A Multi-Radar Single-Target Based On Particle Filter For TBD Algorithm

Track before detect(TBD)based on particle filter(PF)is a very suitable method for detecting the weak signal,this method reduces the detection threshold for the target to track the target,then the data obtained by the sensor is transmitted to the fusion center for processing to obtain the target track,and at the same time to complete the detection of the target,so utilizes the prior information sufficiently.PF-TBD can not only deal with the detection and tracking of weak targets in Gauss system,but also deal with the problem of weak target detection and tracking in nonlinear non Gauss system,which has wide application scope.However,existing TBD methods mainly track and detect targets on the basis of single radar,there is little research on multiple asynchronous radars,and multiple asynchronous radars can make full use of effective information through joint detection to improve tracking detection performance.However,multiple asynchronous radar systems have the problem of different radar geographic locations and different sampling times.They cannot directly fuse the radar echo amplitude information,otherwise it will seriously affect the fusion results.In order to solve the problem of track before detect for multiple asynchronous sensors,the main contributions are given as follows:Firstly,the theoretical principle of PF-TBD is introduced and the detailed implement of PF-TBD algorithm is summarized.Introduced three common basic theories of time registration methods.Secondly,a multiple asynchronous sensor PF-TBD algorithm based on particle filter is proposed.This algorithm is aimed at the problem that multiple asynchronous radar echo amplitude information cannot be directly fused,the state of the particle is pushed back by the formulas of interpolation method to obtain a virtual particle,then calculating the weights of the virtual particle states and selectively combining radars,the problem of serious deviation of a radar's detection performance at a certain point of time is solved.Simulation results verify the effectiveness of the proposed method.Finally,a Quasi Monte Carlo Merging Resampling Intelligent PF-TBD method is proposed.Based on multiple asynchronous sensors,the algorithm uses Halton sequences to generate initial particle groups,and then in order to improve the diversity and effectiveness of the particles,the particle sets of less than the threshold value are treated with cross and mutation.The merging resampling method is used toresampling the particles,which effectively increases the diversity of particles and inhibits the degradation of particles.