Learning Robust Motion Skills via Critical Adversarial Attacks for Humanoid Robots
By: Yang Zhang , Zhanxiang Cao , Buqing Nie and more
Potential Business Impact:
Makes robots walk better on tricky ground.
Humanoid robots show significant potential in daily tasks. However, reinforcement learning-based motion policies often suffer from robustness degradation due to the sim-to-real dynamics gap, thereby affecting the agility of real robots. In this work, we propose a novel robust adversarial training paradigm designed to enhance the robustness of humanoid motion policies in real worlds. The paradigm introduces a learnable adversarial attack network that precisely identifies vulnerabilities in motion policies and applies targeted perturbations, forcing the motion policy to enhance its robustness against perturbations through dynamic adversarial training. We conduct experiments on the Unitree G1 humanoid robot for both perceptive locomotion and whole-body control tasks. The results demonstrate that our proposed method significantly enhances the robot's motion robustness in real world environments, enabling successful traversal of challenging terrains and highly agile whole-body trajectory tracking.
Similar Papers
One-shot Humanoid Whole-body Motion Learning
Robotics
Robots learn new moves from just one example.
Adversarial Locomotion and Motion Imitation for Humanoid Policy Learning
Robotics
Robots move better by splitting body jobs.
Learning Perceptive Humanoid Locomotion over Challenging Terrain
Robotics
Robots walk better on rough ground.