Research On Filtering And Control For Wireless Networked Quantized Systems

Wireless networked control systems (WNCSs), which are also called wireless network-based control systems, are the closed-loop control systems where the plant, sensor, controllerand actuator are connected through wireless shared network. Compared with the traditionalpoint-to-point control systems, the WNCSs provide more advantages such as low cost, reducedweight and power requirements, simple installation and maintenance, etc. Therefore, WNCSshave been widely applied in remote medical treatment, aerospace, industrial automation, mobilerobotics and so on. In addition, quantier become an indispensable part of the WNCSs due tothe bandwidth constraints. Thus, it is of great importance both in theoretical and practicalapplication to study the quantization effects. As a result, considerable attention has been painto the problem of quantized measurement control and estimation.In this thesis, we investigate the problems of network induced delay, data packet dropoutand quantization for the linear WNCSs. Two well-known WNCSs models, which are Bernoullimodel and delayed differential equations (sampling, delay, dropouts) model, are discussed. Thelogarithmic quantizer is introduced. Our purposes are to develop conditions such that the closed-loop systems are stable and satisfy corresponding performance indices, and for the singular sys-tems are also regular, impulse free. The main work of this dissertation are outlined as follows:1. The problem of H∞filtering for singular systems with communication delays is investi-gated. The information limitations, which include signal transmission delay, data packetdropout and measurement quantization, are analyzed. Based on the linear matrix inequal-ity (LMI) approach, a full-order filter is designed to cope with the aforementioned in-formation limitations such that the filtering error singular system is regular, impulse-freeand exponentially stable, and has a prescribed H∞performance as well. Sufficient con-ditions that guarantee the existence of the desired filter are also obtained. Two numericalexamples are given to illustrate the effectiveness of the proposed approach.2. H∞filtering problem for linear systems using quantized measurements is considered.The communication channel we consider encounters in two cases, they are the ideal oneand the unreliable one. For the ideal channel, we design a filter to mitigate the quantiza-tion effects, which ensures not only the asymptotical stability but also a prescribed H∞filtering performance for the filtering error systems. And then, for the unreliable channel,we introduce the stochastic variable satisfying Bernoulli random binary distribution tomodel the lossy measurements. We also design a filter to cope with the losses and mit-igate quantization effects simultaneously such that the filtering error system is not only stochastically stable, but also has a prescribed H∞filtering performance. Further, wederive sufficient conditions for the existence of the aforementioned filters, respectively.Finally, a numerical example is given to illustrate that the proposed approach is effectiveand feasible.3. We address the problem of a reset state observer-based control for linear systems usingquantized measurements. According to the characteristic of the logarithmic quantizer,a reset state observer based control is presented based on the standard one to suppresssensor quantization effects. By using the Lyapunov approach, the closed-loop system isstill asymptotically stable when the reset technique is introduced. The observer and con-troller gains of the closed-loop systems are obtained via solving linear matrix inequalities(LMIs). At last, a numerical example is given to illustrate the effectiveness of the pro-posed results.