The Application Of Nonlinear Filter In Underwater Maneuvering Target Tracking

Underwater maneuvering target tracking,a method for estimating the states of random underwater moving target,is widely used in both military and civil fields.Recently,with the constant development of science and technology,the requirement of underwater target tracking is growing,including higher accuracy,faster efficiency,and stronger robustness.In order to better play its application value in practical engineering.Firstly,this paper describes the background,purpose and significance of the research topic,and introduces the development history and research status of the filtering algorithms and the target motion models in the underwater target tracking system.Then,based on the general form of nonlinear Gauss filter,four nonlinear filtering algorithms are introduced,including Expended Kalman Filtering(EKF),Unscented Kalman Filtering(UKF),Cubature Kalman Filtering(CKF)and Gauss-Hermite Quadrature Filtering(GHQF).Among them,GHQF is a burgeoning nonlinear filtering method which has unique advantage because of its high accuracy and numerical stability,and it is also the focus of this paper.Through simulation analysis,the performance of GHQF is compared with the other three nonlinear filters,and the shortcomings of GHQF are discussed.After that,the novel filtering algorithm is improved in terms of efficiency and robustness.In view of the low efficiency problem caused by the “curse of dimensionality”,the “State-Space Partitioning” concept is introduced,and the Multiple Quadrature Kalman Filtering(MQKF)is obtained.This paper introduces the principle of MQKF,and studies the influence of “State-Space Partitioning” on the performance of MQKF by simulation experiments.The results show that MQKF can not only ensure the accuracy of the estimation,but also greatly improve the efficiency of the algorithm by selecting the appropriate “Subspace Partition Method”.In order to improve the robustness of MQKF,the Strong Tracking Filtering(STF)is introduced to the MQKF so that the STMQKF is obtained.The simulation results show that STMQKF is better than MQKF for the state estimation of the strong maneuvering targets.Finally,from the point of view of the target motion model,the target tracking algorithm is improved again.Based on the theory of Multiple Model(MM),this paper focuses on the Interacting Multiple Model(IMM)algorithm.Then,the STMQKF is embedded into the framework of the IMM algorithm,and the IMM-STMQKF algorithm is obtained.According to the simulation example,the estimation accuracy of IMM-STMQKF is higher than that of CA-STMQKF and IMM-UKF.It shows that IMM-STMQKF has a certain theoretical significance and application value in the field of underwater maneuvering target tracking,because of its high accuracy and matching of the models.The main work of this paper is to propose a new nonlinear filtering algorithm named STMQKF with high efficiency and strong robustness,which is based on GHQF.And apply the IMM-STMQKF algorithm to underwater maneuvering target tracking.In the end,the new algorithm achieves good results.