Neural Network Fractional PID Research And Applications In The Network Control System

With the development and cross—penetration of control of science, computer technology and comunication technology, networked control came into being and is becoming one of the hot topics of the international academic research. However, because of inserting the network into traditional control system, it is inevitable for the existence of the information transmission delay which are the main factors to degrade the performance of control systems.In order to reduce or eliminate the impact of network delay to the control system, the theory at home and abroad on the basis of the network control system is referenced,and system modeling, controllers are designed.The main contents are conduded as follows:Firstly,this paper introduces the background of networked control system,development status and characteristic,and a brief analysis of several problems in networked control systems. Without considering other factors of network, Mathematical modeling of the short delay and long delay of networked control systems is established.Secondly,compared with traditional PID controller, fractional order controller has better adaptability.But the selection of controller parameters is more difficult for a a fractional order controller which introduces additional two parameters λ and μ.RBF neural network has the function of self-adaption and self-learning,with the combination of fractional order controller,Effectively improve the ability of the fractional order controller to adapt to the uncertainty system,and the online parameter adjustment,so that the impact of time-delay on NCS gets mostly compensated.Finally, a NCS simulation platform is established by TrueTime toolbox which provides a simulated network environment to verify the validity of signed controllor. The simulation results evidence that the fractional order PID controller based on RBF neural network not only makes a good compensation to the impact of time-delay on NCS,but also represent stronger robustness and adaptability.The experimental results show that, in case of only considering the network delay, RBF neural network based on the fractional order PID controller has better control effect.