基于自适应PID控制器的平面一级倒立摆的研究  

Research of plane one rod inverted pendulum based on self-adaptive PID controller

作　　者：葛动元[1,2] 姚锡凡[1] 

机构地区：[1]华南理工大学机械与汽车工程学院,广州510640 [2]邵阳学院机械与能源工程系,邵阳422004

出　　处：《机械设计与制造》2009年第3期198-200,共3页Machinery Design & Manufacture

基　　金：国家高技术研究发展计划(863计划),专项经费资助项目(2007)AA04Z111

摘　　要：倒立摆是一个强耦合、严重不稳定的系统,其背景来源于火箭发射等课题。在该系统中,PID控制器常常被采用。由于该系统在建立数学模型时次要的因素被忽略了,实际上是一个非线性系统;为了提高系统的控制性能,根据计算智能逼近非线性系统的功能,设计一个RBF神经网络控制系统,实现对常规PID控制器的参数进行自适应整定。最后使用BC++编写系统的控制程序,通过实物控制验证基于RBF神经网络的PID控制器参数的自适应整定的系统具有较好的瞬态性和鲁棒性。Inverted pendulum system is a close-coupling and serious unstable system,whose back- ground comes from subject such as rocket launch. In this system PID controller was adopted usually. Because of the secondary factar neglected while establishing physical model,the system is non-linear actually. In order to improved control performance of system, in the light of function of computational intelligence approaching nonlinear system,the RBF Neural Network control system was designed for carrying out adaptively tuning of conventional PID controller parameters. Finally the BC＋＋ was adopted to program controlling procedure for the system,and verify the system with adaptive tuning of PID controller based on RBF NN having better robust performance and instantaneous performance through material object control.

关 键 词：PID控制器 RBF神经网络 自适应控制 鲁棒性与瞬态性 BC++ 

分 类 号：TH16[机械工程—机械制造及自动化] TP273.2[自动化与计算机技术—检测技术与自动化装置]