Research And Development Of Multiple Target Tracking System Based On Passive Location And Aircraft Broadcast Information

With the development of science and technology,a measuring device can receive data from multiple targets at the same time.Therefore,most of the existing research in the target tracking field focus on the study of Multiple Targets Tracking(MTT).That has been widely used for military and civil,such as autonomous vehicles,missile tracking and warning monitoring.This thesis is mainly about the research and implementation of an MTT system based on passive radar.First,the algorithm Fast Density Peak-based Clustering(FDPC)is studied for the extended targets existing during the passive radar measuring.For the disadvantage that the FDPC algorithm cannot distinguish between adjacent targets when the target is close to the target in the situation such as formation flying,the Ellipse Threshold-FDPC(ET-FPDC)algorithm is proposed.ET-FPDC sets up the target motion elliptical threshold by carrying a measurement point ADS＿B.The measurement points within the ellipse threshold are clustered.Then the secondary clustering is performed by the FDPC algorithm.Aggregate extension targets in a two-step clustering fashion.Experiments in the system show the ETFDPC algorithm has a better clustering effect on extended targets with the help of ADS＿B information.Experimental results show that the rate of missed detection of ET-FDPC decreased about 75%.the false alarm rate decreased about 48%.Second,to alleviate the measurement and environmental noise in the MTT process,this thesis studies the algorithm system of Kalman Filter.Experimental results show the volumetric Kalman filter has the optimal filtering effect in the nonlinear system.There is no particle degradation problem of the traceless Kalman filter.For cases where traditional single-model filtering algorithms cannot track maneuvering targets.This thesis also studies the interaction multimodel theory,combines interactive multimodel theory and volumetric Kalman filtering algorithm.It gives the derivation and calculation process of volumetric Kalman filtering algorithm based on interactive multimode.Simulation shows the algorithm has a better tracking effect while tracking maneuvering targets.In addition,to solve the problem of track formation in multitarget tracking,the thesis studies the trajectory initiation and trajectory association algorithm.In the dense multiecho environment,the nearest neighbor data correlation algorithm,probability data correlation algorithm and joint probability correlation algorithm are studied.For the above three correlation algorithms,only the limits of one measurement threshold data are considered.Study multihypothesis target tracking theory.It calculates complex scenarios for trackoriented multihypothesis target tracking algorithms.It reduces the number of hypothetical tracks by settling multiple track thresholds.Combined with the ADS＿B information,the judgment method of track confirmation and deletion is given.Besides,it reduces the computational complexity and the lag period of the formation trail of the multihypothesis target tracking algorithm.Systematic experiments show the improved multihypothesis target tracking algorithm is basically unchanged compared with the original algorithm.On average,the confirmed track lag period decreased by 30% and the calculation time decreased by 67%.Finally,based on all of those studies,the above algorithms are integrated into the research and development of a multiple target tracking system based on passive radar.This system can locate,track and display targets within the detection range.And it also completes the task of posture monitoring within the detection range.The system has been developed,tested,and delivered for use,and it has been running stably for three months.