基于多人零和博弈的模块化机器人系统近似最优控制
Approximate optimal control for Modular Robot Manipulators based on multiplayer zero-sum game
董博 1朱新野 1马冰 1安天骄1
作者信息
- 1. 长春工业大学电气与电子工程学院,吉林长春 130012
- 折叠
摘要
提出一种基于多人零和博弈的模块化机器人(Modular Robot Manipulators,MRMs)系统近似最优控制方法.建立了具有交联耦合(Interconnected Dynamic Couplings,IDC)的模块化机器人系统动力学模型.将机器人系统的控制律和IDC效应作为零和博弈的参与者,MRM系统的最优跟踪控制问题转化为多人零和博弈问题.根据自适应动态规划(Adaptive Dynamic Programming,ADP)算法,通过建立评判神经网络求解哈密顿-雅克比-埃塞克斯(Hamilton-Jacobi-Issacs,HJI)方程,推导出最优控制律.基于李雅普诺夫定理,证明了闭环机器人系统是渐近稳定的,最后通过实验验证了所提控制方法的有效性.
Abstract
An approximate optimal control method for Modular Robot Manipulators(MRMs)systems based on multiplayer zero-sum game is proposed.A modular robot system dynamic model with Interconnected Dynamic Couplings(IDC)is developed.The control law and IDC effect are regarded as players in zero-sum game.The approximate optimal tracking control problem of the MRM system can be transformed into a multiplayer zero-sum game.According to the Adaptive Dynamic Programming(ADP)algorithm,the Hamilton-Jacobi-Issacs(HJI)equation can be solved by establishing critic neural network and then the approximate optimal control policy can be derived.Based on the Lyapunov theorem,the closed-loop robotic system is proved to be asymptotic stable.Finally,experiments are conducted to verify the effectiveness of the proposed method.
关键词
自适应动态规划/模块化机器人/多人零和博弈/最优控制Key words
ADP(Adaptive Dynamic Programming)/MRMs(Modular Robot Manipulators)/multiplayer zero-sum game/optimal control引用本文复制引用
基金项目
国家自然科学基金项目(62173047)
吉林省科技发展计划项目(20220201038OX)
吉林省教育厅"十三五"科学计划项目(JJKH20220689KJ)
出版年
2024
NETL
NSTL
万方数据