中文 | English
CLEAR Lab is building an open-source space for robotics, embodied intelligence, simulation, control, and learning-based robot systems.
This page collects our public research code, teaching resources, reproducible experiments, and selected community collaborations.
- Robotics control and learning
- Embodied AI and Vision-Language-Action models
- Simulation, evaluation, and reproducible experiments
- Teaching, tutorials, and research engineering tools
CLEAR Lab is working with Limx on selected open-source projects that connect academic research, engineering practice, reproducible tools, and community collaboration.
Projects in this initiative:
| Project | Status | Scope |
|---|---|---|
| FluxVLA-Tron2 | Experimental research release | FluxVLA / PI0.5 fine-tuning and remote-inference deployment workflows for Tron2 |
| World (Action) Models Tutorial | Tutorial release | From World Models to World Action Models: A Concise Tutorial for Robotics |
Below are selected public repositories from CLEAR Lab. For the full list, please visit the Repositories tab.
| Area | Repository | Scope |
|---|---|---|
| Robot learning resources | Awesome-Legged-Robot-Learning | Curated papers and resources for legged robot learning, reinforcement learning, sim-to-real, and humanoids |
| Wheel-legged locomotion | Wheel-Legged-Gym | Isaac Gym environments and training workflows for wheel-legged robot learning |
| Multi-embodiment locomotion | multi_loco_isaacgym | Unified policy training for TRON-1 pointed-foot and wheeled-foot modes using diffusion and reinforcement learning |
| Bipedal control | Bipedal_MPC | Model predictive control examples for TRON 1 in MuJoCo simulation |
| Task-space control | VF-TSC | Adaptive feedback gain methods for floating-base task-space control |
| Tutorials and teaching | MujocoTutorial | MuJoCo tutorial materials for simulation learning and practice |
| Course code | ACR2025 | Advanced Control for Robotics course code repository |
You can participate through issues, pull requests, documentation improvements, experiment reproduction, tutorials, and technical discussions.
For project-specific questions, please use the issue tracker of the corresponding repository. For community collaboration or broader open-source discussion, contact us by email.
- CLEAR Lab website: https://www.wzhanglab.site/
- CLEAR Lab GitHub: https://github.com/clearlab-sustech
- Limx GitHub: https://github.com/limxdynamics
- Contact: clearlab@sustech.edu.cn