基于分布式深度强化学习的六足机器人步态学习与控制  

Distributed Deep Reinforcement Learning-Based Gait Learning and Control for Hexapod Robots

作　　者：李伦 向国菲[1] 马丛俊 佃松宜[1] LI Lun;XIANG Guofei;MA Congjun;DIAN Songyi(College of Electrical Engineering,Sichuan University,Chengdu 610065,China)

机构地区：[1]四川大学电气工程学院,成都610065

出　　处：《组合机床与自动化加工技术》2024年第8期85-89,共5页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：四川省科技计划资助项目(2023NSFSC1441)。

摘　　要：针对六足机器人系统结构与决策的复杂导致步态规划与控制困难的问题,提出了一种基于分布式强化学习的六足机器人步态学习与控制框架。该框架采用大规模并行学习的方式获取训练数据,通过强化学习的方法对网络进行数据驱动训练,得到最优控制策略,并在IsaacGym平台下构建仿真环境评估和验证六足机器人步态学习与控制的性能。结果表明,训练的六足机器人步态在奖励值、速度跟踪和稳定性都有良好的表现,验证了所提方法的有效性。Aiming at the problem that gait planning and control of hexapod robots are difficult due to complex 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 evaluate 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.

关 键 词：六足机器人 分布式深度强化学习 步态学习 

分 类 号：TH165[机械工程—机械制造及自动化] TG659[金属学及工艺—金属切削加工及机床]