Research On Target Tracking Based On Non-cooperative Passive Radar

Passive radar has become an important development direction of modern radar technology because of its strong anti-stealth characteristics,long detection range,low altitude detection capability and so on.In the passive radar technology,usually using the direction of the angle,delay,Doppler shift and other measurement information to achieve the target positioning and tracking.Different from the active radar,because the target positioning equation is nonlinear,the positioning accuracy is low,the uncertainty of the measurement data and the target state association is high,which makes the research of the passive radar target location tracking technology important.First in this thesis,the method of positioning and positioning accuracy of passive radar is analyzed in detail.After selecting the joint positioning method based on arrival angle and delay information,the particle filter(PF)technique is used to solve the single target tracking problem.Density particle filter algorithm(P-PHD)is adopted to solve the multi-target tracking problem.Finally,the use of Matlab GUI design to achieve singlestation passive location tracking simulation module for system validation.The main work and innovations of this paper are summarized as follows:A study on the Selection of the Precondition-Location Method for Target Tracking.Firstly,the typical positioning method of passive radar is introduced.Then,the influence of different positioning methods,different positioning systems and different transceiver stations on the positioning accuracy of passive radar targets is analyzed in detail by simulation experiments.Finally,we consider the target localization method suitable for passive radar and regard it as the observation equation of the follow-up filtering algorithm.Aiming at the single target tracking scene of passive radar,the single target tracking technique based on particle filter is analyzed.(1)Combining the particle filter algorithm,a new passive radar tracking algorithm is proposed for the traditional nonlinear tracking method in the passive radar positioning and tracking system with poor stability and low estimation accuracy.(2)To further improve the tracking accuracy of the passive radar to the target,the PF algorithm is improved.Based on the Cubic Kalman Filter(CKF)algorithm,the particles after the system resampling are operated to ensure that the sampled particles move to the high likelihood region,so as to improve the accuracy of the PF algorithm and operational efficiency.Aiming at the multi-target tracking scene of passive radar,the probability-assumed density particle filter(P-PHD)algorithm is studied.(1)In this thesis,the ensemble Kalman filter(EnKF)algorithm is used in the basic P-PHD algorithm to solve the problem of missing particle diversity caused by the introduction of the current measurement information in the surviving particle sampling part of the basic P-PHD algorithm.To generate the proposed distribution and use the likelihood function to select the appropriate measurement point in the current measurement set for the correction of the particle state in the EnkF algorithm.(2)A PHD-M algorithm based on Measurementdriven is studied.The problem of estimating the number of new targets in this algorithm is improved.Finally,in order to verify the effectiveness of the proposed algorithm,a simple simulation system for single station passive location tracking filtering is designed by Matlab GUI,and the system is used to verify the effectiveness of the proposed algorithm.