基于近端策略优化算法的四足机器人步态控制研究
|
摘要
足式机器人步态控制是机器人研究领域的难点问题,应用强化学习让机器人自主学习策略提供了一种很好的解决思路.基于ROS机器人操作系统搭建了四足机器人仿真平台,将近端策略优化算法用于四足机器人步态控制,并与其他深度强化学习算法进行了对比分析.仿真实验结果表明,近端策略优化算法在实际应用中具有更好的训练效果.
The gait control of the foot robot is a difficult problem in the field of robot research. Applying reinforcement learning provides a good solution to the robot control policy. Based on the ROS robot operating system,a quadruped robot simulation platform is built. The proximal policy optimization algorithm is applied to the quadruped robot gait control,and compared with other deep reinforcement learning algorithms. Simulation results show that the proximal policy optimization algorithm has better training effect in practical applications.
The gait control of the foot robot is a difficult problem in the field of robot research. Applying reinforcement learning provides a good solution to the robot control policy. Based on the ROS robot operating system,a quadruped robot simulation platform is built. The proximal policy optimization algorithm is applied to the quadruped robot gait control,and compared with other deep reinforcement learning algorithms. Simulation results show that the proximal policy optimization algorithm has better training effect in practical applications.
引文
[1]MNIH V,KAVUKCUOGLU K,SILVER D,et al.Human-level control through deep reinforcement learning[J].Nature,2015,518(7540):529.
[2]SCHULMAN J,LEVINE S,MORITZ P,et al.Trust region policyoptimization[J].Computer Science,2015:1889-1897.
[3]SCHULMAN J,WOLSKI F,DHARIWAL P,et al.Proximal policy optimization algorithms[J].ar Xiv preprint ar Xiv:1707.06347,2017.
[4]BARTO A G.Reinforcement learning[J].A Bradford Book,1998,15(7):665-685.
[5]周志华.机器学习[M].北京:清华大学出版社,2016:377-382.
[6]SUTTON L,RICHARD S,BARTO F,et al.Introduction to reinforcement learning[J].Machine Learning,2005,16(1):285-286.
[7]SUTTON R S,MCALLESTER D A,SINGH S P,et al.Policy gradient methods for reinforcement learning with function approximation[C]∥Neural information processing systems,1999:1057-1063.
[8]LILLICRAP T P,HUNT J J,PRITZEL A,et al.Continuous control with deep reinforcement learning[J].Computer Science,2015,8(6):A187.
[9]ZHANG H Y,XIONG K.Improved deep deterministic policy gradient algorithm based on prioritized sampling[C]∥Chinese Intelligent Systems Conference.Singapore:Springer,2018.
[2]SCHULMAN J,LEVINE S,MORITZ P,et al.Trust region policyoptimization[J].Computer Science,2015:1889-1897.
[3]SCHULMAN J,WOLSKI F,DHARIWAL P,et al.Proximal policy optimization algorithms[J].ar Xiv preprint ar Xiv:1707.06347,2017.
[4]BARTO A G.Reinforcement learning[J].A Bradford Book,1998,15(7):665-685.
[5]周志华.机器学习[M].北京:清华大学出版社,2016:377-382.
[6]SUTTON L,RICHARD S,BARTO F,et al.Introduction to reinforcement learning[J].Machine Learning,2005,16(1):285-286.
[7]SUTTON R S,MCALLESTER D A,SINGH S P,et al.Policy gradient methods for reinforcement learning with function approximation[C]∥Neural information processing systems,1999:1057-1063.
[8]LILLICRAP T P,HUNT J J,PRITZEL A,et al.Continuous control with deep reinforcement learning[J].Computer Science,2015,8(6):A187.
[9]ZHANG H Y,XIONG K.Improved deep deterministic policy gradient algorithm based on prioritized sampling[C]∥Chinese Intelligent Systems Conference.Singapore:Springer,2018.