摘要
提出了一种基于遗传算法和小波神经网络的PID参数整定方法。首先,利用具有自然进化的遗传算法对小波神经网络的初始权值进行优化训练,解决了控制器网络初始权系数对控制效果产生的影响;其次,利用小波神经网络对PID参数进行在线调节;最后,将此算法运用到电机控制系统的PID参数寻优中。仿真结果表明:基于此算法设计的PID控制器可以极大地提高寻优速度,鲁棒性强,改善了电机控制系统的动态性能和稳定性。
A new-type controller based on genetic algorithm and wavelet neural network was presented. The genetic algorithm was utilized to train and optimize the initial weights of wavelet neural network, the method solved the influence of the initial weights of wavelet neural network on the control effectiveness. The wavelet neural network was applied to realize PID parameters self-adjustment on line. At last, the algorithm was applied to electro motor control system PID for parameter optimization. The simulation result shows that the PID controller designed upon this approach improves the search speed and dynamic quality of electro control system.
出处
《仪表技术与传感器》
CSCD
北大核心
2009年第4期132-134,共3页
Instrument Technique and Sensor
关键词
遗传算法
小波神经网络
PID控制
参数寻优
电机控制系统
genetic algorithm
wavelet neural network
PID control
parameter-optimization
electro motor control system
