Sequential Inverse Covariance Cross-fusion Estimation For Stochastic Lag Systems

In recent years,with the wide application of multi-sensor information fusion in military and civil fields,the state estimation fusion of time-delay systems has attracted more and more attention.Due to the aging of system components or the existence of inevitable factors in the system itself,there is a time lag in the system.In the multi-sensor information fusion problem,the time lag of the system is common and inevitable.The transmission is not timely,which affects the evaluation effect of the system,and the uncontrollability of the time lag will cause the system stability to decrease,which ultimately affects the overall control effect of the system.In addition,due to the delay of information transmission and the complexity of time-delay system itself,the error cross covariance between local sensors is difficult to obtain.In order to reduce the large amount of calculation caused by the error cross covariance,we can apply(Sequential inverse covariance intersection,SICI)fusion algorithm is used to deal with the time-delay system with unknown cross covariance,it can significantly reduce the amount of calculation,reduce the requirements of storage space,and its fusion accuracy is higher than that of local sensor estimation,so it has good estimation performance.In this paper,based on the linear minimum variance theory,the multi-sensor stochastic time-delay system is theoretically derived,and the SICI fusion algorithm is used to estimate the state of the system.The main research contents are as follows:Firstly,the ICI fusion algorithm is applied to the multi-sensor information fusion based on the fusion idea of SCI fusion algorithm,and a sequential inverse covariance intersection(SICI)fusion algorithm is proposed.For the stochastic delay multisensor system with uncorrelated noise,based on the local steady-state Kalman filtering algorithm,the SICI fusion algorithm is used to fuse the local sensor information.The fusion results are compared with the SCI fusion algorithm and the matrix weighted fusion algorithm.It is concluded that the SICI fusion algorithm not only inherits the advantages of the small calculation amount of the SCI fusion algorithm,in addition,the accuracy of SICI fusion algorithm is higher than that of SCI fusion algorithm.Compared with the matrix weighted optimal fusion algorithm,SICI fusion algorithm is suboptimal,but its fusion accuracy is close to that of the optimal fusion algorithm.Secondly,the information fusion of multi-sensor system with correlated observation noise and correlated system noise is studied,and first,the original system is transformed into a non time delay system with uncorrelated noise by using decorrelation method and state dimension expansion method.Then,based on local steady-state Kalman filtering,the transformed system is transformed by using sici fusion algorithm,and the fusion estimation results are calculated with other fusion algorithms The results show that SICI fusion algorithm has good fusion accuracy and reduces a lot of computation.Finally,through Matlab simulation of 5 examples to verify the fusion effect of SICI fusion algorithm,and the simulation results show that SICI fusion algorithm is very effective for dealing with time-delay multisensor systems,and the algorithm not only approaches the fusion accuracy of the optimal fusion algorithm in terms of fusion accuracy,but also avoids the calculation of error cross covariance,greatly reducing the calculation amount.