Robust Non-fragile Control Of Linear Networked Control Systems

With the rapid development of the computer networks technology and the increasing dimension of control problem, feedback control systems connected via a serial communication channel, which are called networked control systems(NCS), has obtained more and more attention in control community. Compared with conventional point-to-point control systems, the advantages of NCS are less wiring, lower install cost as well as greater agility in diagnosis and maintenance. However, the introducing communication networks into the control loops makes the analysis and design of NCS complicated, due to random network-induced delays, data packet dropouts as well as single- packet or multiple-packet transmission.The key issues of network-induced delays and data packet dropouts in the networked control systems are studied in this dissertation. Main achievements are listed as follows:(1) Networked control systems with data packet dropout are described by an uncertain switched system model. Based on Lyapunov function method, concerning constrained control, the robust control problem with H_∞performance bound is studied for the networked control systems. By applying H_∞control theory and the linear matrix inequality(LMI) technique, an approach to design the output feedback controller based on state observer is presented. The stability of NCS is also given via LMIs form.(2) The problem of optimal non-fragile H_∞control via state feedback is investigated for a class of networked control systems with network-induced delays and limited external disturbance. By viewing the time-varying delays as time-varying parameter uncertainties, sufficient condition about the existence of suboptimal non-fragile H_∞control law with control gain perturbation has been developed, and the design method of optimal non-fragile H_∞control law is presented.(3) The problem of robust non-fragile guaranteed cost control via state feedback is investigated for a class of networked control systems with network-induced delays and a quadratic cost index. By viewing the time-varying delays as time-varying parameter uncertainties, sufficient conditions about the existence of robust non-fragile guaranteed cost control with two classes of control gain perturbations have been developed for the uncertain systems. Furthermore, the design problem of the suboptimal guaranteed cost controller is turned into a convex optimization problem with linear matrix inequalities constraints.