The Study Of The Generalized PID Algorithm Based On The Neural Network

PID controller has been used in the ninety percent industrial control process today for the reason that the PID controller has the simple structure ,easy to achieve,the good control effect and strong robustness.However,the traditional PID controller could't get better control effect especially when used in the nonlinear systems and the time-varying systems.The intelligent control gets a full development with the rapid development of the computer technology, and combines with the PID controller .For this new controller,the traditional PID can better satisfy the needs of the people and the industrial processes.This paper studies the PID controller system and intelligent PID control system,the main work includes:(1) The generalized PID controller is proposed based on the analysis of conventional PID controller, then discusses the connection between the generalized PID and the conventional PID.(2) Neural network especially the BP neural network and the RBF neural network are also in research in theory.And combinated them with the generalized PID controller to get the neural network GPID controller,finally compar the simulation results with the conventional PID controller.(3) The ball and beam system,inverted pendulum system is studied and get the model contain both of them,then make control processing through the BP neural network;on the other hand make simulation compare between the GPID and the conventional PID used in the model of three dof helicopter.(4) Through comparing the simulation,verify the GPID is better than the conventinal PID control effect.