With Time Delay And Packet Loss System Information Fusion Estimation

Due to the variability and unreliability of the transmission, time delay andpacket dropout are prevalent in the real systems, which cause the control instructionlack of efective implementation and control input and control volume can not updatein time. When seriously, the system may lose stability. At the same time, they alsobring trouble to the controller design. So the study of system with time delay andpacket dropout constitutes urgent problem in net control systems. The studied maincontents are as follows:Firstly, the linear discrete-time stochastic system with random measurement de-lays and multiple packet dropouts was studied. Based on the projection theory, thesystem is transferred into that without delay by using state augmentation method.Against the new system, the linear minimum variance optimal kalman estimatorsare designed for single sensor system. Furthermore, a sufcient condition for the ex-istence of the steady-state flter is given and the asymptotic stability of the optimalflter is analyzed. For multi-sensor systems, the cross-covariance matrix betweenany two sensor subsystems is derived. At last, the distributed linear minimum vari-ance optimal weighted fusion estimators are obtained by using distributed optimalweighted fusion estimation algorithms in the linear minimum variance sense.Secondly, based on the projective theory, the linear minimum variance optimalkalman estimators are presented for the linear discrete stochastic system with ran-dom delays and multiple packet dropouts in control input and observation equationby using state augmentation method.Further, the cross-covariance matrix betweenany two sensor subsystems is derived for the multi-sensor systems. At last, the dis-tributed linear minimum variance optimal weighted fusion estimators are derived.Then, for the linear discrete stochastic system with fnite-step delays in stateequation, frstly, the linear minimum variance optimal kalman estimators are de-signed for single sensor system by using state augmentation method. Further- more,the cross-covariance matrix between any two sensor subsystems is derived forthe multi-sensor systems. At last, the distributed linear minimum variance optimalweighted fusion estimators are designed by applying the weighted fusion estimationalgorithms in the linear minimum variance sense. In addition, the distributed linearminimum variance optimal weighted fusion estimators are obtained for the systemwith fnite-step delay in state equation and with time delay and packet dropout inobservation equation and the system with fnite-step delays in state equation andwith time delays and packet dropout in control input and observation equation.At last, we studied the system with fnite-step delay in state equation and multi-step random time delays in observation equation. Local estimators were studiedbased on the projection theory. Then, we studied the cross-covariance matrixsbetween any two sensor subsystems. Then, the distributed linear minimum varianceoptimal weighted fusion estimators are obtained by applying the weighted fusionestimation algorithms in the linear minimum variance sense.