The Stability And Quantitative Effect Analysis Of Networked Control Systems

With the development of technology and industry, the traditional point-to-point control systemshave been far from meeting the control requirements of industrial production process. With thepopularity of network communication in practical application, the networked control systems whichcombined with network and control systems were born. Compared with the traditionalpoint-to-point control system, the greatest feature of networked system is the usage of network toconnect the sensor, controller and actuator. There exist advantages of simpler structure, higherreliability, lower cost of wiring in networked control systems. It becomes easier for us to upgrade,maintain, manage and control the systems. However, the introduce of network also brings thephysical limitations of the network itself, like existing defects of network-induced delay, packetdropouts, information security and the limitation of channel bandwidth. Due to so many defects innetworked control systems, in order to apply networked control systems to the practical productioncontrol better, it is necessary for us to study the networked control systems.Since G.C Walsh first proposed the concept of networked control systems in1999, scholarshave spent a lot of effort in the study of networked control systems, and have gotten a lot ofachievements. Most scholars study networked control systems based on time-delay system model,quantization model, switching systems model and Markov chain model. The main research contentsare stability analysis, robust control, filtering, quantization control and so on. This dissertationfocuses on the stability analysis of networked control systems.Firstly, the concept, advantages, disadvantages, significance and research status are introducedin the introduction chapter. In the summary of the existing research results, we find that manyscholars ignore some effective information in the treatment of integral cross terms, which leads tomore conservatism of the obtained stability criterion. In addition, many scholars use the delaydecomposition method to analyze the stability of networked control systems, while few peopleapply the delay decomposition method to the study of networked control systems with quantizers.So it needs to be studied further that whether it can reduce conservatism of the result by using thedelay decomposition method to analyze quantitative effect of networked control systems. In thisdissertation, for the question given above, we discuss the stability of networked control systemsbased on time-delay model and quantization model.In the first part, we study the stability problem of nominal networked control systems withnetwork-induced delay and data packet dropout. By using the unequal delay decomposition method,the delay (t)is divided into1(t)and2(t).Based on Lyapunov-Krasovskii functional approach, a new Lyapunov-Krasovskii functional is constructed, and some valid items are retained which areneglected by many other scholars in the treatment of integral terms. Less conservative stabilitycriterions than the existing literatures are obtained by the application of convex combination andconvex optimization theory at last. And through the horizontal comparison between propositions inthis part, we can draw a conclusion: under the premise of ensuring optimum results, we should usethe stability criteria with the simpler form and the better computational efficiency one.Based on the previous research, the delay decomposition method is applied to stability analysisof networked control systems with logarithmic quantizers in the second part. By usingLyapunov-Krasovskii functional approach, a stability criterion is obtained. And a numericalexample is given to show effective of the obtained stability criterion. By comparing with existingliteratures, it proves that the obtained stability criterion which is based on delay decompositionmethod improves the existing results.The stability problem of networked control systems with dynamic quantizers is analyzed in thethird part. A less conservative criterion than the previous part is obtained by employing the delaydecomposition method and Lyapunov-Krasovskii functional approach. And it shows that theobtained stability in this dissertation improves the existing results by comparing to other existingliteratures through a numerical example.Finally, we summarize the research contents of this dissertation and the future direction whichneed to be studied further.