摘要
针对Stewart平台中各关节PID控制自适应性较差,不能实时在线进行参数整定的缺点,结合BP神经网络良好的自学习、自适应性和可以任意精度逼近任意非线性系统的优点,提出PIDNN控制算法,实现运动平台控制器参数的实时在线自整定,提高运动控制系统的控制精度和自适应性。在MATLAB/Simulink中建立Stewart运动平台PIDNN控制系统的仿真模型,对比分析PIDNN和PID运动控制系统的控制效果。仿真结果显示,PIDNN控制算法能够实现PID参数的自适应调节,杆件变化误差大幅下降。表明提出的PIDNN控制算法能够实现参数的实时在线自整定,并且系统具有良好的自适应性和鲁棒性。
Because the PID control for each joints was so poor adaptability in Stewart motion platform,parameters of PID controller can’t be adjusted in real time online.There are the advantages that BP neural network is good self-learning,self-adaptation,and can approach any nonlinear system with arbitrary precision,PIDNN control algorithm is proposed to achieve real-time online self-tuning of controller parameters for motion platform by combining BPNN,to improve the accuracy and adaptability of motion control system.The simulation model of PIDNN motion control system is established in MATLAB,and analyze the control effect of PIDNN and PID motion control system.The result shows that the PIDNN control algorithm can achieve adaptive adjustment of parameters,and system error have dropped significantly after neural network training.It’s proved that the proposed PIDNN control algorithm can achieve real-time online self-tuning of parameters,and the system has good adaptability and Robustness.
作者
朱道扬
段少丽
ZHU Daoyang;DUAN Shaoli(Institute of Intelligent Manufacturing,Wuhan Technical College of Communications,Wuhan 430000,China)
出处
《系统仿真技术》
2019年第3期193-197,共5页
System Simulation Technology
基金
武汉交通职业学院青年项目基金(Q2018001)
中国交通教育研究会基金(1602-6)
创新团队(CX2018A07)
