Aiming at the problem that gait planning and control of hexapod robots are difficult due to com-plex system structure and decision making,this paper proposes a distributed reinforcement learning-based gait learning and control framework for hexapod robots.In this framework,training data is obtained through massively parallel learning and data-driven training is conducted through reinforcement learning to obtain the optimal control strategy.The simulation environment was constructed under IsaacGym platform to eval-uate and verify the performance of hexapod robot gait learning and control.The results show that the trained hexapod robot gait has good performance in reward value,speed tracking and stability,which verifies the effectiveness of the proposed method.
hexapod robotdistributed deep reinforcement learninggait learning
NETL
NSTL
万方数据
