基于BP神经网络的球杆控制算法设计  

Design of Ball and Beam Control Algorithm Based on the BP Neural Network

作　　者：余松灿 刘永信[1] 王玲琳[1] YU Song-can;LIU Yong-xin;WANG Ling-lin(College of Electronic and Information Engineering,Inner Mongolia University,Hohhot 010021,China)

机构地区：[1]内蒙古大学电子信息工程学院

出　　处：《内蒙古大学学报（自然科学版）》2019年第3期309-316,共8页Journal of Inner Mongolia University：Natural Science Edition

基　　金：国家自然科学基金项目资助(61362002)

摘　　要：针对球杆系统定位控制问题,基于BP神经网络设计了BP神经网络控制器和BP神经网络PID参数自整定两种智能控制器.完成了两种控制器的网络结构与实现方法,并在Simulink环境中仿真.仿真结果显示出BP神经网络PID参数自整定控制器的稳定性优于BP神经网络控制器,将BP神经网络PID参数自整定控制器算法移植到GBB1004球杆系统,实现了对该系统的控制.实验结果显示,该控制器响应快,有一定的抗干扰能力,获得系统调节时间小于16s,稳态误差小于1cm.Aimed at the position control for ball-beam system,two intelligent controllers are designed based on the BP neural network.One is a BP neural network controller,and the other is a BP neural network self-tuning PID parameter controller.The structure of neural network and implementation are fulfilled.The simulated results in Simulink environment show that the BP neural network controller with the self-tuning PID parameter is more stable than the BP neural network,so this algorithm is transplanted to GBB1004 ball-beam system and control to this system has been implemented.The experimental results show that the respond of BP neural network controller with the self-tuning PID parameter is fast and also there is more anti-interference ability.The time that system reachs stable is less than 16 s and the steady state error is less than 1 cm.

关 键 词：球杆系统 BP神经网络 智能控制 PID SIMULINK 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]