Learning agility and adaptive legged locomotion via curricular hindsight reinforcement learning

作者：Sicen Li, Gang Wang, Yiming Pang, Panju Bai, Shihao Hu, Zhaojin Liu, Liquan Wang, Jiawei Li · 发表于：Scientific Reports · 年份：2024 · DOI：10.1038/s41598-024-79292-4 · 被引用次数：9 · 研究领域：Robotic Locomotion and Control、Muscle activation and electromyography studies、Prosthetics and Rehabilitation Robotics

Agile and adaptive maneuvers such as fall recovery, high-speed turning, and sprinting in the wild are challenging for legged systems. We propose a Curricular Hindsight Reinforcement Learning (CHRL) that learns an end-to-end tracking controller that achieves powerful agility and adaptation for the legged robot. The two key components are (i) a novel automatic curriculum strategy on task difficulty and (ii) a Hindsight Experience Replay strategy adapted to legged locomotion tasks. We demonstrated successful agile and adaptive locomotion on a real quadruped robot that performed fall recovery autonomously, coherent trotting, sustained outdoor running speeds up to 3.45 m/s, and a maximum yaw rate of 3.2 rad/s. This system produces adaptive behaviors responding to changing situations and unexpected disturbances on natural terrains like grass and dirt.