基于小波基神经网络PID的直流电机伺服控制

DC Motor's Servo Controller Based on WNN PID

通过一种新颖的小波基神经网络对未知数学模型的对象进行在线辨识,得到对象的数学模型———Jacob ian信息,并提出了神经网络与模糊控制算法共同在线调整PID参数的方法,从而实现电机位置的准确、快速、实时地跟踪.通过仿真和实验表明:使用该自适应控制方法,能够对位置准确跟踪,基本克服了一般神经网络控制对初始权值的依赖,大大提高了对未知模型的辨识精度,改善了系统的动态响应品质,增强了系统的鲁棒性.

A novel topology network of WNN is introduced to identify on line the object of a certain unknown mathematical model, and obtain the object＇ s mathematical model the Jacobian information. A method is suggested of PID parameters co-adjusted on line by WNN and fuzzy arithmetic, and thus realize the accurate, quick and real - time track. The sim together self-tuning using is reported to track the motor＇ s position precisely. The simulation and the experiment results show that the adaptive PID control method can be used to track accurately the position, and overcome the dependence of general neural network on initial weight, and greatly increase the precision of identifying the unknown model, and imorove the control.