripper1_joint", "psm_tool_gripper2_joint", "psm_pitch_back_joint", "psm_pitch_bottom_joint", "psm_pitch_top_joint", "psm_pitch_front_joint"]
+observation.state_joints_vel: ["psm_yaw_joint", "psm_pitch_end_joint", "psm_main_insertion_joint", "psm_tool_roll_joint", "psm_tool_pitch_joint", "psm_tool_yaw_joint", "psm_tool_gripper1_joint", "psm_tool_gripper2_joint", "psm_pitch_back_joint", "psm_pitch_bottom_joint", "psm_pitch_top_joint", "psm_pitch_front_joint"]
+```
+### 📋 Recommended Additional Representations
+
+*Even if not your primary action/state representation, we strongly encourage including these standardized formats for maximum compatibility:*
+
+**Recommended Action Fields:**
+- **`action.cartesian_absolute`**: Absolute Cartesian pose with absolute quaternions
+  ```
+  [x, y, z, qx, qy, qz, qw, gripper_angle]
+  ```
+
+**Available**, these are the same as action values: [tip_pos_x, tip_pos_y, tip_pos_z, tip_rot_x, tip_rot_y, tip_rot_z, tip_rot_w, gripper]
+
+**Recommended State Fields:**
+- **`observation.state.joint_positions`**: Absolute positions for all articulated joints
+  ```
+  [joint_1, joint_2, ..., joint_n]
+  ```
+
+  **Available** as the simulation does not include computed joints.
+
+
+---
+
+## ⏱️ Data Synchronization Approach
+
+*Describe how you achieved proper data synchronization across different sensors, cameras, and robotic systems during data collection. This is crucial for ensuring temporal alignment of all modalities in your dataset.*
+
+All sensors and cameras are inherently perfectly synchronized due to the simulated nature of the dataset. At the start of the frame, instruments are updated from the most recent available input data (from connected controllers). Camera views are rendered later, but the instrument positions will not be updated until the next frame starts processing. Therefore the sensor/action data is guaranteed to be synchronized with the camera data. 
+
+---
+
+## 👥 Attribution & Contact
+
+*Please provide attribution for the dataset creators and a point of contact.*
+
+| | |
+| :--- | :--- |
+| **Dataset Lead** | `Mateusz Wójcikowski, Przemysław Korzeniowski, Sabina Martyniak, Michał Naskręt, Prof Filippo Filicori, Dr Aditya Amit Godbole, Dr Maria Clara Morais, Dr Mattia Ballo` |
+| **Institution** | `Sano - Centre for Computational Medicine` |
+| **Contact Email** | `m.wojcikowski@sanoscience.org, p.korzeniowski@sanoscience.org` |
+| **Citation (BibTeX)** | <pre><code>@misc{SANO_VR_Cholecystectomy_2025,<br>  author = {Mateusz Wójcikowski, Przemysław Korzeniowski, Sabina Martyniak, Michał Naskręt, Prof Filippo Filicori, Dr Aditya Amit Godbole, Dr Maria Clara Morais, Dr Mattia Ballo},<br>  title = {SANO_VR_Cholecystectomy},<br>  year = {2025},<br>  publisher = {Open-H-Embodiment},<br>  note = {Simulated Domain: Robotic Cholecystectomy}<br>}</code></pre> |

