Analysis And Synthetic Study For Net-worked Systems Using Quantized Measurements

Networked control systems(NCSs), are the closed-loop control systems and result of the tech-nology of network communication and control, where the plant, sensor, controller and actuator are connected through shared network. The NCSs have been widely applied in such as remote medical treatment,aerospace, industrial automation and so on because of its low cost, simple installation and maintenance. Although NCSs provide so many advantages, it also brings new challenge and diffi-cult problems. Because of the data-loss in the communicated channel, specially when net jam come and data collision in the interface. For some digit signals are not proper to be coded, then we bring quantizer to our NCSs which also brings quantized error, and the main purpose in our research is research of the two problem of quantized error and data-loss. Thus many method is provided to dis-cuss the problems, the first is to view the packet losses as a binary switching sequence which obeys a Bernoulli distributed white sequence taking on values of zero and one with certain probability; The second approach is to use a discrete-time linear system with Markov jumping parameter to represent random packet-loss model for network.there are also approaches for dealing with the quantization error, the first is to translate the logarithmic quantization error to robust problem, the second is that linear quantization error can be translated to solve error variance which is minimum.In this thesis, by means of strict LMIs, we study the stability and stabilization for networked systems, including network-delay, data dropout and data quantized, while data quantized consist of logarithmic one and linear one. We have two models which is Bernoulli and Markov models, and the main work of this dissertation are outlined as follows:1. The problem of the design of state observer is considered for networked control systems using linear quantization in this paper. The observer we consider encounters in two cases, this is, the quantized measurements and data-loss, which occurs when the quantized signal is transmitted in the unreliable channel. To cope with the quantization error and the loses, the observer we design is to guarantee not only the bounded-input and bounded-output stability, but also the optical gain L,for the closed-loop systems.Finally, the numerical examples illustrate that the proposed approach is effective and feasible, and show the relationship between the quantized density, the probability of event packet-loss and the gain matrix L.2. The problems of stability and stabilization for networked time-delay systems are studied. During our research two cases must be solved,that is,the quantized measurement and date-loss,which occurs when the quantized signal is transmitted in the unreliable channel. To cope with the quantization error and the loses, the observer and controller we design is to make the closed-loop systems asymptotic stability. Finally,the numerical examples illustrate that the proposed approach is effective and feasible,and relationship between the quantized density, the probability of event packet-loss and the control gain matrix is shown.3. We study the effect of communication packet losses and quantized error in the feedback loop of a control system.The sensor-to-controller(S-C) and controller-to-actuator(C-A) random net-work date losses are modeled as Markov chains, and the resulting closed-loop systems are jump linear systems with two models. The goal is to find a observer and a controller such that the closed loop is mean square stable for a give packet loss rate and quantization error. Also, a numerical example is provide to demonstrate the effectiveness of the proposed method.