The Comprehensive Research On Modeling And Control Of Networked Control Systems With Time Delay

With the developing of modern science technology, constantly cross andintegration of communications engineering, computer engineering and automaticcontrol engineering, networked control systems(NCSs) come into being. Compared withthe traditional point to point control systems, NCSs have many attractive advantages,such as resources sharing, easy operation and maintenance, simple wiring, low cost,high reliability, high flexibility and so on. However, the insertion of the network, it hasbrought many new problems and challenges. Communication network inevitablyappears network latency, packet loss, packet disorder and so on. It is known that theoccurrence of these problems degrade the stability and control performance of thecontrol loop system, what’s worse, may lead to the system instability. Traditionalcontrol theories and methods are not applicable to the NCSs. Therefore, it is necessaryto explore new control methods and strategies to study the NCSs in the control areas.In this paper, the study mainly focuses on networked control systems with timedelay. Firstly, a detailed analysis is given for the impact of time delay on systemperformance. Secondly, for two types of delay, constant delay and random delay,modeling and control research are presented, respectively. For the constant delay, thecontinuous system state space model is established, considering the delay of the inputsignal. Based on Lyapunov stability theory and Riccati inequalities, system stability isanalyzed, and controllers are designed, ensuring that the system is Delay-Independentstable. A feasible solution of the controllers is derived through LMI toolbox. For therandom delays both in the forward channel and the feedback channel, they are describedas two different Markov chains. And the distribution of the delay state transitionprobability is considered as two decimal places. Then, based on Lyapunov theory,LMI toolbox and Markov jump systems theory, the sufficient condition for thestochastic stability is established, and a two-mode-dependent state feedback controller isdesigned. A nonlinear linear matrix inequality approach is used to calculate thecontrollers. Using the cone complementarity linearization algorithm, the nonlinearlinear matrix inequality is transformed into inequality optimization problem.Finally, networked control system simulation platform is built via the modules ofTruetime toolbox and Simulink. A comprehensive simulation study is concentrated on these factors, such as time delay, packet loss, interference and different networkprotocols. It is shown that the influence of controller-to-actuator (C-A) delay on thesystem performance is more intense than sensor-to-controller (S-C) delay when longdelay considered. In addition, the system has a certain tolerance range for bandwidthutilization and packet loss rate, when they are out of the range, the system performancewill be destroyed.