Surgical/sanoscience/sanoscience_v1_2_merged/nonexpert_stereo/README.md

@@ -0,0 +1,259 @@
+# SANO_VR_Cholecystectomy - README
+
+---
+
+## 📋 At a Glance
+
+*Provide a one-sentence summary of your dataset.*
+
+Cholecystectomy procedure performed in a Virtual Reality simulator.
+
+---
+
+## 📖 Dataset Overview
+
+*Briefly describe the purpose and content of this dataset. What key skills or scenarios does it demonstrate?*
+
+This dataset contains recordings of both experts and non-experts performing robotic cholecystectomy in a Virtual Reality simulated environment. It includes successful attempts as well as failures and recovery attempts.
+The data includes multiple camera angles, different lighting scenarios, different quality levels of simulated anatomy, and each camera view includes additional depth data, segmentation masks, approximate surface normals and approximated optical flow.
+
+Data breakdown:
+15 372 total trajectories, of which:
+- 5 454 are performed by experts
+- 9 918 are performed by non-experts
+
+- 15 120 are successes
+- 126 + 126 are failures + recoveries
+
+Episodes vary in length from 2 seconds to 10 seconds.
+The data contains both per-episode and per-frame labels. Both are generated automatically from simulation data based on current instrument state, including activation state and physics contacts. Per-frame labels are generated directly in this way, while per-episode labels are generated by finding the closest event in the near future. If task can't be determined, "Unknown" is set as the task. As these labels are generated automatically, they may not always accurately represent the intention of the operator.
+
+The observations and actions for dVRK PSM manipulator joints are obtained from nVidia Newton inverse kinematic simulation and may not be precise.
+IMPORTANT - PATCH: In order to use the joints data, the videos and images folders from the v1.1 directory should be copied here.
+
+| | |
+| :--- | :--- |
+| **Total Trajectories** | `[15 372]` |
+| **Total Hours** | `[~0.94h augmented x18, total ~17h]` |
+| **Data Type** | `[ ] Clinical` `[ ] Ex-Vivo` `[ ] Table-Top Phantom` `[X] Digital Simulation` `[ ] Physical Simulation` `[ ] Other` |
+| **License** | CC BY 4.0 |
+| **Version** | `[1.1]` |
+
+---
+
+## 🎯 Tasks & Domain
+
+### Domain
+
+*Select the primary domain for this dataset.*
+
+- [X] **Surgical Robotics**
+- [ ] **Ultrasound Robotics**
+- [ ] **Other Healthcare Robotics** 
+
+### Demonstrated Skills
+
+*List the primary skills or procedures demonstrated in this dataset.*
+
+- Robotic cholecystectomy
+- Tissue grasping
+- Tissue prepping
+- Vessels clipping and cutting
+- Gallbladder dissection
+- Tissue diathermy / electrocautery
+
+---
+
+## 🔬 Data Collection Details
+
+### Collection Method
+
+*How was the data collected?*
+
+- [ ] **Human Teleoperation**
+- [ ] **Programmatic/State-Machine**
+- [ ] **AI Policy / Autonomous**
+- [X] **Other (Virtual Reality Simulator)**
+
+### Operator Details
+
+| | Description |
+| :--- | :--- |
+| **Operator Count** | `[4]` |
+| **Operator Skill Level** | `[X] Expert (e.g., Surgeon, Sonographer)` <br> `[ ] Intermediate (e.g., Trained Researcher)` <br> `[X] Novice (e.g., ML Researcher with minimal experience)` <br> `[ ] N/A` |
+| **Collection Period** | From `[2026-01-01]` to `[2026-01-16]` |
+
+### Recovery Demonstrations
+
+*Does this dataset include examples of recovering from failure?*
+
+- [X] **Yes**
+- [ ] **No**
+
+**If yes, please briefly describe the recovery process:**
+
+126 failure case trajectories and a matching number of recovery trajectories are included. The failures relate to the process of clipping vessels. A failure occurs if the user cuts a vessel or damages it excessively by improperly using diathermy. The recovery process consists of clipping the leaking vessel.
+
+---
+
+## 💡 Diversity Dimensions
+
+*Check all dimensions that were intentionally varied during data collection.*
+
+- [X] **Camera Position / Angle**
+- [X] **Lighting Conditions**
+- [ ] **Target Object** (e.g., different phantom models, suture types)
+- [ ] **Spatial Layout** (e.g., placing the target suture needle in various locations)
+- [ ] **Robot Embodiment** (if multiple robots were used)
+- [ ] **Task Execution** (e.g., different techniques for the same task)
+- [ ] **Background / Scene**
+- [ ] **Other** (Please specify: `[Your Dimension]`)
+
+*If you checked any of the above please briefly elaborate below.*
+
+Each attempt was rendered with 6 virtual cameras, each with 3 lighting scenarios:
+- 1 camera follows the user's view while performing the procedure, with FOV matched to the real FOV of the VR HMD.
+- 1 camera follows the user's view, with FOV set at a fixed values, to better represent the important areas during the procedure.
+- 4 remaining cameras are placed in static spots around the area of interest.
+
+The lighting scenarios include a spot light following the user's view, a spot light positioned at the currently rendering camera, and a fixed point light setup around the area of interest.
+Additionally, instrument starting positions and equipped instruments naturally vary between operators and between runs.
+
+
+---
+
+## 🛠️ Equipment & Setup
+
+### Robotic Platform(s)
+
+N/A (simulated environment)
+dVRK (simulated joints)
+
+### Sensors & Cameras
+
+*List the sensors and cameras used. Specify model names where possible. (Add and remove rows as needed)*
+
+| Type | Model/Details |
+| :--- | :--- |
+| **Virtual Mono Camera** | `[Simulated 3D camera, 1920x1080 @ 25-30fps]` |
+| **Virtual Stereo Camera** | `[Simulated 3D Stereo camera, 2x1080x1080 @ 30fps]` |
+
+
+In addition to shaded color views, additional camera views are provided:
+- Depth (reversed, 1 = closest, 0 = furthest)
+- View-space normals (scaled from [-1, 1] range into [0, 1] on each axis for video encoding purposes)
+- Segmentation mask
+- Optical flow generated from motion vectors (format based on https://www.mdpi.com/1424-8220/10/4/2975, hue represents angle while value represents the motion strength; value is in UV-space (UV coordinates, not pixel offsets) scaled 10x to reduce the impact of compression artifacts)
+
+Most trajectories (13 770) are provided using the mono camera. 
+1 602 trajectories are provided in stereo format, lenses are spaced 8 mm apart. The two camera views are provided side-by-side in one video and are not distorted. Due to technical constraints, stereo trajectories do not include optical flow views.
+
+---
+
+## 🎯 Action & State Space Representation
+
+*Describe how actions and robot states are represented in your dataset. This is crucial for understanding data compatibility and enabling effective policy learning.*
+
+### Action Space Representation
+
+**Primary Action Representation:**
+- [X] **Absolute Cartesian** (position/orientation relative to robot base)
+- [ ] **Relative Cartesian** (delta position/orientation from current pose)
+- [X] **Joint Space** (direct joint angle commands)
+- [ ] **Other** (Please specify: `[Your Representation]`)
+
+**Orientation Representation:**
+- [X] **Quaternions** (x, y, z, w)
+- [ ] **Euler Angles** (roll, pitch, yaw)
+- [ ] **Axis-Angle** (rotation vector)
+- [ ] **Rotation Matrix** (3x3 matrix)
+- [ ] **Other** (Please specify: `[Your Representation]`)
+
+**Reference Frame:**
+- [ ] **Robot Base Frame**
+- [ ] **Tool/End-Effector Frame**
+- [X] **World/Global Frame**
+- [X] **Camera Frame**
+- [ ] **Other** (Please specify: `[Your Frame]`)
+
+Both world and camera frames are provided as separate fields.
+
+**Action Dimensions:**
+*List the action space dimensions and their meanings.*
+
+**Example:**
+```
+action: [tip_pos_x, tip_pos_y, tip_pos_z, tip_rot_x, tip_rot_y, tip_rot_z, tip_rot_w,
+tip_cam_x, tip_cam_y, tip_cam_z, tip_cam_x, tip_cam_y, tip_cam_z, tip_cam_w,
+handle, gripper]
+- tip_pos_x, tip_pos_y, tip_pos_z: Absolute position of the end effector tip in world frame (meters, approximate)
+- tip_rot_x, tip_rot_y, tip_rot_z, tip_rot_w: Absolute rotation (quaternion) of the end effector tip in world frame
+- tip_cam_x, tip_cam_y, tip_cam_z: Relative position of the end effector tip in camera frame (meters, approximate)
+- tip_cam_x, tip_cam_y, tip_cam_z, tip_rot_w: Relative rotation (quaternion) of the end effector tip in camera frame
+- handle: user handle value, controls graspers, scissors, clipper activation. (0-1, with 0 being fully released and 1 being fully pressed)
+- gripper: Gripper opening angle (radians), this is valid only for instruments which have a gripper
+
+action.joint_positions: ["psm_yaw_joint", "psm_pitch_end_joint", "psm_main_insertion_joint", "psm_tool_roll_joint", "psm_tool_pitch_joint", "psm_tool_yaw_joint", "psm_tool_gripper1_joint", "psm_tool_gripper2_joint", "psm_pitch_back_joint", "psm_pitch_bottom_joint", "psm_pitch_top_joint", "psm_pitch_front_joint"]
+
+These actions are repeated 4 times, as there are up to 4 instruments available in the simulated procedure.
+
+```
+
+### State Space Representation
+
+**State Information Included:**
+- [X] **Joint Positions** (all articulated joints)
+- [ ] **Joint Velocities**
+- [X] **End-Effector Pose** (Cartesian position/orientation)
+- [ ] **Force/Torque Readings**
+- [X] **Gripper State** (position, force, etc.)
+- [ ] **Other** (Please specify: `[Your State Info]`)
+
+**State Dimensions:**
+State space is identical to action space with additional joint velocity data
+
+```
+observation.state_joints_pos: ["psm_yaw_joint", "psm_pitch_end_joint", "psm_main_insertion_joint", "psm_tool_roll_joint", "psm_tool_pitch_joint", "psm_tool_yaw_joint", "psm_tool_gripper1_joint", "psm_tool_gripper2_joint", "psm_pitch_back_joint", "psm_pitch_bottom_joint", "psm_pitch_top_joint", "psm_pitch_front_joint"]
+observation.state_joints_vel: ["psm_yaw_joint", "psm_pitch_end_joint", "psm_main_insertion_joint", "psm_tool_roll_joint", "psm_tool_pitch_joint", "psm_tool_yaw_joint", "psm_tool_gripper1_joint", "psm_tool_gripper2_joint", "psm_pitch_back_joint", "psm_pitch_bottom_joint", "psm_pitch_top_joint", "psm_pitch_front_joint"]
+```
+### 📋 Recommended Additional Representations
+
+*Even if not your primary action/state representation, we strongly encourage including these standardized formats for maximum compatibility:*
+
+**Recommended Action Fields:**
+- **`action.cartesian_absolute`**: Absolute Cartesian pose with absolute quaternions
+  ```
+  [x, y, z, qx, qy, qz, qw, gripper_angle]
+  ```
+
+**Available**, these are the same as action values: [tip_pos_x, tip_pos_y, tip_pos_z, tip_rot_x, tip_rot_y, tip_rot_z, tip_rot_w, gripper]
+
+**Recommended State Fields:**
+- **`observation.state.joint_positions`**: Absolute positions for all articulated joints
+  ```
+  [joint_1, joint_2, ..., joint_n]
+  ```
+
+  **Available** as the simulation does not include computed joints.
+
+
+---
+
+## ⏱️ Data Synchronization Approach
+
+*Describe how you achieved proper data synchronization across different sensors, cameras, and robotic systems during data collection. This is crucial for ensuring temporal alignment of all modalities in your dataset.*
+
+All sensors and cameras are inherently perfectly synchronized due to the simulated nature of the dataset. At the start of the frame, instruments are updated from the most recent available input data (from connected controllers). Camera views are rendered later, but the instrument positions will not be updated until the next frame starts processing. Therefore the sensor/action data is guaranteed to be synchronized with the camera data. 
+
+---
+
+## 👥 Attribution & Contact
+
+*Please provide attribution for the dataset creators and a point of contact.*
+
+| | |
+| :--- | :--- |
+| **Dataset Lead** | `Mateusz Wójcikowski, Przemysław Korzeniowski, Sabina Martyniak, Michał Naskręt, Prof Filippo Filicori, Dr Aditya Amit Godbole, Dr Maria Clara Morais, Dr Mattia Ballo` |
+| **Institution** | `Sano - Centre for Computational Medicine` |
+| **Contact Email** | `m.wojcikowski@sanoscience.org, p.korzeniowski@sanoscience.org` |
+| **Citation (BibTeX)** | <pre><code>@misc{SANO_VR_Cholecystectomy_2025,<br>  author = {Mateusz Wójcikowski, Przemysław Korzeniowski, Sabina Martyniak, Michał Naskręt, Prof Filippo Filicori, Dr Aditya Amit Godbole, Dr Maria Clara Morais, Dr Mattia Ballo},<br>  title = {SANO_VR_Cholecystectomy},<br>  year = {2025},<br>  publisher = {Open-H-Embodiment},<br>  note = {Simulated Domain: Robotic Cholecystectomy}<br>}</code></pre> |

