基于神经网络的自适应PID控制器设计  

The Design of Adaptive PID Controller Based on Neural Network

作　　者：欧艳华[1] 

机构地区：[1]柳州职业技术学院,广西柳州545006

出　　处：《机械设计与制造》2014年第6期263-265,共3页Machinery Design & Manufacture

基　　金：广西自然科学基金项目(桂科攻0815003-6)

摘　　要：为了提高PID控制器参数根据实际误差自动调整能力,提出了基于神经网络的自适应PID控制器的设计。详细分析了神经网络自适应控制原理,设计了通过神经网络对PID控制器参数自动调整的方案。利用Matlab仿真工具,对自适应PID控制器进行了仿真测试,给出了神经网络训练参数,通过大量数据的模拟测试,结果表明基于神经网络的自适应PID控制器能够根据控制误差自动调整PID参数,对输入的阶跃信号响应平稳,综合性能大大提高。To improve the auto-adjust ability to practical error,it raised the design of adaptive PID controller based on neural network.It specifically analyzed the adaptive principle,and designed the auto-adjust implementation scheme of PID controller parameters by neural network.Using Matlab as a simulation tool,the adaptive PID controller was emulation tested and the neural network training parameter was given.By the simulation test of mass data,it shows that the adaptive PID controller which based on neural network can automatically adjust PID parameter by controlling error,and has stable response to step signal.Its comprehensive performance is greatly improved.

关 键 词：神经网络 PID 自适应 参数调整 阶跃信号 

分 类 号：TH16[机械工程—机械制造及自动化] TM57[电气工程—电器]