Multiple Target Tracking Algorithm Research Before Testing

With the development of aircraft stealth technology, the RCS of targets become smaller and smaller. So the problem of weak target detection is becoming more and more important. It is an effective way of Track-Before-Detect Technology to deal with the problem of weak target detection. In practical, multiple targets often simultaneously exist, therefore it is necessary for us to do the reaseach of multi-target track-before-detect technology.The dissertation focuses on multi-target TBD method and knowledge base multi-target TBD technology. It mainly comprises:1) For solving the problem of unknown initial velocity of targets, the detection and tracking performance of DPA in defferent state space is analyzed. For solving the problem of unknown number of targets and huge interference of adjacent targets, a SIC-DPA is proposed, by which not only the number of targets can be estimate exactly with lower SNR, but also adjacent targets can be distinguished better. For solving the problem of high computation berden an improved angorithm is porposed, by which the states need to be searched are tremendously reduced, computation complexity is nearly independent of the number of targets.2) For solving the problem of how to use the priori knowledge of targets with constant amplitude, Likelihood ratio accumulation method is used to improve the performance of muti-target TBD. For the targets with fluctuate amplitude, a model of fluctuate amplitude is proposed. Morever, kalman filter is used for the amplitude estimation. The detection and tracking performance of targets with deferent kind of amplitude is impoved by using priori knowledge.3) For solving the problem of using priori knowledge of the road. The search erea of the targets on the road is greatly readuced. For the aircraft near the road, it comes to be high clutter eara, a method is used to make road clutter zero and do compensation. Detection performance is improved. For using the priori information of clutter figure, the merite function in deferent ereas is accumulate deferently, in this way the influence of the strong clutter is greatly reduced. The effectiveness of the methods above is proved by simulations. Compared to the TBD algorithms before, the multi-target TBD algorithms and knowledge based multi-target TBD algorithms proposed in this dissertation, multi target can be better easily tracked and with better detetion and tracking performance.