Surgical/tud/260131_tundra_dataset/endoscope_guidance/README.md

@@ -0,0 +1,260 @@
+<!--
+Open-H Embodiment Dataset README Template (v1.0)
+-->
+
+# TUNDRA (**TU**D & **N**CT **D**resden **R**obot **A**ssistant) Dataset - README
+
+---
+
+## 📋 At a Glance
+
+Teleoperated demonstrations of a Universal Robots UR5e robot performing endoscope guidance, grasping and retraction of organs in cooperation with a surgeon during *in vivo* laparoscopic surgeries on pigs.
+
+---
+
+## 📖 Dataset Overview
+
+This dataset contains 196 trajectories of a resident surgeon using a Universal Robots UR5e to perform surgical assistance tasks (grasping, retraction and endoscope guidance) during laparoscopic surgeries on *in vivo* porcine models. It includes variable lighting, trocar positions and angles as well as target organs. Fail-states are followed by recoveries. This dataset can be used to train imitation learning policies.
+
+| | |
+| :--- | :--- |
+| **Total Trajectories** | `196` |
+| **Total Hours** | `0.96 hours (104,237 frames)` |
+| **Data Type** | `[ ] Clinical` `[ ] Ex-Vivo` `[ ] Table-Top Phantom` `[ ] Digital Simulation` `[ ] Physical Simulation` `[X] In-Vivo` |
+| **License** | CC BY 4.0 |
+| **Version** | `1.0` |
+
+---
+
+## 🎯 Tasks & Domain
+
+### Domain
+
+- [X] **Surgical Robotics**
+- [ ] **Ultrasound Robotics**
+- [ ] **Other Healthcare Robotics**
+
+### Demonstrated Skills
+
+**Endoscope guidance**
+
+`0. Overview`: Broad visualization of the operative field for context setting. Starts when the camera switches to a wide view without following or focusing on something specific.
+
+`1. Focus` Narrowing the field of view, centering on a specific anatomical structure or task, reducing irrelevant background. Starts when camera movement clearly targets a specific area of interest. If the surgeon is working and the camera is only zooming in, it is Focus (not Follow).
+
+`2. Follow where the surgeon is working` Maintenance of optimal working distance and tracking of active instruments (tip). Starts when surgeon begins manipulating tissue or instruments within the focused area.
+
+`3. Follow the target*` Emphasis on exposed anatomical structures with camera movement aligned with anatomical progression. Starts when surgical workflow shifts from tool-driven movement to anatomical exploration or exposure.
+
+`4. Relevel the horizon` Camera rotation to restore visual orientation. Starts when background is visibly tilted or anatomical orientation is unclear. Ends when background horizon appears level and camera rotation stops.
+
+`5. Stop` Static camera position with no intentional camera movement. Starts with the end of another phase and ends with the start of another phase.
+
+`6. Correction` Fixes to framing, focus, zoom, or orientation when camera framing, focus or orientation is suboptimal. Typically followed by Confirmation.
+
+`7. Confirmation` Surgeon accepts camera view (event).
+
+`8. End` when the endoscope is pulled back inside the trocar after after task completion.
+
+**Grasping**
+
+`0. Wait` - Staying in place as robot waits for the surgeon to grasp the target* and offers it for handover to the robot.
+
+`1. Approach` - Movement toward the target* being held by the surgeon and grasper opening for handover.
+
+`2. Grasp` -  Grasper jaws closing to grasp target* and not moving until surgeon secures the other end of the target*.
+
+**Retraction**
+
+`3. Retract` With both ends of the target* grasped, tensioning to straighten the target* in the camera frame completely.
+
+`4. Hold` Maintaining grip and position on the straightened target*.
+
+`5. Stop` - Stop the grasper movement and stay.
+
+`6. Correct` Correct the previous movement or position of the grasper.
+
+*target: `[ small intestine | colon | gallbladder | stomach | omentum ]`
+
+---
+
+## 🔬 Data Collection Details
+
+### Collection Method
+
+- [X] **Human Teleoperation**
+- [ ] **Programmatic/State-Machine**
+- [ ] **AI Policy / Autonomous**
+- [ ] **Other**
+
+### Operator Details
+
+| | Description |
+| :--- | :--- |
+| **Operator Count** | `1` |
+| **Operator Skill Level** | `[X] Expert (resident surgeon with 3 years of experience in surgical assistance and in vivo porcine surgery)` <br> `[ ] Intermediate (e.g., Trained Researcher)` <br> `[ ] Novice (e.g., ML Researcher with minimal experience)` <br> `[ ] N/A` |
+| **Collection Period** | From `2025-11-06` to `2025-12-04` |
+
+### Recovery Demonstrations
+
+- [X] **Yes**
+- [ ] **No**
+
+Multiple demonstrations start with naturalistically observed failure occurrences with subsequent recoveries:
+- Missed grasps
+- False grasps
+- Tissue slips
+
+Many demonstrations also include the successful handling of difficult situations:
+- Robot instrument out-of-frame
+- Obstruction of robot instrument (e.g. by surgeon instrument or tissue)
+- Obstruction of target (i.e. partially covered or hidden, no full obstruction)
+
+---
+
+## 💡 Diversity Dimensions
+
+- [X] **Camera Position / Angle**
+- [X] **Lighting Conditions**
+- [X] **Target Object** 
+- [X] **Spatial Layout** 
+- [ ] **Robot Embodiment**
+- [ ] **Task Execution**
+- [X] **Background / Scene** 
+- [ ] **Other**
+
+We changed the position and angle of the endoscope inside the trocar every 5 to 10 demonstrations. 
+We varied the light intensity from 10% to 100% of the maximum of our endoscope's light source.
+We targeted small bowel, large bowel, stomach, gallbladder and omentum.
+We varied the targeted organ segments (e.g., which part of the small bowel) and starting positions of the instruments.
+We targeted organs in a total of three different abdominal quadrants (gallbladder and bowel in RUQ, stomach, omentum and bowel in LUQ, bowel in RLQ) with substantially different backgrounds.
+
+---
+
+## 🛠️ Equipment & Setup
+
+### Robotic Platform(s)
+
+- **Robot:** `Universal Robots UR5e (Universal Robots, Odense, Denmark)`
+
+### Sensors & Cameras
+
+| Type | Model/Details |
+| :--- | :--- |
+| **Primary Camera** | `Karl Storz TIPCAM 1S stereo endoscope (Karl Storz SE & Co. KG, Tuttlingen, Germany), left and right streams with a resolution of 960x540 @ 30 Hz` |
+| **Instrument interface** | `Custom-made KOALA-Grasp mechatronic interface for Karl Storz CLICKLINE instruments (Karl Storz SE & Co. KG, Tuttlingen, Germany)` |
+
+---
+
+## 🎯 Action & State Space Representation
+
+### Action Space Representation
+
+**Primary Action Representation:**
+- [ ] **Absolute Cartesian** (position/orientation relative to robot base)
+- [X] **Relative Cartesian** (delta position/orientation from current pose, depending on action type, see below)
+- [ ] **Joint Space** (direct joint angle commands)
+- [ ] **Other**
+
+**Orientation Representation:**
+- [X] **Quaternions** (x, y, z, w)
+- [ ] **Euler Angles** (roll, pitch, yaw)
+- [ ] **Axis-Angle** (rotation vector)
+- [ ] **Rotation Matrix** (3x3 matrix)
+- [X] **Other** (`Roll`)
+
+**Reference Frame:**
+- [X] **Robot Base Frame** (for endoscope guidance)
+- [ ] **Tool/End-Effector Frame**
+- [ ] **World/Global Frame**
+- [X] **Camera Frame** (for grasping and retraction)
+- [ ] **Other**
+
+**Action Dimensions:**
+
+For endoscope guidance:
+```
+[dx, dy, dz, dr]
+- dx, dy, dz: relative tip position in robot base frame (meters)
+- dr: relative tip roll in robot base frame (radiants)
+```
+
+For grasping and retraction:
+```
+[dx, dy, dz, dr, grasper]
+- dx, dy, dz: relative tip position in endoscope frame (meters)
+- dr: relative tip roll in endoscope frame (radiants)
+- grasper: grasper opening condition (1=open / 0=closed)
+```
+
+### State Space Representation
+
+**State Information Included:**
+- [X] **Joint Positions** (all articulated joints)
+- [X] **Joint Velocities**
+- [X] **End-Effector Pose** (cartesian position/orientation)
+- [X] **Force/Torque Readings** (joint efforts)
+- [X] **Grasper State** (opened/closed)
+- [X] **Quaternion between endoscope frame and robot base frame**
+- [X] **Initial tip position in robot base frame**
+- [X] **Position of the remote centre of motion in robot base frame**
+
+
+**State Dimensions:**
+
+#### Grasping and retraction
+
+**Observations**
+
+- `observation.images.laparoscope_left`: Left stereo camera view (540x960x3)
+- `observation.images.laparoscope_right`: Right stereo camera view (540x960x3)
+- `observation.state`: Robot state vector (35 dimensions for endoscope guidance, 40 dimensions for grasping and retraction):
+  - `base_to_camera_quat_x/y/z/w`* (4): Camera orientation quaternion
+  - `gripper_opened`* (1): Grasper state (open/closed)
+  - `relative_tip_position_x/y/z` (3): Tip position relative to RCM
+  - `shoulder_pan/lift_joint_pos`, `elbow_joint_pos`, `wrist_1/2/3_joint_pos` (6): UR5e joint positions
+  - `shoulder_pan/lift_joint_vel`, `elbow_joint_vel`, `wrist_1/2/3_joint_vel` (6): UR5e joint velocities
+  - `shoulder_pan/lift_joint_effort`, `elbow_joint_effort`, `wrist_1/2/3_joint_effort` (6): UR5e joint efforts
+  - `base_frame_ee_position_x/y/z` (3): End-effector position in base frame
+  - `base_frame_ee_orientation_x/y/z/w` (4): End-effector orientation quaternion in base frame
+  - `base_frame_initial_tip_position_x/y/z` (3): Initial tip position in base frame
+  - `base_frame_rcm_position_x/y/z` (3): RCM position in base frame
+  - `roll_angle` (1): Rotation around shaft axis
+
+**Actions**
+
+- `action`: control vector (4 dimensions for endoscope guidance, 5 dimensions for grasping and retaction)
+  - `delta_camera_frame_tip_position_x/y/z` (3): Delta tip position in camera frame
+  - `delta_roll_angle` (1): Delta roll angle around shaft axis
+  - `open_gripper`* (1): Gripper command
+
+**Metadata**
+
+- `timestamp`: Time in seconds
+- `episode_index`: Episode identifier
+- `frame_index`: Frame number within episode
+- `index`: Global frame index across all episodes
+- `task_index`: Task category (0-4)
+- `next.done`: Episode termination flag
+
+*only for grasping and retraction
+
+---
+
+## ⏱️ Data Synchronization Approach
+
+The clocks of all the involved machines and therefore the data timestamps were synchronized using the Precision Time Protocol (PTP).
+The gripper state (open/closed) was annotated manually in post-processing using the video.
+
+---
+
+## 👥 Attribution & Contact
+
+| | |
+| :--- | :--- |
+| **Dataset Lead (Principal Investigators)** | `[Stefanie Speidel, Martin Wagner]` |
+| **Dataset Lead (MD/PhD)** | `[Rayan Younis, Ariel Rodriguez, Lorenzo Mazza, Ortrun Hellig]` |
+| **Institution** | `[Department for Translational Surgical Oncology, National Center for Tumor Diseases, Partner Site Dresden (NCT/UCC): German Cancer Research Center (DKFZ), Helmholtz-Zentrum Dresden - Rossendorf (HZDR); Department of Visceral, Thoracic and Vascular Surgery, Faculty of Medicine and University Hospital Carl Gustav Carus, TUD Dresden University of Technology; Centre for the Tactile Internet with Human-in-the-Loop (CeTI), TUD Dresden University of Technology]` |
+| **Contact Email** | `[stefanie.speidel@nct-dresden.de, martin.wagner@ukdd.de]` |
+| **Acknowledgements** | `[Funded by the German Research Foundation (DFG, Deutsche Forschungsgemeinschaft) as part of Germany’s Excellence Strategy – EXC 2050/2 – Project ID 390696704 – Cluster of Excellence “Centre for Tactile Internet with Human-in-the-Loop” (CeTI) of TUD Dresden University of Technology. The authors acknowledge the financial support by the Federal Ministry of Research, Technology and Space of Germany in the programme of “Souverän. Digital. Vernetzt.”. Joint project 6G-life, project identification number: 16KISK001K. We thank the team of the Experimental Operating Room at the NCT Dresden for their support.]` |
+| **Citation (BibTeX)** | <pre><code>@misc{[TUNDRA_dataset_2026],<br>  author = {[Rodriguez, Ariel and Younis, Rayan and Mazza, Lorenzo and Hellig, Ortrun and Wagner, Martin and Speidel, Stefanie]},<br>  title = {[TUD & NCT Dresden Robot Assistant Dataset]},<br>  year = {2026},<br>  publisher = {Open-H-Embodiment},<br>  note = {}<br>}</code></pre> |

