Guaranteed Cost Control Of Networked Control Systems With Uncertainty And Long Time Delay

Networked control systems (NCS) are closed-loop control systems which use proprietary or public data communications network to replace the traditional point-to-point connection system. Networked control systems break the space constraints of physical location in traditional control system, broaden the control sites, reduce the complexity of links in system, cut the running and maintenance cost, facilitate the achievement of control incorporation, and enhance the information integration. But a series of network problems can be brought simultaneously, such as time delay, data packet loss, packet disorder and so on. The main contributions of this dissertation for issues of time delay and packet loss are listed as follows:(1) Background of the network control system and current research methods on time delay in network control system are introduced. And theoretical analysis and examples are given to illustrate the effect of time delay towards stability of network control system.(2) The problem of guaranteed cost control with emergence of fixed state delay in object and random transmission delay in control system is investigated. Based on integral inequality and freedom weight matrix approach, a sufficient condition for the existence of guaranteed cost state feedback controller is given in the form of linear matrix inequalities and a design method is also presented, ensuring that network control system is asymptotically stable and that performance index possesses an upper bound. A numerical example illustrates effectiveness and superiority of freedom weight matrix approach.(3) By modeling networked control systems with random network-induced delay into Markov jump linear system, the problem of guaranteed cost control for uncertain systems with distribution of delay in the form of Markov chain is studied. A sufficient condition for asymptotic stability and upper bound for performance of the system is given in the form of linear matrix inequality by using cone complementary iterative algorithm. Numerical examples illustrate the effectiveness of this method for the guaranteed cost control law.(4) For network control systems with long delays and packet losses, by modeling it as an asynchronous system and using stability criteria of asynchronous dynamical system theory, conditions of exponential stability of the network control system with uncertainty are given, and the scope of data transmission success rate is also presented ensuring system exponentially stable. A numerical example verifies the effectiveness of this method.