Filtering And Control Of Wireless Networked Control System

Wireless networked control systems(WNCSs) are feedback control systems with data commu-nication, detection and closed control loops through wireless network communication channels, It is the syncretic result of automatic control technology, communication technology and computer science. Gaining attention is due to their advantages in reducing the complexity of wiring con-nections, decreasing the costs of cables and power, simplifying the installation and maintenance of the whole system. Wireless networked control system has been widely employed by the industries like chemical, automobiles, aircrafts and so on. Meanwhile, the research on networked control system has also attracted the attention of many scholars. However, the insertion of communication channels makes the analysis and design of wireless networked control systems complicated and hence raises new interesting and challenging problems such as time delays, time-varying network topologies and packet losses.This thesis focuses on the details of networked control systems. It mainly deals with the stabilization and filtering problem for networked control system with induced time delays, packet loss or randomly occurred sensor nonlinearities.Specifically, the main contribution of this thesis can be summarized as follows:(1) To figure out the characteristics of network induced time delay, the design and realization of a data transmission test platform based on ZigBee wireless network is introduced. Hundreds of hours of data transmission test was performed on this platform for three frequently used network topologies. In addition, the cause and the composition of the network induced delays are analysed with the data acquired from the experiments;(2) In order to study the relation between sample rate and the finite wireless communication bandwidth, the thesis focuses on the stabilization algorithm of networked control system under multiple sampling rates in three different scenarios and the effectiveness of the given method is illustrated by numerical examples; (3) Lyapunov-Krasovskii functionals are used to design an H∞filter for the networked control systems with randomly occurred sensor nonlinearities and packet dropouts. The filter ob-tained not only makes the filtering error system stable but also achieves the prescribed H∞level;(4) The thesis presents the concepts of multiple network environments by Markov transition matrix. The stabilizing controller gain can be designed by solving a series of bilinear matrix inequalities;(5) In order to reduce the number of switching between different controllers for different net-work environment and the computational complexity, the interval element transition matrix method is used in the modeling. One controller would be enough for multiple network en-vironments and as a result, this approach effectively reduces the computational complexity. The corresponding controller can be obtained by solving linear matrix inequalities through cone complementary linearization algorithm.Finally, a brief summary of this research work and perspective of future studies are presented as the end of this thesis.