Surgical/tud/260131_tundra_dataset/grasping_retraction/README.md

@@ -0,0 +1,260 @@
+<!--
+Open-H Embodiment Dataset README Template (v1.0)
+-->
+
+# TUNDRA (**TU**D & **N**CT **D**resden **R**obot **A**ssistant) Dataset - README
+
+---
+
+## 📋 At a Glance
+
+Teleoperated demonstrations of a Universal Robots UR5e robot performing endoscope guidance, grasping and retraction of organs in cooperation with a surgeon during *in vivo* laparoscopic surgeries on pigs.
+
+---
+
+## 📖 Dataset Overview
+
+This dataset contains 196 trajectories of a resident surgeon using a Universal Robots UR5e to perform surgical assistance tasks (grasping, retraction and endoscope guidance) during laparoscopic surgeries on *in vivo* porcine models. It includes variable lighting, trocar positions and angles as well as target organs. Fail-states are followed by recoveries. This dataset can be used to train imitation learning policies.
+
+| | |
+| :--- | :--- |
+| **Total Trajectories** | `196` |
+| **Total Hours** | `0.96 hours (104,237 frames)` |
+| **Data Type** | `[ ] Clinical` `[ ] Ex-Vivo` `[ ] Table-Top Phantom` `[ ] Digital Simulation` `[ ] Physical Simulation` `[X] In-Vivo` |
+| **License** | CC BY 4.0 |
+| **Version** | `1.0` |
+
+---
+
+## 🎯 Tasks & Domain
+
+### Domain
+
+- [X] **Surgical Robotics**
+- [ ] **Ultrasound Robotics**
+- [ ] **Other Healthcare Robotics**
+
+### Demonstrated Skills
+
+**Endoscope guidance**
+
+`0. Overview`: Broad visualization of the operative field for context setting. Starts when the camera switches to a wide view without following or focusing on something specific.
+
+`1. Focus` Narrowing the field of view, centering on a specific anatomical structure or task, reducing irrelevant background. Starts when camera movement clearly targets a specific area of interest. If the surgeon is working and the camera is only zooming in, it is Focus (not Follow).
+
+`2. Follow where the surgeon is working` Maintenance of optimal working distance and tracking of active instruments (tip). Starts when surgeon begins manipulating tissue or instruments within the focused area.
+
+`3. Follow the target*` Emphasis on exposed anatomical structures with camera movement aligned with anatomical progression. Starts when surgical workflow shifts from tool-driven movement to anatomical exploration or exposure.
+
+`4. Relevel the horizon` Camera rotation to restore visual orientation. Starts when background is visibly tilted or anatomical orientation is unclear. Ends when background horizon appears level and camera rotation stops.
+
+`5. Stop` Static camera position with no intentional camera movement. Starts with the end of another phase and ends with the start of another phase.
+
+`6. Correction` Fixes to framing, focus, zoom, or orientation when camera framing, focus or orientation is suboptimal. Typically followed by Confirmation.
+
+`7. Confirmation` Surgeon accepts camera view (event).
+
+`8. End` when the endoscope is pulled back inside the trocar after after task completion.
+
+**Grasping**
+
+`0. Wait` - Staying in place as robot waits for the surgeon to grasp the target* and offers it for handover to the robot.
+
+`1. Approach` - Movement toward the target* being held by the surgeon and grasper opening for handover.
+
+`2. Grasp` -  Grasper jaws closing to grasp target* and not moving until surgeon secures the other end of the target*.
+
+**Retraction**
+
+`3. Retract` With both ends of the target* grasped, tensioning to straighten the target* in the camera frame completely.
+
+`4. Hold` Maintaining grip and position on the straightened target*.
+
+`5. Stop` - Stop the grasper movement and stay.
+
+`6. Correct` Correct the previous movement or position of the grasper.
+
+*target: `[ small intestine | colon | gallbladder | stomach | omentum ]`
+
+---
+
+## 🔬 Data Collection Details
+
+### Collection Method
+
+- [X] **Human Teleoperation**
+- [ ] **Programmatic/State-Machine**
+- [ ] **AI Policy / Autonomous**
+- [ ] **Other**
+
+### Operator Details
+
+| | Description |
+| :--- | :--- |
+| **Operator Count** | `1` |
+| **Operator Skill Level** | `[X] Expert (resident surgeon with 3 years of experience in surgical assistance and in vivo porcine surgery)` <br> `[ ] Intermediate (e.g., Trained Researcher)` <br> `[ ] Novice (e.g., ML Researcher with minimal experience)` <br> `[ ] N/A` |
+| **Collection Period** | From `2025-11-06` to `2025-12-04` |
+
+### Recovery Demonstrations
+
+- [X] **Yes**
+- [ ] **No**
+
+Multiple demonstrations start with naturalistically observed failure occurrences with subsequent recoveries:
+- Missed grasps
+- False grasps
+- Tissue slips
+
+Many demonstrations also include the successful handling of difficult situations:
+- Robot instrument out-of-frame
+- Obstruction of robot instrument (e.g. by surgeon instrument or tissue)
+- Obstruction of target (i.e. partially covered or hidden, no full obstruction)
+
+---
+
+## 💡 Diversity Dimensions
+
+- [X] **Camera Position / Angle**
+- [X] **Lighting Conditions**
+- [X] **Target Object** 
+- [X] **Spatial Layout** 
+- [ ] **Robot Embodiment**
+- [ ] **Task Execution**
+- [X] **Background / Scene** 
+- [ ] **Other**
+
+We changed the position and angle of the endoscope inside the trocar every 5 to 10 demonstrations. 
+We varied the light intensity from 10% to 100% of the maximum of our endoscope's light source.
+We targeted small bowel, large bowel, stomach, gallbladder and omentum.
+We varied the targeted organ segments (e.g., which part of the small bowel) and starting positions of the instruments.
+We targeted organs in a total of three different abdominal quadrants (gallbladder and bowel in RUQ, stomach, omentum and bowel in LUQ, bowel in RLQ) with substantially different backgrounds.
+
+---
+
+## 🛠️ Equipment & Setup
+
+### Robotic Platform(s)
+
+- **Robot:** `Universal Robots UR5e (Universal Robots, Odense, Denmark)`
+
+### Sensors & Cameras
+
+| Type | Model/Details |
+| :--- | :--- |
+| **Primary Camera** | `Karl Storz TIPCAM 1S stereo endoscope (Karl Storz SE & Co. KG, Tuttlingen, Germany), left and right streams with a resolution of 960x540 @ 30 Hz` |
+| **Instrument interface** | `Custom-made KOALA-Grasp mechatronic interface for Karl Storz CLICKLINE instruments (Karl Storz SE & Co. KG, Tuttlingen, Germany)` |
+
+---
+
+## 🎯 Action & State Space Representation
+
+### Action Space Representation
+
+**Primary Action Representation:**
+- [ ] **Absolute Cartesian** (position/orientation relative to robot base)
+- [X] **Relative Cartesian** (delta position/orientation from current pose, depending on action type, see below)
+- [ ] **Joint Space** (direct joint angle commands)
+- [ ] **Other**
+
+**Orientation Representation:**
+- [X] **Quaternions** (x, y, z, w)
+- [ ] **Euler Angles** (roll, pitch, yaw)
+- [ ] **Axis-Angle** (rotation vector)
+- [ ] **Rotation Matrix** (3x3 matrix)
+- [X] **Other** (`Roll`)
+
+**Reference Frame:**
+- [X] **Robot Base Frame** (for endoscope guidance)
+- [ ] **Tool/End-Effector Frame**
+- [ ] **World/Global Frame**
+- [X] **Camera Frame** (for grasping and retraction)
+- [ ] **Other**
+
+**Action Dimensions:**
+
+For endoscope guidance:
+```
+[dx, dy, dz, dr]
+- dx, dy, dz: relative tip position in robot base frame (meters)
+- dr: relative tip roll in robot base frame (radiants)
+```
+
+For grasping and retraction:
+```
+[dx, dy, dz, dr, grasper]
+- dx, dy, dz: relative tip position in endoscope frame (meters)
+- dr: relative tip roll in endoscope frame (radiants)
+- grasper: grasper opening condition (1=open / 0=closed)
+```
+
+### State Space Representation
+
+**State Information Included:**
+- [X] **Joint Positions** (all articulated joints)
+- [X] **Joint Velocities**
+- [X] **End-Effector Pose** (cartesian position/orientation)
+- [X] **Force/Torque Readings** (joint efforts)
+- [X] **Grasper State** (opened/closed)
+- [X] **Quaternion between endoscope frame and robot base frame**
+- [X] **Initial tip position in robot base frame**
+- [X] **Position of the remote centre of motion in robot base frame**
+
+
+**State Dimensions:**
+
+#### Grasping and retraction
+
+**Observations**
+
+- `observation.images.laparoscope_left`: Left stereo camera view (540x960x3)
+- `observation.images.laparoscope_right`: Right stereo camera view (540x960x3)
+- `observation.state`: Robot state vector (35 dimensions for endoscope guidance, 40 dimensions for grasping and retraction):
+  - `base_to_camera_quat_x/y/z/w`* (4): Camera orientation quaternion
+  - `gripper_opened`* (1): Grasper state (open/closed)
+  - `relative_tip_position_x/y/z` (3): Tip position relative to RCM
+  - `shoulder_pan/lift_joint_pos`, `elbow_joint_pos`, `wrist_1/2/3_joint_pos` (6): UR5e joint positions
+  - `shoulder_pan/lift_joint_vel`, `elbow_joint_vel`, `wrist_1/2/3_joint_vel` (6): UR5e joint velocities
+  - `shoulder_pan/lift_joint_effort`, `elbow_joint_effort`, `wrist_1/2/3_joint_effort` (6): UR5e joint efforts
+  - `base_frame_ee_position_x/y/z` (3): End-effector position in base frame
+  - `base_frame_ee_orientation_x/y/z/w` (4): End-effector orientation quaternion in base frame
+  - `base_frame_initial_tip_position_x/y/z` (3): Initial tip position in base frame
+  - `base_frame_rcm_position_x/y/z` (3): RCM position in base frame
+  - `roll_angle` (1): Rotation around shaft axis
+
+**Actions**
+
+- `action`: control vector (4 dimensions for endoscope guidance, 5 dimensions for grasping and retaction)
+  - `delta_camera_frame_tip_position_x/y/z` (3): Delta tip position in camera frame
+  - `delta_roll_angle` (1): Delta roll angle around shaft axis
+  - `open_gripper`* (1): Gripper command
+
+**Metadata**
+
+- `timestamp`: Time in seconds
+- `episode_index`: Episode identifier
+- `frame_index`: Frame number within episode
+- `index`: Global frame index across all episodes
+- `task_index`: Task category (0-4)
+- `next.done`: Episode termination flag
+
+*only for grasping and retraction
+
+---
+
+## ⏱️ Data Synchronization Approach
+
+The clocks of all the involved machines and therefore the data timestamps were synchronized using the Precision Time Protocol (PTP).
+The gripper state (open/closed) was annotated manually in post-processing using the video.
+
+---
+
+## 👥 Attribution & Contact
+
+| | |
+| :--- | :--- |
+| **Dataset Lead (Principal Investigators)** | `[Stefanie Speidel, Martin Wagner]` |
+| **Dataset Lead (MD/PhD)** | `[Rayan Younis, Ariel Rodriguez, Lorenzo Mazza, Ortrun Hellig]` |
+| **Institution** | `[Department for Translational Surgical Oncology, National Center for Tumor Diseases, Partner Site Dresden (NCT/UCC): German Cancer Research Center (DKFZ), Helmholtz-Zentrum Dresden - Rossendorf (HZDR); Department of Visceral, Thoracic and Vascular Surgery, Faculty of Medicine and University Hospital Carl Gustav Carus, TUD Dresden University of Technology; Centre for the Tactile Internet with Human-in-the-Loop (CeTI), TUD Dresden University of Technology]` |
+| **Contact Email** | `[stefanie.speidel@nct-dresden.de, martin.wagner@ukdd.de]` |
+| **Acknowledgements** | `[Funded by the German Research Foundation (DFG, Deutsche Forschungsgemeinschaft) as part of Germany’s Excellence Strategy – EXC 2050/2 – Project ID 390696704 – Cluster of Excellence “Centre for Tactile Internet with Human-in-the-Loop” (CeTI) of TUD Dresden University of Technology. The authors acknowledge the financial support by the Federal Ministry of Research, Technology and Space of Germany in the programme of “Souverän. Digital. Vernetzt.”. Joint project 6G-life, project identification number: 16KISK001K. We thank the team of the Experimental Operating Room at the NCT Dresden for their support.]` |
+| **Citation (BibTeX)** | <pre><code>@misc{[TUNDRA_dataset_2026],<br>  author = {[Rodriguez, Ariel and Younis, Rayan and Mazza, Lorenzo and Hellig, Ortrun and Wagner, Martin and Speidel, Stefanie]},<br>  title = {[TUD & NCT Dresden Robot Assistant Dataset]},<br>  year = {2026},<br>  publisher = {Open-H-Embodiment},<br>  note = {}<br>}</code></pre> |

