# LeggedGym-Ex
A [legged_gym](https://github.com/leggedrobotics/legged_gym) based framework for training legged robots in [Genesis](https://github.com/Genesis-Embodied-AI/Genesis/tree/main), [IsaacGym](https://developer.nvidia.com/isaac-gym) and [IsaacSim](https://developer.nvidia.com/isaac/sim).
## Features
- **Totally based on [legged_gym](https://github.com/leggedrobotics/legged_gym)**
This framework keeps most APIs and conventions consistent with legged_gym, providing good readability and better control over training pipeline.
- **Integration of multiple simulators**
We support training in either of three simulators: IsaacGym, Genesis and IsaacSim.
Quick tips for choosing among three simulators:
- Faster training but worse rendering -> IsaacGym
- Both training speed and support for fluid, soft materials -> Genesis
- More realistic rendering at the cost of training speed -> IsaacSim.
- **Incorporation of various methods in published RL papers**
| Method | Paper Link | Code |
|--------|------------|----------|
| Periodic Gait Reward | [Sim-to-Real Learning of All Common Bipedal Gaits via Periodic Reward Composition](https://arxiv.org/abs/2011.01387) | [go2_wtw](https://github.com/lupinjia/LeggedGym-Ex/blob/main/legged_gym/envs/go2/go2_wtw/go2_wtw.py#L322) |
| Walk These Ways | [Walk These Ways: Tuning Robot Control for Generalization with Multiplicity of Behavior](https://gmargo11.github.io/walk-these-ways/) | [go2_wtw](https://github.com/lupinjia/LeggedGym-Ex/blob/main/legged_gym/envs/go2/go2_wtw) |
| System Identification | [Learning Agile Bipedal Motions on a Quadrupedal Robot](https://arxiv.org/abs/2311.05818) | [go2_sysid](https://github.com/lupinjia/LeggedGym-Ex/tree/main/legged_gym/envs/go2/go2_sysid) |
| Teacher-Student | [Rapid Locomotion via Reinforcement Learning](https://agility.csail.mit.edu/) | [go2_ts](https://github.com/lupinjia/LeggedGym-Ex/tree/main/legged_gym/envs/go2/go2_ts) |
| Explicit Estimator | [Concurrent Training of a Control Policy and a State Estimator for Dynamic and Robust Legged Locomotion](https://arxiv.org/abs/2202.05481) | [go2_ee](https://github.com/lupinjia/LeggedGym-Ex/tree/main/legged_gym/envs/go2/go2_ee) |
| Constraints as Terminations | [CaT: Constraints as Terminations for Legged Locomotion Reinforcement Learning](https://constraints-as-terminations.github.io/) | [go2_cat](https://github.com/lupinjia/LeggedGym-Ex/tree/main/legged_gym/envs/go2/go2_cat) |
| DreamWaQ | [DreamWaQ: Learning Robust Quadrupedal Locomotion With Implicit Terrain Imagination via Deep Reinforcement Learning](https://arxiv.org/abs/2301.10602) | [go2_dreamwaq](https://github.com/lupinjia/LeggedGym-Ex/tree/main/legged_gym/envs/go2/go2_dreamwaq) |
| SPO (Simple Policy Optimization) | [Simple Policy Optimization](https://github.com/MyRepositories-hub/Simple-Policy-Optimization) | [`legged_robot_config.py`](https://github.com/lupinjia/LeggedGym-Ex/tree/main/legged_gym/envs/base/legged_robot_config.py) |
| CTS (Concurrent Teacher Student) | [CTS: Concurrent Teacher-Student Reinforcement Learning for Legged Locomotion](https://clearlab-sustech.github.io/concurrentTS/) | [go2_cts](https://github.com/lupinjia/LeggedGym-Ex/tree/main/legged_gym/envs/go2/go2_cts) |
| DeepMimic | [DeepMimic: Example-Guided Deep Reinforcement Learning of Physics-Based Character Skills](https://arxiv.org/abs/1804.02717) | [g1_deepmimic](https://github.com/lupinjia/LeggedGym-Ex/tree/main/legged_gym/envs/g1/g1_deepmimic) |
| AMP (Adversarial Motion Priors) | [AMP: Adversarial Motion Priors for Stylized Physics-Based Character Control](https://arxiv.org/abs/2104.02180) | [k1_amp](https://github.com/lupinjia/LeggedGym-Ex/tree/main/legged_gym/envs/k1/k1_amp) |
## Gallery
| Robot | Sim | Real |
|--- | --- | --- |
| Unitree Go2 | 
| [video](https://www.bilibili.com/video/BV1FPedzZEdi/) |
| TRON1_PF | 
| [video](https://www.bilibili.com/video/BV1MdePzcEvk/) |
| TRON1_SF | 
| |
| Unitree G1 DeepMimic | 
| |
| Booster K1 | 
| [video](https://www.bilibili.com/video/BV1GyXgBmEa9/) |
| Go2 Parkour with Depth Image | 
| [video](https://www.bilibili.com/video/BV1riwMzmEZT/?spm_id_from=333.1387.homepage.video_card.click) |
## Acknowledgements
- [Genesis](https://github.com/Genesis-Embodied-AI/Genesis/tree/main)
- [Genesis-backflip](https://github.com/ziyanx02/Genesis-backflip)
- [legged_gym](https://github.com/leggedrobotics/legged_gym)
- [rsl_rl](https://github.com/leggedrobotics/rsl_rl)
- [unitree_rl_gym](https://github.com/unitreerobotics/unitree_rl_gym)
- [tron1-rl-isaacgym](https://github.com/limxdynamics/tron1-rl-isaacgym)
- [isaaclab](https://github.com/isaac-sim/IsaacLab)
## TODO
- [x] Add support for TRON1_SF (2026/02/13)
- [x] Add support for IsaacSim simulator (2026/02/15)
- [x] Add support for DeepMimic Implementation (2026/02/28)
- [x] Add support for Booster K1 Robot (2026/03/18)
- [x] Add support for AMP Implementation (2026/03/28)
- [ ] Add support for warp-based depth camera
- [ ] Add support for TRON1_WF
## Contacts
You can add our Feishu group to get latest updates or ask questions:
---
```{toctree}
:maxdepth: 1
:hidden:
user_guide/index
developer_guide/index
```