Networked control systems (NCSs) are spatially distributed systems wherein the communication among system components is supported by a shared communication network. Recently, NCSs have been successfully applied in a broad range of areas and attracted increasing attention in control community. However, the insertion of networks in control systems inevitably brings some new challenges, such as network-induced delays, packet dropouts, data quantization and media access constraints. Due to these network issues, the performance of NCSs will be degraded and systems may even become unstable. Therefore, it is of great theoretical and practical significance to develop more advanced approach to analysis and synthesis the NCSs taking into consideration those network-induced issues. This thesis focuses on several analysis and synthesis problems of NCSs with different types of network-induced constraints which have not been discussed in the open literature.
Firstly, we investigate fuzzy decentralized control of a class of large-scale networked dynamic systems with uncertain network-induced time delays, missing measurements and external disturbances. Nonlinear systems are modeled in Takagi-Sugeno (T-S) form. Bernoulli stochastic variables are used to describe the packet dropouts. A sufficient condition on stochastic stability with prescribed H∞ performance for the closed-loop control system is obtained by using the common Lyapunov function method. It is shown that the controller can be obtained by solving linear matrix inequalities (LMIs). Numerical simulations are finally provided to illustrate the effectiveness of the proposed results.
Then the H∞ fuzzy decentralized control problem for a class of NCSs with time delays, missing measurements and external disturbances is considered based on piecewise Lyapunov function approach, which is less conservative than the common one. Decentralized static output feedback controllers are designed so that the closed-loop NCS is stochastically stable with a guaranteed H∞ performance. It is also shown that the output feedback controllers can be obtained by LMI techniques. Simulation results are provided to show the effectiveness of this method.
Thirdly, the problem of fuzzy decentralized control for NCSs with time delays and packet dropouts via the delta operator approach is studied. Different from the shift operator which is used in Chapters 2 and 3, the delta operator method has advantages of unifying both continuous and discrete-time systems into a single framework, and is particularly effective when the sampling period is small. Fuzzy decentralized controllers are designed by using the Lyapunov functional and LMIs techniques so that the resulting closed-loop system is stochastically stable. It is noted that the sampling period here in delta operator is an explicit parameter which can be chosen accordingly based on networks load conditions or system requirements. Simulations are given to demonstrate the effectiveness of the control approach.
Lastly, finite-level logarithmic quantizers are first designed for NCSs with event-triggered mechanism (ETM). Dynamic quantizers are utilized in both sensor-to-controller (S/C) and controller-to-actuator (C/A) sides. The event-triggered method is adopted to reduce data transmission and energy consumptions. A sufficient condition is presented to guarantee the asymptotic stability of the closed loop system by using the Lyapunov functional method. The constructive design of the controllers is expressed in terms of linear matrix inequality (LMI). Then, network-induced time delays are taken into consideration based on the event-based model. These time delays are considered as long time delays which are more reasonable for practical applications. The considered networked dynamic system is modeled as a switched system with both stable and unstable subsystems. Criteria for exponential stability are presented for the closed-loop system via the average dwell time method and quadratic Lyapunov-Krasovskii functional. The state feedback controller is then designed by solving a set of LMIs. Finally, Simulation examples are included to illustrate the effectiveness of the proposed approaches.