Surgical/ubc/knottying_merged/README.md

@@ -0,0 +1,185 @@
+<!--
+Open-H Embodiment Dataset README Template (v1.0)
+Please fill out this template and include it in the ./metadata directory of your LeRobot dataset.
+This file helps others understand the context and details of your contribution.
+-->
+
+# [UBC - Knot Tying] - README
+
+---
+
+## 📋 At a Glance
+
+*Teleoperated demonstrations of a da Vinci robot performing suturing with intracorporeal knot on a penrose drain.*
+
+---
+
+## 📖 Dataset Overview
+
+*This dataset contains 523 trajectories of expert surgeons using the da Vinci Si to perform surgical knot tying tasks. It includes successful trials, failures, and recovery attempts to provide a robust dataset for training imitation learning policies.*
+
+| | |
+| :--- | :--- |
+| **Total Trajectories** | `523` |
+| **Total Hours** | `0.71` |
+| **Data Type** | `Table-Top Phantom` |
+| **License** | MIT |
+
+---
+
+## 🎯 Tasks & Domain
+
+### Domain
+
+- [✅] **Surgical Robotics**
+- [ ] **Ultrasound Robotics**
+- [ ] **Other Healthcare Robotics** (Please specify: `[Your Domain]`)
+
+### Demonstrated Skills
+
+*Knot Tying*
+
+---
+
+## 🔬 Data Collection Details
+
+### Collection Method
+
+- [✅] **Human Teleoperation**
+- [ ] **Programmatic/State-Machine**
+- [ ] **AI Policy / Autonomous**
+- [ ] **Other** (Please specify: `[Your Method]`)
+
+### Operator Details
+
+| | Description |
+| :--- | :--- |
+| **Operator Count** | `6` |
+| **Operator Skill Level** | `3 Expert (e.g., Surgeon, Sonographer)` <br> `2 Intermediate (e.g., Trained Researcher)` <br> `1 Novice (e.g., ML Researcher with minimal experience)` |
+| **Collection Period** | From `[2025-11-15]` to `[2026-02-03]` |
+
+### Recovery Demonstrations
+
+- [ ] **Yes**
+- [✅] **No**
+
+---
+
+## 💡 Diversity Dimensions
+
+- [✅] **Camera Position / Angle**
+- [ ] **Lighting Conditions**
+- [ ] **Target Object** (e.g., different phantom models, suture types)
+- [✅] **Spatial Layout** (e.g., placing the target suture needle in various locations)
+- [ ] **Robot Embodiment** (if multiple robots were used)
+- [ ] **Task Execution** (e.g., different techniques for the same task)
+- [ ] **Background / Scene**
+- [ ] **Other** (Please specify: `[Your Dimension]`)
+
+---
+
+## 🛠️ Equipment & Setup
+
+### Robotic Platform(s)
+
+- **Robot 1:**  `da Vinci Si`
+
+### Sensors & Cameras
+
+| Type | Model/Details |
+| :--- | :--- |
+| **Primary Camera** | `Stereo Endoscopic Camera, 640x480 @ 30fps` |
+| **Room/3rd Person Camera** | `Intel RealSense D435i, 640x480 @ 30fps` |
+| **Other** | `Eye Gaze Tracker Made by RCL@UBC, 640x480 @ 30fps` |
+
+---
+
+## 🎯 Action & State Space Representation
+
+*We use ISI dVAPI recording robotic action and state. The data is following a research agreement between Intuitive Surgical and UBC RCL. We can only release the data when Intuitive Surgical allows us to do so.*
+
+### Action Space Representation
+
+**Primary Action Representation:**
+- [ ] **Absolute Cartesian** (position/orientation relative to robot base)
+- [ ] **Relative Cartesian** (delta position/orientation from current pose)
+- [ ] **Joint Space** (direct joint angle commands)
+- [ ] **Other** (Please specify: `[Your Representation]`)
+
+**Orientation Representation:**
+- [ ] **Quaternions** (x, y, z, w)
+- [ ] **Euler Angles** (roll, pitch, yaw)
+- [ ] **Axis-Angle** (rotation vector)
+- [ ] **Rotation Matrix** (3x3 matrix)
+- [ ] **Other** (Please specify: `[Your Representation]`)
+
+**Reference Frame:**
+- [ ] **Robot Base Frame**
+- [ ] **Tool/End-Effector Frame**
+- [ ] **World/Global Frame**
+- [ ] **Camera Frame**
+- [ ] **Other** (Please specify: `[Your Frame]`)
+
+**Action Dimensions:**
+*List the action space dimensions and their meanings.*
+
+**Example:**
+```
+action: [x, y, z, qx, qy, qz, qw, gripper]
+- x, y, z: Absolute position in robot base frame (meters)
+- qx, qy, qz, qw: Absolute orientation as quaternion
+- gripper: Gripper opening angle (radians)
+```
+
+### State Space Representation
+
+**State Information Included:**
+- [ ] **Joint Positions** (all articulated joints)
+- [ ] **Joint Velocities**
+- [ ] **End-Effector Pose** (Cartesian position/orientation)
+- [ ] **Force/Torque Readings**
+- [ ] **Gripper State** (position, force, etc.)
+- [ ] **Other** (Please specify: `[Your State Info]`)
+
+**State Dimensions:**
+*List the state space dimensions and their meanings.*
+
+**Example:**
+```
+observation.state: [j1, j2, j3, j4, j5, j6, j7, gripper_pos]
+- j1-j7: Absolute joint positions for 7-DOF arm (radians)
+- gripper_pos: Current gripper opening (meters)
+```
+
+### 📋 Recommended Additional Representations
+
+*Even if not your primary action/state representation, we strongly encourage including these standardized formats for maximum compatibility:*
+
+**Recommended Action Fields:**
+- **`action.cartesian_absolute`**: Absolute Cartesian pose with absolute quaternions
+  ```
+  [x, y, z, qx, qy, qz, qw, gripper_angle]
+  ```
+
+**Recommended State Fields:**
+- **`observation.state.joint_positions`**: Absolute positions for all articulated joints
+  ```
+  [joint_1, joint_2, ..., joint_n]
+  ```
+
+
+---
+
+## ⏱️ Data Synchronization Approach
+
+*We record timestamps for all sensors' data. We develop a software to synchronize and annotate the data. You can find it [here](https://github.com/UBCRCL-Surgery/Surgical-Robot-Data-Processing).*
+
+---
+
+## 👥 Attribution & Contact
+
+| | |
+| :--- | :--- |
+| **Dataset Lead** | `Zijian Wu` |
+| **Institution** | `University of British Columbia` |
+| **Contact Email** | `zijianwu@ece.ubc.ca` |

