The Study Of Weak Target Track Before Detect Algorithm Basing On Dynamic Progrmming

Track Before Detect(Track-Before-Detect,TBD)technology is the important means of radar detecting enemy's situation in the international war.The radar that working principle is largely based on track-before-detect method is to track the movement of the object.Whether Radar can detect enemy's attack in time is very important to have enough time to prepare,take advantages of combat,and even win the war.This paper mainly studies the detection before tracking algorithm based on dynamic programming,the main contents are:(1)It has introduced the traditional Dynamic Programming track before detect(Dynamic Programming Track-Before-Detect DP-TBD)algorithm which is used for tracking the point target,moving or curvilinear moving under the background of gauss.(2)Put forward improvement of the Dynamic Programming.When radar detect target under low SNR environment,the testing time delay,even missing target problem comes out,this paper proposes a dynamic programming algorithm based on the direction weighted and displacement weighted.Make use of the previous target location information from former four frames detected to design a current direction probability density,at the same time,according to the distance from candidates target which coming from candidate status gather to the former target state,design a displacement function to inhibit or enhance,finally constructed a kind of stage value function based on the direction weighted and displacement weighted.This method shorten the time delay of the algorithm which detect and track moving target under low SNR,and detection ability of the algorithm has been improved,as well as tracking performance and detection capability.Last,design a threshold of the algorithm and combine with two threshold,the point which energy is relatively weak will be out,and computational complexity of the algorithm is largely reduced.(3)In this paper,use Matlab software to simulate the designed algorithm.There are three parts in the system: noise modeling,target motion trajectory modeling,and algorithm to detect and track.By changing the target trajectory and signal-to-noise ratio,The tracking and detection performance of the algorithm has been checked.