基于RBF神经网络逆系统的机械手轨迹跟踪控制  

Trajectory Tracking Control of Manipulator Based on RBF Neural Network Inverse System

作　　者：缸明义[1] 陈立辛 乔印虎 夏兴国[1] GANG Mingyi;CHEN Lixin;QIAO Yinhu;XIA Xingguo(Department of Electrical Engineering,Maanshan Technical College,Maanshan 243031,China;College of Mechanical Engineering,Anhui Science and Technology University,Fengyang 233100,China)

机构地区：[1]马鞍山职业技术学院电气工程系,安徽马鞍山243031 [2]安徽科技学院机械工程学院,安徽凤阳233100

出　　处：《安徽科技学院学报》2021年第3期68-74,共7页Journal of Anhui Science and Technology University

基　　金：安徽省高校自然科学研究重点项目(KJ2019A1245);安徽省高校优秀青年人才支持计划重点项目(gxyqZD2018105)。

摘　　要：目的:针对机械手系统的高度耦合、非线性等动力学特性,系统结构和参数在实际工作中存在诸多不可预知因素,研究机械手的轨迹跟踪问题。方法:利用神经网络来构建机械手的逆系统模型,将其与被控对象串联构成伪线性系统,从而将非线性问题转化为线性问题,实现了机械手的在线建模、解耦控制。结果:仿真实验结果表明,迅速的跟踪给定的期望轨迹,脉冲干扰并未系统带来明显影响。结论:方案对控制系统有较强的适应性、稳定性以及抗干扰性能,有效地解决了机械手的轨迹跟踪问题。Objective:In view of the highly coupled and nonlinear dynamic characteristics of the manipulator system,there are many unpredictable factors in the system structure and parameters in practical work,so the trajectory tracking problem of the manipulator is studied.Methods:The inverse system model of the manipulator is constructed by using neural network,which is connected with the controlled object in series to form a pseudo-linear system,thus transforming the nonlinear problem into a linear problem,and realizing the online modeling and decoupling control of the manipulator.Results:The simulation results show that the impulse interference has no obvious influence on the system when the desired trajectory is tracked quickly.Conclusion:The scheme has strong adaptability,stability and anti-interference performance against the control system,and effectively solves the trajectory tracking problem in the manipulator.

关 键 词：机械手 RBF神经网络 逆控制 轨迹跟踪 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程]