机器人关节机械驱动下手臂位置的闭环控制  

Research on Closed⁃loop Control Method for Arm Position Under Mechanical Drive of Robot Joints

作　　者：苏永华 李佳阳[2] 韩蓉[2] 田易之[3] 戴璐 SU Yonghua;LI Jiayang;HAN Rong;TIAN Yizhi;DAI Lu(School of College of Intelligent Manufacturing Engineering,Guangxi Vocational and Technical College of Manufacture Engineering,Nanning 530105,China;Harbin Institute of petroleum,College of Mechnical Engineering,Harbin 150000,China;School of Electrical Engineering,Xinjiang University,Urumqi 830000,Chian;Guangxi Academy of Oceanography,Nanning 530001,China)

机构地区：[1]广西制造工程职业技术学院智能制造工程学院,南宁530105 [2]哈尔滨石油学院机械工程学院,哈尔滨150000 [3]新疆大学电气工程学院,乌鲁木齐830000 [4]广西壮族自治区海洋研究院,南宁530022

出　　处：《机械设计与研究》2024年第1期121-124,135,共5页Machine Design And Research

基　　金：黑龙江省自然科学基金项目(LHI2022E094)。

摘　　要：由于机器人的手臂运动涉及到非线性动力学效应,使得手臂的运动模式复杂,手臂位置闭环控制难度较大、精度较低。为此,提出机器人关节机械驱动下手臂位置闭环控制方法。利用力矩传感器在关节上展开力矩反馈来完成机械驱动机器人的动力学建模,通过动力学模型准确估计动力学参数;基于模糊神经网络改进PID控制器,模糊神经网络可以作为PID控制器的前馈部分,导致输入动力学参数可提供更准确的控制信号。将期望的轨迹参数、关节力矩参数等动力学参数作为PID控制器控制目标,通过粒子群算法通过不断迭代和优化找到最优解,实现对机器人手臂位置展开闭环控制。实验结果表明,该方法的控制稳定性强、控制精度高,关节角误差在1%以内,且关节力矩变化幅度较小。Due to the nonlinear dynamic effects involved in the arm movement of robots,the motion mode of the arm is complex,and the closed⁃loop control of the arm position is difficult and has low accuracy.Therefore,a closed⁃loop control method for arm position under mechanically driven of robot joints is proposed.Torque sensors are utilized to expand torque feedback on the joints to complete the dynamic modeling of mechanically driven robots.The dynamic parameters are estimated through the dynamic model.Based on improvement to the PID controller using fuzzy neural network,the fuzzy neural network can serve as the feedforward part of the PID controller,resulting in more accurate control signals provided by the input dynamic parameters.Taking the expected trajectory parameters,joint torque parameters,and other dynamic parameters as the control objectives of the PID controller,the particle swarm optimization algorithm is used to continuously iterate to find the optimal solution,to achieve closed⁃loop control of the robot arm position.The experimental results show that the method has strong control stability,high control accuracy,with a joint angle error within 1%,and small joint torque variation.

关 键 词：机械驱动 闭环控制 改进粒子群算法 PID控制 模糊神经网络 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]