Contents 1 INTRODUCTION
In networked control systems (NCSs), multiple sensors and actuators are connected to a centralized controller via a shared communication networks. NCSs have brought us many great advantages, and have been successfully implemented in many industry processes [1]–[3], However, many new problems have also emerged due to the unreliable communication networks, and the network-induced delay which is regarded as the major cause of the deterioration of system dynamic performance and potential system instability is among these new problems. See the structure of the NCS shown in Fig. 1, in which the time-varying network-induced delays are assumed to be smaller than one sampling period of the NCS. An earlier approach to deal with the short (smaller than one sampling period) tim-varying delay is to introduce a buffer between the controller and the actuator to change the time-varying delay as a constant one[4]. Obviously, such an approach is conservative since the delays are unnecessarily enlarged. Some other approaches, such as the stochastic system approach[5], [6], the Markovian system approach[6] and the Arbitrary switching system approach[7] have been presented in the past decade. However, the stochastic system and the Markovian switched system approaches are not applicable when the delay distribution laws are not explicitly known. Besides, the switched system approach which requires that the overall closed-loop NCS is stable under arbitrary switching is conservative, especially, when the values of the delays do not change frequently. Recently, a new uncertain-system-like approach was presented in [8], however, only stability results were presented.
