Optimal Control Of Networked Control Systems With Transmission Delays And Packet Losses

A networked control system (NCS) is composed of multiple feedback control loops that share a serial communication channel. As networked control systems introducing network into a control system bring many advantages over conventional point-to-point control systems, for example:lower cost, simple installation and maintenance, and better system flexibility. However, data exchange via network inevitably results in transmission delay, transmission intervals, packet losses and variable sampling and so on, which will degrade system performance and even lead to closed-loop instability.This paper we focus on the transmission delay and packet losses, develop a method to approximate optimal control for nonlinear systems. Iterative methods are primary tools in ADP to obtain the approximate solution of the Hamilton-Jacobi-Bellman (HJB) equation. The main content of this paper is as follows:Firstly, summarize simply the research background and significance of this topic, the fundamental issues in the research and the present situation of the research. Then, this paper introduces some concepts of optimal control, adaptive dynamic programming and neural networks.Secondly, describe the issue of our research in this paper. The monotonic and convergence characteristics are analyzed. As we can not directly solve the optimization problem for nonlinear systems, we use adaptive dynamic programming algorithm to approximate of function structure. Lately, general algorithm design is given. Then, the optimal tracking control is analyzed by mathematical transformations, which change the tracking problem into optimal control problem. And convergence characteristics are analyzed.Finally, we summarize and prospect the full text, point out the contribution of this paper and the next research direction.This paper discussed optimal control of the nonlinear system with transmission delay and packet losses. We use adaptive dynamic programming algorithm to approxi-mate of function structure, as we can not directly solve the optimization problem for nonlinear systems. Numerical examples of later chapters show that ADP and neural network can solve the problem of networked control systems with transmission delays and packet losses.