Research On Track Before Detect Technology For Multiple Sensors Based On Dynamic Programming

When long-range early warning radars detect the targets above ground,above the sea or under the sea,weak targets are always covered by strong clutter.Moreover,as the development of aircraft invisible performance and maneuverability,the early warning time and discovery probability is decreased in a large extent.It enhances the breakthrough probability of air defense system.Considering the problems of weak targets tracking,there are main two solutions.The first method is Track Before Detect.By multiple frames accumulation,it enable to increase the tracking performance.The second way is using radar network to track targets.By disposing multiple radars in a proper way,it can increase space gain.In this paper,we propose a DP-TBD algorithm aiming for multiple sensors.Through using the advantages of track-before-detect and multiple sensors fusion,the purpose is to increase weak targets tracking performance.We consider DP-TBD algorithms for various sensors in this paper and propose centralized and distributed fusion methods respectively.The main work is listed below:1 Considering the tracking problems of multiple sensors centralized fusion,we raise a centralized asynchronous fusion algorithm for sensors with different resolution.A different resolution sensors alignment method and an adaptive change of transition area method are proposed.In this way,different resolution centralized fusion can be realized and the tracking performance of weak targets can be enhanced.2 In view of the tracking problem of multiple sensors distributed fusion.In this paper,we raise a recursive filter distributed fusion method which is based on DP-TBD.Via recursive filter and asynchronous fusion,it is able to significantly improve weak targets tracking results,decrease track loss rate and improve system robust.3 Considering the detection problems of multiple sensors distributed fusion,we propose a particle filter fusion algorithm which is based on DP-TBD.By particle filter and asynchronous conversion methods,it realizes multiple sensors asynchronous fusion.Comparing to recursive filter algorithm,this method can dramatically increase the tracking performance of weak targets.All these methods proposed above are demonstrated by simulation verification.In addition,the proposed method have compared with tradition algorithms.All the results prove the efficiency of multiple sensors fusion in weak targets tracking field.