Surgical/ubc/needlepassing_merged/README.md

@@ -0,0 +1,185 @@
+<!--
+Open-H Embodiment Dataset README Template (v1.0)
+Please fill out this template and include it in the ./metadata directory of your LeRobot dataset.
+This file helps others understand the context and details of your contribution.
+-->
+
+# [UBC - Needle Passing] - README
+
+---
+
+## 📋 At a Glance
+
+*Teleoperated demonstrations of a da Vinci robot performing needle passing on a training phantom with little metal rings.*
+
+---
+
+## 📖 Dataset Overview
+
+*This dataset contains 685 trajectories of expert surgeons using the da Vinci Si to perform surgical suturing tasks. It includes successful trials, failures, and recovery attempts to provide a robust dataset for training imitation learning policies.*
+
+| | |
+| :--- | :--- |
+| **Total Trajectories** | `685` |
+| **Total Hours** | `0.65` |
+| **Data Type** | `Table-Top Phantom` |
+| **License** | MIT |
+
+---
+
+## 🎯 Tasks & Domain
+
+### Domain
+
+- [✅] **Surgical Robotics**
+- [ ] **Ultrasound Robotics**
+- [ ] **Other Healthcare Robotics** (Please specify: `[Your Domain]`)
+
+### Demonstrated Skills
+
+*Needle Passing*
+
+---
+
+## 🔬 Data Collection Details
+
+### Collection Method
+
+- [✅] **Human Teleoperation**
+- [ ] **Programmatic/State-Machine**
+- [ ] **AI Policy / Autonomous**
+- [ ] **Other** (Please specify: `[Your Method]`)
+
+### Operator Details
+
+| | Description |
+| :--- | :--- |
+| **Operator Count** | `7` |
+| **Operator Skill Level** | `4 Expert (e.g., Surgeon, Sonographer)` <br> `2 Intermediate (e.g., Trained Researcher)` <br> `1 Novice (e.g., ML Researcher with minimal experience)` |
+| **Collection Period** | From `[2025-11-15]` to `[2026-02-03]` |
+
+### Recovery Demonstrations
+
+- [ ] **Yes**
+- [✅] **No**
+
+---
+
+## 💡 Diversity Dimensions
+
+- [✅] **Camera Position / Angle**
+- [ ] **Lighting Conditions**
+- [ ] **Target Object** (e.g., different phantom models, suture types)
+- [✅] **Spatial Layout** (e.g., placing the target suture needle in various locations)
+- [ ] **Robot Embodiment** (if multiple robots were used)
+- [ ] **Task Execution** (e.g., different techniques for the same task)
+- [ ] **Background / Scene**
+- [ ] **Other** (Please specify: `[Your Dimension]`)
+
+---
+
+## 🛠️ Equipment & Setup
+
+### Robotic Platform(s)
+
+- **Robot 1:**  `da Vinci Si`
+
+### Sensors & Cameras
+
+| Type | Model/Details |
+| :--- | :--- |
+| **Primary Camera** | `Stereo Endoscopic Camera, 640x480 @ 30fps` |
+| **Room/3rd Person Camera** | `Intel RealSense D435i, 640x480 @ 30fps` |
+| **Other** | `Eye Gaze Tracker Made by RCL@UBC, 640x480 @ 30fps` |
+
+---
+
+## 🎯 Action & State Space Representation
+
+*We use ISI dVAPI recording robotic action and state. The data is following a research agreement between Intuitive Surgical and UBC RCL. We can only release the data when Intuitive Surgical allows us to do so.*
+
+### Action Space Representation
+
+**Primary Action Representation:**
+- [ ] **Absolute Cartesian** (position/orientation relative to robot base)
+- [ ] **Relative Cartesian** (delta position/orientation from current pose)
+- [ ] **Joint Space** (direct joint angle commands)
+- [ ] **Other** (Please specify: `[Your Representation]`)
+
+**Orientation Representation:**
+- [ ] **Quaternions** (x, y, z, w)
+- [ ] **Euler Angles** (roll, pitch, yaw)
+- [ ] **Axis-Angle** (rotation vector)
+- [ ] **Rotation Matrix** (3x3 matrix)
+- [ ] **Other** (Please specify: `[Your Representation]`)
+
+**Reference Frame:**
+- [ ] **Robot Base Frame**
+- [ ] **Tool/End-Effector Frame**
+- [ ] **World/Global Frame**
+- [ ] **Camera Frame**
+- [ ] **Other** (Please specify: `[Your Frame]`)
+
+**Action Dimensions:**
+*List the action space dimensions and their meanings.*
+
+**Example:**
+```
+action: [x, y, z, qx, qy, qz, qw, gripper]
+- x, y, z: Absolute position in robot base frame (meters)
+- qx, qy, qz, qw: Absolute orientation as quaternion
+- gripper: Gripper opening angle (radians)
+```
+
+### State Space Representation
+
+**State Information Included:**
+- [ ] **Joint Positions** (all articulated joints)
+- [ ] **Joint Velocities**
+- [ ] **End-Effector Pose** (Cartesian position/orientation)
+- [ ] **Force/Torque Readings**
+- [ ] **Gripper State** (position, force, etc.)
+- [ ] **Other** (Please specify: `[Your State Info]`)
+
+**State Dimensions:**
+*List the state space dimensions and their meanings.*
+
+**Example:**
+```
+observation.state: [j1, j2, j3, j4, j5, j6, j7, gripper_pos]
+- j1-j7: Absolute joint positions for 7-DOF arm (radians)
+- gripper_pos: Current gripper opening (meters)
+```
+
+### 📋 Recommended Additional Representations
+
+*Even if not your primary action/state representation, we strongly encourage including these standardized formats for maximum compatibility:*
+
+**Recommended Action Fields:**
+- **`action.cartesian_absolute`**: Absolute Cartesian pose with absolute quaternions
+  ```
+  [x, y, z, qx, qy, qz, qw, gripper_angle]
+  ```
+
+**Recommended State Fields:**
+- **`observation.state.joint_positions`**: Absolute positions for all articulated joints
+  ```
+  [joint_1, joint_2, ..., joint_n]
+  ```
+
+
+---
+
+## ⏱️ Data Synchronization Approach
+
+*We record timestamps for all sensors' data. We develop a software to synchronize and annotate the data. You can find it [here](https://github.com/UBCRCL-Surgery/Surgical-Robot-Data-Processing).*
+
+---
+
+## 👥 Attribution & Contact
+
+| | |
+| :--- | :--- |
+| **Dataset Lead** | `Zijian Wu` |
+| **Institution** | `University of British Columbia` |
+| **Contact Email** | `zijianwu@ece.ubc.ca` |

