基于强化学习的连续型机械臂自适应跟踪控制  

Reinforcement-learning-based Adaptive Tracking Control for a Space Continuum Robot Based on Reinforcement Learning

作　　者：江达 蔡志勤[1] 刘忠振 彭海军[1,2] 吴志刚[2] Jiang Da;Cai Zhiqin;Liu Zhongzhen;Peng Haijun;Wu Zhigang(Dalian University of Technology,Dalian 116024,China;State Key Laboratory of Structural Analysis for Industrial Equipment,Dalian 116024,China)

机构地区：[1]大连理工大学,辽宁大连116024 [2]工业装备结构分析国家重点实验室,辽宁大连116024

出　　处：《系统仿真学报》2022年第10期2264-2271,共8页Journal of System Simulation

基　　金：国家自然科学基金重大研究计划重点项目(91748203);国家自然科学基金优秀青年项目(11922203)。

摘　　要：针对空间主动碎片清除操作中连续型三臂节机器人系统跟踪问题,提出一种基于强化学习的自适应滑模控制算法。基于数据驱动的建模方法,采用BP神经网络对三臂节连续型机械臂进行建模,并作为预测模型指导强化学习实时调节所提出滑模控制器的控制参数,从而实现连续型机器人运动的实时跟踪控制。仿真结果表明:提出的数据驱动的预测模型对随机轨迹预测的相对误差保持在±1%以内,能够高精度地反映系统动态特性。对比固定参数的滑模控制器,提出的自适应控制器在保证系统达到控制目标的同时具有更低的超调量和更短的调节时间,表现出更好的控制效果。Aiming at the tracking control for three-arm space continuum robot in space active debris removal manipulation, an adaptive sliding mode control algorithm based on deep reinforcement learning is proposed. Through BP network, a data-driven dynamic model is developed as the predictive model to guide the reinforcement learning to adjust the sliding mode controller’s parameters online, and finally realize a real-time tracking control. Simulation results show that the proposed data-driven predictive model can accurately predict the robot’s dynamic characteristics with the relative error within ±1% to random trajectories. Compared with the fixed-parameter sliding mode controller, the proposed adaptive controller has a lower overshoot and shorter settling time and can achieve a better tracking performance.

关 键 词：空间连续型机器人 强化学习 预测控制 滑模控制 轨迹跟踪 

分 类 号：TP273.2[自动化与计算机技术—检测技术与自动化装置]