Control Theory Research Of Reset Quantized Station

With the network communication techniques developing, networked control systems (NC-Ss) appeared. The NCSs is mainly consists of plant, sensor, controller and actuator in sharingnetwork. However, the shared-network system is limited by communication bandwidth, carry-ing capacity and service capacity.The quantization as a encoder becomes an indispensable part of the NCSs to deal with thebandwidth constraints. Quantized encoder can solve bandwidth limitation problem, simple inconstruction and easy to manufacture. However, the encoder brings two effect in system. Oneis called saturation, the other is called degrade the achievable position accuracy. Facing to thedegrade the achievable position accuracy, we can solve it in two pars. One way is that choosinghigh performance sensor quantization to replace the old one. Another cost effective way is toremove the quantization noise by some numerical algorithms. Reset quantization state is a newsolution to suppress the quantization effects.In this thesis, we study the network control problems which based on reset quantization.From easy to complex, we divided our research to three parts. The first part, we study quanti-zation effects and loss measurements effect. The second part, we consider the effects of quan-tized jump for linear system. The third part, we design the controller to suppress effect of thequantized jump. According to logarithmic quantizer, Bernoulli model, probability theory and s-tochastic processing, we proved the designed closed-loop systems are asymptotically stable andsatisfy corresponding controller performance indices. The main work of this thesis is outlinedas follows:1. In the practical networked control systems, we design a controller to reduce the quantiza-tion effects and data packet loss of influence. Using the Lyapunov approach, we get theclosed-loop system satisfy asymptotically stable and also satisfy the H_∞performance.Finally, a numerical example is given to prove the effectiveness of proposed system.2. In networked control system, there exists the phenomenon of quantitative data jumpingwhen the data is quantized in the segment which caused by low accuracy of controller orall kinds of noise. We provide a reset state observer for linear systems to suppress effectsof quantitative data jumping. Using the Lyapunov approach, we can prove the closed loopcontrol system is asymptotically stable. Finally, a numerical example is given to provereset state observer is better performance than traditional observer.3. The quantitative data plays an important role to deal with network control system with to suppress the jumped quantitative data effects. Firstly, a reset state observer based onthe standard quantizer is introduced to guaranteed the performance of the closed loopcontrol systems when the quantized jumping occurs in the segment. Secondly, combingthe probability theory and stochastic process with character of reset quantized state, areset quantized state controller is introduced to improve the performance of the system,which leads to the fact that the reset system is still asymptotically stable. Furthermore,the controller gain can be obtained by solving a class of convex linear matrix inequalities.At last, a numerical example to be illustrated the effectiveness of the proposed approach.