Surgical/ubc/pickandplace_merged/README.md

@@ -0,0 +1,185 @@
+<!--
+Open-H Embodiment Dataset README Template (v1.0)
+Please fill out this template and include it in the ./metadata directory of your LeRobot dataset.
+This file helps others understand the context and details of your contribution.
+-->
+
+# [UBC - Pick and Place] - README
+
+---
+
+## 📋 At a Glance
+
+*Teleoperated demonstrations of a da Vinci robot performing pick and place tasks on a training phantom.*
+
+---
+
+## 📖 Dataset Overview
+
+*This dataset contains 702 trajectories of expert surgeons using the da Vinci Si to perform pick and place tasks. It includes successful trials, failures, and recovery attempts to provide a robust dataset for training imitation learning policies.*
+
+| | |
+| :--- | :--- |
+| **Total Trajectories** | `702` |
+| **Total Hours** | `0.33` |
+| **Data Type** | `Table-Top Phantom` |
+| **License** | MIT |
+
+---
+
+## 🎯 Tasks & Domain
+
+### Domain
+
+- [✅] **Surgical Robotics**
+- [ ] **Ultrasound Robotics**
+- [ ] **Other Healthcare Robotics** (Please specify: `[Your Domain]`)
+
+### Demonstrated Skills
+
+*Pick and Place*
+
+---
+
+## 🔬 Data Collection Details
+
+### Collection Method
+
+- [✅] **Human Teleoperation**
+- [ ] **Programmatic/State-Machine**
+- [ ] **AI Policy / Autonomous**
+- [ ] **Other** (Please specify: `[Your Method]`)
+
+### Operator Details
+
+| | Description |
+| :--- | :--- |
+| **Operator Count** | `5` |
+| **Operator Skill Level** | `3 Expert (e.g., Surgeon, Sonographer)` <br> `1 Intermediate (e.g., Trained Researcher)` <br> `1 Novice (e.g., ML Researcher with minimal experience)` |
+| **Collection Period** | From `[2025-11-15]` to `[2026-02-03]` |
+
+### Recovery Demonstrations
+
+- [ ] **Yes**
+- [✅] **No**
+
+---
+
+## 💡 Diversity Dimensions
+
+- [✅] **Camera Position / Angle**
+- [ ] **Lighting Conditions**
+- [ ] **Target Object** (e.g., different phantom models, suture types)
+- [✅] **Spatial Layout** (e.g., placing the target suture needle in various locations)
+- [ ] **Robot Embodiment** (if multiple robots were used)
+- [ ] **Task Execution** (e.g., different techniques for the same task)
+- [ ] **Background / Scene**
+- [ ] **Other** (Please specify: `[Your Dimension]`)
+
+---
+
+## 🛠️ Equipment & Setup
+
+### Robotic Platform(s)
+
+- **Robot 1:**  `da Vinci Si`
+
+### Sensors & Cameras
+
+| Type | Model/Details |
+| :--- | :--- |
+| **Primary Camera** | `Stereo Endoscopic Camera, 640x480 @ 30fps` |
+| **Room/3rd Person Camera** | `Intel RealSense D435i, 640x480 @ 30fps` |
+| **Other** | `Eye Gaze Tracker Made by RCL@UBC, 640x480 @ 30fps` |
+
+---
+
+## 🎯 Action & State Space Representation
+
+*We use ISI dVAPI recording robotic action and state. The data is following a research agreement between Intuitive Surgical and UBC RCL. We can only release the data when Intuitive Surgical allows us to do so.*
+
+### Action Space Representation
+
+**Primary Action Representation:**
+- [ ] **Absolute Cartesian** (position/orientation relative to robot base)
+- [ ] **Relative Cartesian** (delta position/orientation from current pose)
+- [ ] **Joint Space** (direct joint angle commands)
+- [ ] **Other** (Please specify: `[Your Representation]`)
+
+**Orientation Representation:**
+- [ ] **Quaternions** (x, y, z, w)
+- [ ] **Euler Angles** (roll, pitch, yaw)
+- [ ] **Axis-Angle** (rotation vector)
+- [ ] **Rotation Matrix** (3x3 matrix)
+- [ ] **Other** (Please specify: `[Your Representation]`)
+
+**Reference Frame:**
+- [ ] **Robot Base Frame**
+- [ ] **Tool/End-Effector Frame**
+- [ ] **World/Global Frame**
+- [ ] **Camera Frame**
+- [ ] **Other** (Please specify: `[Your Frame]`)
+
+**Action Dimensions:**
+*List the action space dimensions and their meanings.*
+
+**Example:**
+```
+action: [x, y, z, qx, qy, qz, qw, gripper]
+- x, y, z: Absolute position in robot base frame (meters)
+- qx, qy, qz, qw: Absolute orientation as quaternion
+- gripper: Gripper opening angle (radians)
+```
+
+### State Space Representation
+
+**State Information Included:**
+- [ ] **Joint Positions** (all articulated joints)
+- [ ] **Joint Velocities**
+- [ ] **End-Effector Pose** (Cartesian position/orientation)
+- [ ] **Force/Torque Readings**
+- [ ] **Gripper State** (position, force, etc.)
+- [ ] **Other** (Please specify: `[Your State Info]`)
+
+**State Dimensions:**
+*List the state space dimensions and their meanings.*
+
+**Example:**
+```
+observation.state: [j1, j2, j3, j4, j5, j6, j7, gripper_pos]
+- j1-j7: Absolute joint positions for 7-DOF arm (radians)
+- gripper_pos: Current gripper opening (meters)
+```
+
+### 📋 Recommended Additional Representations
+
+*Even if not your primary action/state representation, we strongly encourage including these standardized formats for maximum compatibility:*
+
+**Recommended Action Fields:**
+- **`action.cartesian_absolute`**: Absolute Cartesian pose with absolute quaternions
+  ```
+  [x, y, z, qx, qy, qz, qw, gripper_angle]
+  ```
+
+**Recommended State Fields:**
+- **`observation.state.joint_positions`**: Absolute positions for all articulated joints
+  ```
+  [joint_1, joint_2, ..., joint_n]
+  ```
+
+
+---
+
+## ⏱️ Data Synchronization Approach
+
+*We record timestamps for all sensors' data. We develop a software to synchronize and annotate the data. You can find it [here](https://github.com/UBCRCL-Surgery/Surgical-Robot-Data-Processing).*
+
+---
+
+## 👥 Attribution & Contact
+
+| | |
+| :--- | :--- |
+| **Dataset Lead** | `Zijian Wu` |
+| **Institution** | `University of British Columbia` |
+| **Contact Email** | `zijianwu@ece.ubc.ca` |

Surgical/ubc/wirechasing_merged/README.md

@@ -0,0 +1,185 @@
+<!--
+Open-H Embodiment Dataset README Template (v1.0)
+Please fill out this template and include it in the ./metadata directory of your LeRobot dataset.
+This file helps others understand the context and details of your contribution.
+-->
+
+# [UBC - Wire Chasing] - README
+
+---
+
+## 📋 At a Glance
+
+*Teleoperated demonstrations of a da Vinci robot performing the 2D wire chasing task on a standard wire chaser board.*
+
+---
+
+## 📖 Dataset Overview
+
+*This dataset contains 615 trajectories of expert surgeons using the da Vinci Si to perform 2D wire chasing tasks. It includes successful trials, failures, and recovery attempts to provide a robust dataset for training imitation learning policies.*
+
+| | |
+| :--- | :--- |
+| **Total Trajectories** | `615` |
+| **Total Hours** | `0.55` |
+| **Data Type** | `Table-Top Phantom` |
+| **License** | MIT |
+
+---
+
+## 🎯 Tasks & Domain
+
+### Domain
+
+- [✅] **Surgical Robotics**
+- [ ] **Ultrasound Robotics**
+- [ ] **Other Healthcare Robotics** (Please specify: `[Your Domain]`)
+
+### Demonstrated Skills
+
+*2D Wire Chasing*
+
+---
+
+## 🔬 Data Collection Details
+
+### Collection Method
+
+- [✅] **Human Teleoperation**
+- [ ] **Programmatic/State-Machine**
+- [ ] **AI Policy / Autonomous**
+- [ ] **Other** (Please specify: `[Your Method]`)
+
+### Operator Details
+
+| | Description |
+| :--- | :--- |
+| **Operator Count** | `6` |
+| **Operator Skill Level** | `3 Expert (e.g., Surgeon, Sonographer)` <br> `2 Intermediate (e.g., Trained Researcher)` <br> `1 Novice (e.g., ML Researcher with minimal experience)` |
+| **Collection Period** | From `[2025-11-15]` to `[2026-02-03]` |
+
+### Recovery Demonstrations
+
+- [ ] **Yes**
+- [✅] **No**
+
+---
+
+## 💡 Diversity Dimensions
+
+- [✅] **Camera Position / Angle**
+- [ ] **Lighting Conditions**
+- [ ] **Target Object** (e.g., different phantom models, suture types)
+- [✅] **Spatial Layout** (e.g., placing the target suture needle in various locations)
+- [ ] **Robot Embodiment** (if multiple robots were used)
+- [ ] **Task Execution** (e.g., different techniques for the same task)
+- [ ] **Background / Scene**
+- [ ] **Other** (Please specify: `[Your Dimension]`)
+
+---
+
+## 🛠️ Equipment & Setup
+
+### Robotic Platform(s)
+
+- **Robot 1:**  `da Vinci Si`
+
+### Sensors & Cameras
+
+| Type | Model/Details |
+| :--- | :--- |
+| **Primary Camera** | `Stereo Endoscopic Camera, 640x480 @ 30fps` |
+| **Room/3rd Person Camera** | `Intel RealSense D435i, 640x480 @ 30fps` |
+| **Other** | `Eye Gaze Tracker Made by RCL@UBC, 640x480 @ 30fps` |
+
+---
+
+## 🎯 Action & State Space Representation
+
+*We use ISI dVAPI recording robotic action and state. The data is following a research agreement between Intuitive Surgical and UBC RCL. We can only release the data when Intuitive Surgical allows us to do so.*
+
+### Action Space Representation
+
+**Primary Action Representation:**
+- [ ] **Absolute Cartesian** (position/orientation relative to robot base)
+- [ ] **Relative Cartesian** (delta position/orientation from current pose)
+- [ ] **Joint Space** (direct joint angle commands)
+- [ ] **Other** (Please specify: `[Your Representation]`)
+
+**Orientation Representation:**
+- [ ] **Quaternions** (x, y, z, w)
+- [ ] **Euler Angles** (roll, pitch, yaw)
+- [ ] **Axis-Angle** (rotation vector)
+- [ ] **Rotation Matrix** (3x3 matrix)
+- [ ] **Other** (Please specify: `[Your Representation]`)
+
+**Reference Frame:**
+- [ ] **Robot Base Frame**
+- [ ] **Tool/End-Effector Frame**
+- [ ] **World/Global Frame**
+- [ ] **Camera Frame**
+- [ ] **Other** (Please specify: `[Your Frame]`)
+
+**Action Dimensions:**
+*List the action space dimensions and their meanings.*
+
+**Example:**
+```
+action: [x, y, z, qx, qy, qz, qw, gripper]
+- x, y, z: Absolute position in robot base frame (meters)
+- qx, qy, qz, qw: Absolute orientation as quaternion
+- gripper: Gripper opening angle (radians)
+```
+
+### State Space Representation
+
+**State Information Included:**
+- [ ] **Joint Positions** (all articulated joints)
+- [ ] **Joint Velocities**
+- [ ] **End-Effector Pose** (Cartesian position/orientation)
+- [ ] **Force/Torque Readings**
+- [ ] **Gripper State** (position, force, etc.)
+- [ ] **Other** (Please specify: `[Your State Info]`)
+
+**State Dimensions:**
+*List the state space dimensions and their meanings.*
+
+**Example:**
+```
+observation.state: [j1, j2, j3, j4, j5, j6, j7, gripper_pos]
+- j1-j7: Absolute joint positions for 7-DOF arm (radians)
+- gripper_pos: Current gripper opening (meters)
+```
+
+### 📋 Recommended Additional Representations
+
+*Even if not your primary action/state representation, we strongly encourage including these standardized formats for maximum compatibility:*
+
+**Recommended Action Fields:**
+- **`action.cartesian_absolute`**: Absolute Cartesian pose with absolute quaternions
+  ```
+  [x, y, z, qx, qy, qz, qw, gripper_angle]
+  ```
+
+**Recommended State Fields:**
+- **`observation.state.joint_positions`**: Absolute positions for all articulated joints
+  ```
+  [joint_1, joint_2, ..., joint_n]
+  ```
+
+
+---
+
+## ⏱️ Data Synchronization Approach
+
+*We record timestamps for all sensors' data. We develop a software to synchronize and annotate the data. You can find it [here](https://github.com/UBCRCL-Surgery/Surgical-Robot-Data-Processing).*
+
+---
+
+## 👥 Attribution & Contact
+
+| | |
+| :--- | :--- |
+| **Dataset Lead** | `Zijian Wu` |
+| **Institution** | `University of British Columbia` |
+| **Contact Email** | `zijianwu@ece.ubc.ca` |

