The Research On Track Before Detect Algorithms For Weak Targets In The Sky Wave Over The Horizon Radar

Weak targets detection and tracking is one of the huge challenges in the radar processing system.The traditional detect algorithms based on the single frame data can hardly give out the reliable detection results under low signal to noise ratio conditions.But track before detect(TBD)technology can solve this problem.The TBD algorithms based on particle filter(PF-TBD)are researched deeply and widely for solving the nonlinear and non-Gaussian questions.But the researches on radar domain are still in the primary stage.Therefore,this paper that aimed at the weak targets detections of the sky wave over-the-horizon radar(SKW-OTHR)researches the PF-TBD algorithms.Firstly,this paper based on the model of the SKW-OTHR researches the state vector of the target and the system measurement model.Based on the characteristics of the SKW-OTHR,the state vector is chose reasonably with range,radial velocity and amplitude.This paper chooses the range-doppler(RD)pectrum which is the result of radar echo processing as the point spread model.Then,the formulas for calculating the weight of particles in the process of particle filter are given in the background of Gaussian complex noise.After a variety of PF algorithms are studied.This paper studied emphatically basic particle filter algorithm based on sequential importance resampling(SIR),auxiliary particle filter(APF)algorithm and cost reference particle filter(CRPF)algorithm.The APF algorithm can solve the problem of particle degradation in the resampling algorithm by means of auxiliary variables.The CRPF algorithm does not need to obtain the prior information of the noise distribution of the dynamic system,and estimate the target state through the risk and cost function.Then,the corresponding PF-TBD algorithm is studied for various PF algorithms,and an improved particle filter TBD algorithm(ISPF-TBD)is proposed based on the standard particle filter(SPF-TBD)algorithm.The algorithm can solve the problem that the target probability curve is slow to converge due to the low target probability of the newborn particle hit in the initial stage.Finally,the PF-TBD algorithm for multiple target is studied.Multiple mode multiple target PF-TBD(MM-MT-PF-TBD)is studied in this paper.The algorithm based on a three-layer particle filter algorithm recursively estimate the existence probability of the target and the target state,which is based on the possible situation of all targets.Simulation results show the effectiveness of the proposed algorithm.