Research Of Underwater Targets Track-before-detect Algorithm Based On Passive Array

With the development of modern ships,deep diving equipment and other underwater equipments,the water area could not be detected in the past gradually entered human's vision.The ocean has been paid more and more attentions for its rich resources.With the constant competition of the countries in the world,it's urgent to improve the ability to protect the rights and interests of marine environments.Passive sonar has been focused by the scientific personnel for its high covertness.Influenced by the characteristics of passive sonar systems,the received signals are more likely to be low signal-noise-ratio signals.The traditional tracking methods do not have sufficient performance to process these signals.However,the track-before-detect algorithms for passive sonar systems still need to be developed.Aiming at the problems above,the particle filtering track-before-detect algorithm based on passive array sonar has been studied in this paper.The main contributions of this paper are as follows:1.Aiming at the problem that the study of passive sonar system target detection and tracking algorithm is insufficient,the basic model of passive sonar system and the basic theory of multi-target particle filter algorithm are studied.A suitable multi-target tracking framework and array signal model are established,which laid a theoretical foundation for the algorithm of this paper.2.Aiming at the low SNR problem of passive sonar target,a particle filtering track-before-detect algorithm based on passive sonar array with dual-parameter likelihood for underwater acoustic target has been proposed.The algorithm establishes the target motion and measurement model of passive sonar bearing-only tracking,designs the dual-parameter spatial spectral likelihood by using minimum CV distance criterion and realizes the track-before-detect for passive sonar targets.The effectiveness of the proposed method are demonstrated by numerical experiment.3.To solve the problem that the prior information of the passive sonar targets is unknown and the number of targets is time-varying,a particle filter track-before-detect algorithm based on weight hypothesis test is proposed.The algorithm analyzes the priori dependence problem of the proposed PF-TBD algorithm for passive sonar.The target “start” and “end” particle filter models based on the weight hypothesis test are established by studying the characteristics of target and clutter in the spatial spectral likelihood function.The algorithm has gained the ability to handle the priori unknown targets.The advantages of tracking and real-time performance of the algorithm are illustrated by comparing proposed algorithm and traditional tracking method.Finally,the validity and practicability of the algorithm are further verified by the real data.