Surgical/uic/uic_crcd_lerobot/README.md

@@ -0,0 +1,155 @@
+---
+license: cc-by-4.0
+task_categories:
+- robotics
+- reinforcement-learning
+configs:
+- config_name: default
+  data_files:
+  - split: train
+    path: data/*/train-*
+dataset_info:
+  features:
+  - name: video_id
+    dtype: string
+  - name: frame_n
+    dtype: int64
+  - name: timestamp
+    dtype: float64
+  - name: frame_left
+    dtype: image
+  - name: frame_right
+    dtype: image
+  - name: /ECM/custom/local/setpoint_cp
+    sequence: float64
+  - name: /ECM/custom/setpoint_cp
+    sequence: float64
+  - name: /ECM/measured_js
+    sequence: float64
+  - name: /MTML/gripper/measured_js
+    sequence: float64
+  - name: /MTML/local/measured_cp
+    sequence: float64
+  - name: /MTML/measured_cp
+    sequence: float64
+  - name: /MTML/measured_js
+    sequence: float64
+  - name: /MTMR/gripper/measured_js
+    sequence: float64
+  - name: /MTMR/local/measured_cp
+    sequence: float64
+  - name: /MTMR/measured_cp
+    sequence: float64
+  - name: /MTMR/measured_js
+    sequence: float64
+  - name: /PSM1/custom/local/setpoint_cp
+    sequence: float64
+  - name: /PSM1/custom/setpoint_cp
+    sequence: float64
+  - name: /PSM1/jaw/measured_js
+    sequence: float64
+  - name: /PSM1/measured_js
+    sequence: float64
+  - name: /PSM2/custom/local/setpoint_cp
+    sequence: float64
+  - name: /PSM2/custom/setpoint_cp
+    sequence: float64
+  - name: /PSM2/jaw/measured_js
+    sequence: float64
+  - name: /PSM2/measured_js
+    sequence: float64
+  - name: /pedals/camera
+    dtype: bool
+  - name: /pedals/clutch
+    dtype: bool
+  - name: /pedals/monopolar
+    dtype: bool
+  splits:
+  - name: train
+    num_bytes: 78489990786
+    num_examples: 755891
+    shard_lengths: [4776, 5475, 5707, 5032, 4032, 3700, 5077, 4546, 4046, 3546, 3692, 3100, 3046, 3445, 3945, 3345, 2945, 3045, 3045, 4545, 5104, 4202, 3902, 3702, 3302, 3200, 3902, 4493, 4091, 3190, 3190, 2990, 3000, 3580, 3890, 3890, 3490, 3490, 4290, 5362, 4262, 4362, 3962, 4062, 3961, 4061, 3861, 3661, 4061, 4161, 3861, 4261, 4635, 4417, 3917, 4134, 3517, 3317, 3017, 3217, 3017, 3117, 2917, 2900, 2717, 2617, 3317, 3217, 9718, 9601, 7901, 8901, 11700, 11758, 12358, 6168, 4211, 3611, 3610, 4910, 3610, 3010, 3310, 4110, 4710, 3210, 3210, 4010, 5138, 4638, 3438, 3238, 4843, 5142, 5042, 4442, 3642, 4142, 6698, 5956, 4756, 4156, 4456, 5456, 4956, 4956, 4456, 4956, 6549, 6493, 5993, 4693, 4893, 5193, 5093, 4593, 3893, 4093, 3493, 3400, 6548, 6956, 6456, 6956, 6656, 5856, 5856, 5055, 5911, 5211, 5411, 5221, 5210, 5710, 4010, 4900, 7360, 5350, 6650, 6550, 6050, 6350, 7050, 5950, 6250, 5650, 6950, 4750, 7149, 7618, 8217, 7917, 7517, 6117, 6517, 2617]
+---
+
+# Comprehensive Robotic Cholecystectomy Dataset (CRCD)
+
+The **Comprehensive Robotic Cholecystectomy Dataset (CRCD)** is a large-scale, multimodal dataset for **robot-assisted surgery (RAS)** research.  
+It provides synchronized **endoscopic videos, da Vinci surgical robot kinematics, and pedal usage signals**, making it one of the most comprehensive open datasets for studying **robotic cholecystectomy procedures**.  
+
+CRCD supports research in:
+- Medical robotics and surgical automation  
+- Computer vision for endoscopic surgery  
+- Surgical workflow analysis and phase recognition  
+- Instrument tracking and tissue segmentation  
+- AI and machine learning in healthcare  
+
+---
+
+## Updates
+
+- Sep 30, 2025: Corrected video timestamps for improved synchronization and cleaned up the dataset for easier use.
+
+
+---
+
+## Dataset Info
+
+- **Curated by:** Ki-Hwan Oh, Leonardo Borgioli, Alberto Mangano, Valentina Valle, Marco Di Pangrazio, Francesco Toti, Gioia Pozza, Luciano Ambrosini, Alvaro Ducas, Miloš Žefran, Liaohai Chen, Pier Cristoforo Giulianotti  
+
+---
+
+## Dataset Sources
+
+<table align="center">
+  <tr>
+    <td><a href="https://github.com/sitleng/CRCD"><img src="https://img.shields.io/badge/GitHub-Repo-blue?logo=github"/></a><br/><b>GitHub</b></td>
+    <td><a href="https://arxiv.org/abs/2412.12238"><img src="https://img.shields.io/badge/arXiv-Paper-red?logo=arxiv"/></a><br/><b>Journal (Expanded)</b></td>
+    <td><a href="https://ieeexplore.ieee.org/abstract/document/10585836"><img src="https://img.shields.io/badge/IEEE-Paper-blue?logo=ieee"/></a><br/><b>Conference Paper</b></td>
+  </tr>
+</table>
+
+- **Raw Dataset:** Endoscopic videos, da Vinci kinematics, and console pedal usage ([link](https://uofi.box.com/s/p3aocj6yzq4ctwc0s635a2dfyk9zdv5j))  
+- **Annotated Dataset:** Frames with annotated **tissue segmentation** and **instrument keypoints** ([link](https://uofi.box.com/s/f9bg69ve6fkwktr3o33ahmp620w8jth6))  
+- **Additional Information:** Stereo endoscopic camera calibration and surgeon background data ([link](https://uofi.box.com/s/w65rui5ylm0i4v4jvlkpacpi4q6jkdpe))  
+
+---
+
+## Dataset Creation
+
+CRCD was collected during **robotic cholecystectomy procedures** performed on the **da Vinci surgical system**.  
+Each case includes:
+- **High-resolution endoscopic video**  
+- **Robot kinematic data** (ECM, MTML, MTMR, PSM1, PSM2)  
+- **Surgeon pedal signals** (clutch, camera, monopolar, bipolar)  
+
+Several surgeons with different levels of expertise participated, enabling research on **skill assessment, workflow modeling, and training.**
+
+---
+
+### License
+
+This dataset is licensed under the **Creative Commons Attribution 4.0 International License**.
+
+---
+
+## Citation
+
+If you use CRCD, please cite:  
+
+```bibtex
+@INPROCEEDINGS{oh2024crcd,
+  author={Oh, Ki-Hwan and Borgioli, Leonardo and Mangano, Alberto and Valle, Valentina and Di Pangrazio, Marco and Toti, Francesco and Pozza, Gioia and Ambrosini, Luciano and Ducas, Alvaro and Žefran, Miloš and Chen, Liaohai and Giulianotti, Pier Cristoforo},
+  booktitle={2024 International Symposium on Medical Robotics (ISMR)}, 
+  title={Comprehensive Robotic Cholecystectomy Dataset (CRCD): Integrating Kinematics, Pedal Signals, and Endoscopic Videos}, 
+  year={2024},
+  pages={1-7},
+  doi={10.1109/ISMR63436.2024.10585836}
+}
+
+@article{oh2024crcdexpanded,
+  author = {Oh, Ki-Hwan and Borgioli, Leonardo and Mangano, Alberto and Valle, Valentina and Pangrazio, Marco Di and Toti, Francesco and Pozza, Gioia and Ambrosini, Luciano and Ducas, Alvaro and Žefran, Miloš and Chen, Liaohai and Giulianotti, Pier Cristoforo},
+  title = {Expanded Comprehensive Robotic Cholecystectomy Dataset},
+  journal = {Journal of Medical Robotics Research},
+  doi = {10.1142/S2424905X25500060},
+  URL = {https://doi.org/10.1142/S2424905X25500060}
+}