Analysis And Design Of Networked Control Systems Based On Asynchronous Dynamical And Fault Tolerant Model

ABSTRACT:Networked control system is a feedback control system whose system components are connected over the shared network. Compared with the traditional point-to-point control system, networked control system has obvious advantages:for example, less wiring, lower installation costs, higher fault diagnostic capability as well as easy to re-construct and maintain etc., allowing its wide range of applications in coordinated control of mobile robots, advanced aircraft airborne control and areas in power, rail transport services system and so on. But because of the insertion of network into the control system, the system is restricted by network bandwidth and carrying capacity, resulting in many issues, such as network-induced delay, packet dropout, communication constraints, fault, and quantization error during the process of data transmission, which becomes complicated to analyze and design the controller. The related research is done in this thesis for part of the problems existing above.Firstly, the background and significance of the research, the concept and structure of networked control system and essential problems are presented in this thesis. The current research on different issues, especially on the co-design of control and schedule is summarized. Besides, theoretical basics of asynchronous dynamical systems, switched system, fault tolerant control and linear matrix inequality toolbox are introduced respectively.Secondly, because not all the states of controlled plant are measurable, the design of observer for networked control system with random delay and packet loss is given here. This algorithm consists of two parts:the system is modeled as an asynchronous dynamical system based on whether packet loss occurs and then piece-wise strategy is used for the random delay existing in the system. The timestamp technique is used to measure the current actual delay values, enabling the control system with the idea of "switch" to take different control values according to current time delay values. Simulation with TrueTime toolbox is also made to prove the results.Thirdly, taking the constraints of limited bandwidth in networked control system into consideration, a co-design method of fault tolerant control and dynamic schedule is designed for networked control system with random delay, media access constraints and actuator failures simultaneously. Only a limited number of nodes can access the network at a time, a better weighted compensation strategy with more degree of freedom is used for the data that cannot access the network. Then the dynamic scheduling protocol based on a given quadratic index is chosen to select the scheduling policy. Robust fault-tolerant controller is designed by Lyapunov function and linear matrix inequality toolbox, allowing the system to remain stable even in the presence of actuator faults. Then it is applied to a wheeled robot and Matlab is used to make principle simulation and TrueTime toolbox is also used to simulate network environment, verifying the effectiveness of the algorithm.Finally, the results in this thesis are summarized and further work is also discussed.