Multi-target Tracking Method With Passive Multistatic Radar

Passive Multistatic Radar(PMR)uses the commercial or military transmitter as the opportunity illuminator to detect and track the target by coherent processing of the direct signal and the target echo signal.It also known as Passive Bistatic Radar(PBR)when a single transmitter and a single receiver are used.Compared with the active radar,PMR has the advantage of high security and low cost.It can use the diversity of spatial distribution and frequency to improve the detection performance of stealth targets.The PMR has important military value and scientific research significance and attracts more and more attentions.In this thesis,we focus on the problem of low observable multi-target tracking with PMR.The main contributions are as follows:1.Aiming at the problem that the low accuracy of multi-target association in PMR leads to the decreasement of tracking accuracy,a joint probabilistic data association(JPDA)based on Kullback-Leibler Divergence(KLD)is proposed.First,the posterior probability density function of the associated event,and the KLD between the function and the single Gaussian probability density function are calculated,respectively.Secondly,KLD is used as a cost function to optimize the posterior probability density function of related events.Finally,the target state is estimated according to the optimized posterior probability density function.The simulation results show that compared with traditional data association algorithms,this algorithm can effectively improve the accuracy of target tracking.2.In order to balance the tracking accuracy and computational complexity when the number of targets is unknown,a second-order Gaussian mixture probability hypothesis density(SO-GMPHD)algorithm is proposed.First,the target prediction process is changed from the Possion point process to the Panjer point process.Secondly,the variance of the number of targets is used to propagate the second-order moment.Finally,the state and number of targets are estimated by second-order statistics.The simulation results show that the proposed algorithm can achieve acceptable tracking accuracy and computational complexity simultaneously.3.Aiming at the problem that the tracking performance of PMR with bistatic configuration degrades under low detection probability,a covariance interaction(CI)based adaptive SO-GMPHD multistatic fusion algorithm is proposed.First,the adaptive birth Gaussian component estimation method is used to extract the possible new target measurements from total measurement set,and the birth Gaussian component is updated with the target set estimated by the measurement set.Then,the Gaussian component of each receiver is updated using SO-GMPHD to obtain the local posterior intensity.Finally,the local Gaussian component is used to obtain the state and number of targets with CI.Simulation results show that the tracking accuracy of the proposed algorithm is improved under the low detection probability.