Chasing Stability: Humanoid Running via Control Lyapunov Function Guided Reinforcement Learning
Chasing Stability: Humanoid Running via Control Lyapunov Function Guided Reinforcement Learning
This paper presents CLF-RL, a method that integrates nonlinear control theory with reinforcement learning to achieve stable running on humanoid robots. By embedding Control Lyapunov Functions into the reward structure and using full-order dynamics for trajectory optimization, the approach eliminates the need for hand-crafted heuristic rewards. The resulting policy enables a Unitree G1 robot to perform reliable full-flight running in tight corridors and outdoor settings while remaining robust to disturbances and utilizing only on-board sensors for global reference tracking.
The authors propose a method combining Control Lyapunov Functions with Reinforcement Learning to stabilize humanoid running, demonstrating robust performance on a Unitree G1 robot in both treadmill and outdoor environments.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
Reported numbers
position tracking improvement
1.7
View original evidence
wards. We further develop a trajectory optimization routine that can handle the hybrid structure of the system to produce dynamically feasible, multi-domain, trackable references; using full-order dynamics rather than a kinematic approximation yields up to a 1.7x improvement in position tracking. The resulting policy runs on a Unitree G1 humanoid with a full flight phase, operating reliably in tight corridors such as
Open source S4
- paper: not_reported
- code: not_reported
- weights: not_reported
- dataset: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
wards. We further develop a trajectory optimization routine that can handle the hybrid structure of the system to produce dynamically feasible, multi-domain, trackable references; using full-order dynamics rather than a kinematic approximation yields up to a 1.7x improvement in position tracking. The resulting policy runs on a Unitree G1 humanoid with a full flight phase, operating reliably in tight corridors such as
Open source S4
a treadmill as well as in outdoor environments, all while remaining robust to disturbances applied to the torso and feet. Moreover, it achieves accurate global reference tracking utilizing only on-board sensors, making a critical step toward integrating these dynamic motions into a full autonomy stack.
Open source S5
Chasing Stability: Humanoid Running via Control Lyapunov Function Guided Reinforcement Learning
Open source S1
arXiv:2509.19573v2 Announce Type: replace Abstract: Achieving highly dynamic behaviors on humanoid robots, such as running, requires controllers that are both robust and precise, and hence difficult to design. Classical control theory offers principled tools for stabilizing such systems but struggles to synthesize real-time controllers for their nonlinear, hybrid dynamics; reinforcement learning (RL) handles these dy
Open source S2
namics well but typically depends on extensive hand-tuning of heuristic rewards. In this work, we embed ideas from nonlinear control theory, specifically control Lyapunov functions (CLFs), along with optimized reference trajectories for the hybrid system into the reinforcement learning reward. This approach, CLF-RL, replaces the need to handcraft and tune heuristic reward terms with easy to tune stability informed re
Open source S3
Source：arXiv Robotics — research abstracts · arxiv.org