Multi-sensor Under-observation System Fusion Self-correcting Incremental Kalman Filter

Due to the development of high-speed electronic computers and the need to determine technical problems such as satellite orbits and navigation,the theory of multi-sensor information fusion filter has received widespread attention.In order to seek an estimate of the unknown real signal or state from the polluted observation signal,a series of Kalman filter algorithms have been proposed to solve such problems,and its real-time recursive property is superior to other general filte algorithms.However,due to the complexity of the system model,uncertain factors in the working environment,and errors in the measurement equipment itself,the system model has unknown and uncertain parameters,or the measurement system error occurs,that is,a poor observation condition.Using the traditional Kalman filtering algorithm directly will cause the filter performance to deteriorate,and even cause the filter to diverge.The self-tunning fusion filter and incremental Kalman filter are proposed to solve the above problems.This paper is based on the existing incremental filter model,and conducts related research on multi-sensor weighted fusion self-tunning incremental Kalman filter algorithm.The main contents are as follows:Firstly,new parameter estimation algorithms are proposed,including the recursive instrumental variable(RIV)algorithm based on the incremental observation model,the recursive extended least squares(RELS)algorithm based on the incremental observation model,and two-stage recursive least squares-recursive extended least squares(RLS-RELS)based on the incremental observation.The algorithm can solve the situation that the model parameters are unknown in the under-observed system,and effectively identify the unknown model parameters.Subsequent simulation examples verified the feasibility of the algorithm.Secondly,referring to the new parameter estimation method based on the incremental model proposed in Chapter 2 and combining with the existing incremental Kalman filter,the self-tuning incremental Kalman algorithm for single-sensor under-observed system is proposed.The dynamic error system analysis(DESA)method proves that the incremental Kalman filter is effectively solved when the model parameters are unknown.Subsequent simulation examples verified the feasibility of the algorithm.Finally,in practical applications,the environment is complex,and target recognition and control cannot be completed by a single sensor.At the same time,in order to further improve the accuracy of state estimation of under-observed systems,multi-sensor weighted state fusion and weighted observation fusion self-correcting incremental Kalman filter are proposed respectively.Subsequent simulation examples verified the feasibility of the algorithm.