Research On Filter For Wireless Networked Control System

In the wireless networked control system, with the constraints of storage space, limited communication bandwidth and CPU processor, network delay, data packet dropouts are caused, which seriously affect the control performance and stability of the system, however, the network delay is one of the essential factors that affect the performance of the wireless networked control system. Because the signal is susceptible to external interference of noise factors in the process of measurement and transmission and some system states are unobservable, estimation of the object states objects to minimize the impact of measurement noise and process noise on the system is necessary. Therefore, combined with the latest research literature in related fields, two kinds of filters——Kalman filter and H∞ filter were designed, and they were simulated under the hardware-in-loop simulation through ZigBee technology and virtual instrument technology.The main researches of this paper are outlined as follows:(1) For linear discrete system with Gauss white noise and packet dropouts, packet dropouts are modeled as stochastic processes with Bernoulli distribution. Through the method of hold input to compensate, design of the Kalman filter was done to eliminate the interference of noise and packet dropouts on system effectiveness, finally numerical example is given to illustrate the validity of the algorithm;(2) In view of wireless networked control system with the existence of noise interference, one-step delays, packet dropouts, and noise in the system is bounded energy, the delay and packet dropouts are modeled as independent distributed Bernouli process, and solving parameters by using the linear matrix inequality method for H∞ filter to draw the conclusion that the filtering error system is stable under the meaning to square sense exponential and make the H∞ filter have a certain performance index. The results of numerical simulation showed that the algorithm can eliminate the influence to the noise, time delays and data packet dropout.(3) Based on ZigBee and virtual instrument technology, the hardware-in-loop simulation platform is designed in the wireless network control system. Conclusion, further verify the proposed results show that under the condition of the statistical characteristics of the unknown H∞ filtering performance is better than Kalman filtering.