Stabilityanalysis, Control And Filtering For Networked Systems

Computer network is the concentrated reflection of information technology, and constitutes the ecological environment of modern industry. Meanwhile, networked control systems (NCSs) have been applied in all fields of economic construction and national defense. In light of this, this dissertation is concerned with the analysis and synthesis for NCSs under present communication protocol and topology settings. Around the structure and basic problems existed in networked systems, the feedback models are established firstly, from which the asymptotical stability conditions of the closed-loop systems under particular performance indexes are analyzed by delay system approach; then on this basis, the stabilization problem, the quantized feedback control, the H_?filter design and H_?output tracking control problems for NCSs are synthesized.The main research works mentioned in this dissertation are listed as follows:1. With consideration of network-induced delay and nonlinear feedback disturbance in network channels, the absolute stability and control subject to sector-bounded condition for multi-input and multi-output (MIMO) NCSs are investigated by making use of time-delay Lurie system approach. Meanwhile, the maximum allowable delay bound for globally stable running of the closed-loop system is also obtained.2. With consideration of the effect of both network-induced delays and data packet dropouts, the robust H_?control problems of nonlinear networked systems which contain the rational nonlinear parameter uncertainties are discussed by utilizing a T-S fuzzy approach. It is worth noting that the every successfully transmitted date packet to the actuator will update the event driven ZOH (zero-order holder) once, and by limiting the ZOH's updating periods, the systems'stability can be guaranteed in the sense that at least one packet will be transmitted successfully as input signal in the prescriptive time. With consideration of these facts, the development and perfection of modeling method of closed-loop feedback control system based on network driving end is obtained theoretically.3. Considering the three important issues to be addressed for communication-based systems-signal transmission delays, data packet dropouts, and signal quantization, the quantized feedback control problems in four cases for networked systems is addressed with delay system approach. (1) Due to the data format conversion in different transmission media interface and the limited precision approximation ability of quantization process, the quantized feedback H_?control problem for networked systems with a dynamic quantizer on the sampler side is carried out. (2) For the wireless or low bandwidth network which is referred to as limited communication capacity system, the stabilization problem is investigated for the NCSs with two logarithmic quantizers as one is on the sampler side and another on the controller side. Moreover, the quantized feedback control criterion for the NCSs with one quantizer on the sampler side only is also deduced as corollary accordingly. (3) Owing to the digital communication characteristic of network, the state quantization feedback H_?control for nonlinear networked systems is proposed based on a T-S fuzzy approach, and in the course of this process, the modeling method based on the sampler end is perfected, following which it is proven that the network characteristic parameter reflected in the closed-loop feedback systems and the networked model obtained from the aforementioned driving-end-modeling method is equal in normal conditions. (4) Under the condition that not all states is measurable or available, the output quantization feedback stabilization problem based on observer for networked systems is derived, where a logarithmic quantizer on the sampler side is supposed.4. When the statistical characteristic of the external disturbance and the measurement noise is not identified, the H_?filters for linear and nonlinear networked systems are designed based on the filtering error system, respectively. According to parallel distributed compensation (PDC) strategy, the H_?fuzzy filter with the same fuzzy rules of the plant is designed based on T-S fuzzy approach5. By the augmented state matrix method, the network-based H_?output tracking control is concerned with simultaneous consideration of signal transmission delays, data packet dropouts, and signal quantization. Moreover, the utilization of piecewise delay system approach makes the less conservative for the existence of admissible H_?tracking controllers.