Estimation And Optimal Control Of Network Control Systems With Packet Loss

The study on networked control systems(NCSs)is one of key research areas in control theory,which has a broad range of applications in communication,economics,aerospace and other fields.However,the phenomenon of packet losses will occur in the information transmission of the network control system,which may inevitably lead to the performance degradation or seriously lead to the instability of the system.Therefore,this paper studies two kinds of networked control systems with packet losses by using probability theory,optimal filtering and optimal control as the basic knowledge,and analyzes the corresponding estimation and output feedback control problem,which is conducive to improving the performance of the system.The main contents and results are as follows:1.First of all,the background and research status of networked control systems are introduced.Then the current achievements and research directions for estimation and optimal control problems of NCSs with packet losses are summarized.Finally,the basic theoretical knowledge and the existing ideas and methods for solving the control problem are expounded.2.For networked control systems with Bernoulli packet losses,the estimation is studied according to different measurement equations.On the one hand,the optimal estimation is derived strictly by a recursive method.On the other hand,in order to facilitate the practical application,a sub-optimal approximate estimator is developed and compared with LMMSE.It is verified that the performance of the sub-optimal estimator proposed in this paper is better.3.For the networked control system with Markovian packet losses,the basic linear quadratic optimal control problem is solved.Firstly,some preliminary results are given,and the optimal recursive estimator and the error covariance matrices are designed.Accordingly,the value function is defined.The optimal control law and the related value function are derived by using the method of dynamic programming.Finally,we show that the optimal controls are given in terms of a modified Riccati equation,which can be calculated backwardly under the basic assumption made in this paper.The obtained results can be regarded as an important implementation of NCSs with Markovian packet losses.