File size: 4,565 Bytes
222da1d | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 | # Explainability Subcard
| Field | Response |
| ---------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| Intended Task/Domain: | Locomotion API for Unitree H2 Plus, enabling integration with upper-body controllers for whole-body teleoperation and loco-manipulation data collection. |
| Model Type: | Feed-forward PPO actor MLP exported as ONNX. |
| Intended Users: | Roboticists and researchers in academia and industry. |
| Output: | 14 leg-and-waist joint-position targets, proportional and derivative gains, and feedback state. |
| Describe how the model works: | The MLP maps a five-step history of robot proprioception, commands and previous actions to target joint positions. It was trained with PPO in Isaac Lab simulation. |
| List the groups (or group characteristics) for which this was tested to produce comparable outcomes: | Not Applicable |
| Technical Limitations & Mitigation: | The policy may fail on floor types outside its training domain, under external disturbances, or with highly dynamic upper-body controllers. Validate representative friction and disturbances, use independent safety controls, and expand upper-body motion gradually. |
| Verified to have met prescribed NVIDIA quality standards: | Yes |
| Performance Metrics: | Robustness, latency and control frequency are relevant evaluation dimensions. Final training telemetry reported 0.125 m/s XY velocity error and 0.313 rad/s yaw error. |
| Potential Known Risks: | This model is not tested or intended for use in mission critical applications that require functional safety. The use of the model in those applications is at the user's own risk and sole responsibility, including taking the necessary steps to add needed guardrails or safety mechanisms prior to deployment. This model was trained in simulation and evaluated in simulation and the real world. Physical deployment requires independent safety and behavior qualification. Incorrect joint mapping, stale feedback, out-of-range commands or policy failure can cause the robot to fall or make unexpected movements. |
| Licensing: | NVIDIA Open Model License ([license text](https://www.nvidia.com/en-us/agreements/enterprise-software/nvidia-open-model-license/)) |
|