Research On The Problem Of Time Delay And Packet Loss For Wireless Networked Control System

Wireless networked control system was a nonlinear closed-loop feedback control system. Each node was flexible and highly integrated because they were connected through the wireless transmission network. The system could be extended and maintained convenient. wiring cost was low. But a series of problems appeared including delay, packet loss, time variant gain, interference and limited bandwidth. The first two were most important. The system had been greatly threatened.This paper considers about predicting and compensating the time delay and packet loss to improve the real-time stability of the system from the perspective of control strategy. Smith delay predictive compensation and variable forgetting factor packet loss compensation were respectively combined with the adaptive fuzzy PID control.The conventional PID had been widely used in the field of control system. However, it was obviously inadequate in wireless networked control system. Fuzzy control could collect previous experience as a knowledge base. they were used to handle the inaccuracy information by fuzzy logic. It was suitable for complex system. This paper firstly adjusted three PID dynamic parameters by using self-adaptive fuzzy PID control. It was real-time and dynamic for the system.However, fuzzy controller could not solve the steady-state problem completely. It was inaccurate. Smith prediction compensator had excellent anti-interference performance for time-varying system. It could track output better on the premise of a accurate model.This paper adapted an improved self-adaptive fuzzy PID control design to solve the time delay problem. A Smith time delay compensation was added It was based on the controller side. This design did not need to establish exact model. It also could deal with the adverse effects of the random time delay. The simulation results based on TrueTime showed that this design could compensate the time delay better and improved the system quality.This paper adopted a kind of forgetting factor method for packet loss. This method was originally used in the least square method for identification system. In this paper, a number of historical data packets were collected to predict and replace the current packet loss. Every historical data packets had a time-varying forgetting factor as weight. Packet loss was compensated. The value of the forgetting factor depended on the error between the output and the reference value. Similarly, the design was improved. Variable forgetting factor was used for packet loss compensation. The simulation based on the TrueTime had proved the reliability of the method.