Research On Track-Before-Detect Algorithm Of Maneuvering Target Based On Particle Filter

In order to solve the problem that Particle Filter Track-Before-Detect(PF-TBD)can not get better detection and tracking performance due to particle degradation,many scholars have proposed a variety of improvement methods,and to improve the processing ability of PF-TBD to mobile targets,Track-Before-Detect Based on Multi Model Particle Filter(MMPF-TBD)algorithm is proposed,however,the algorithm has the defects of low probability of using effective model.Based on the theory of Particle Filter(PF)and Multi Model(MM),this thesis takes the MMPF-TBD algorithm as the main line,proposes the following three algorithms respectively from the two aspects of optimizing filter method and optimizing multi model structure.The main research work and achievements are as follows:(1)Research on Auxiliary Particle Filter Track-Before-Detect Based on Optimized Genetic Resampling(OGRAPF-TBD)algorithm.In order to solve the problem that Auxiliary Particle Filter Track-Before-Detect(APF-TBD)can not achieve good detection and tracking performance due to the lack of particles,OGRAPF-TBD algorithmis proposed.In the resampling of APF-TBD,the Optimized Genetic Resampling(OGR)method is applied to select high-quality particles according to the weight,and new particles are obtained by sorting grouping crossover and mutation.This method not only retains the advantage of APF-TBD in improving the accuracy of sampling particles by optimizing the importance distribution function,but also introduces OGR idea into resampling,which can effectively solve the problem of particle shortage and increase the number of effective particles.Simulation results show that compared with APF-TBD and PF-TBD,OGRAPF-TBD has higher target detection probability and tracking accuracy,and stronger applicability.(2)Research on Multi Model Auxiliary Particle Filter Track-Before-Detect Based on Optimized Genetic Resampling(MMOGRAPF-TBD)algorithm.In order to solve the problem that MMPF-TBD can not detect and track maneuvering targets well due to the use of Sequential Importance Resampling(SIR)filtering method,a MMOGRAPF-TBD algorithmis proposed.By replacing SIR with OGRAPF-TBD,MMPF-TBD is improved from the perspective of optimal filtering method,which ensures the effectiveness of particles and solves the problem of lack of particles,making the weighted average particles closer to the real state of the target.Simulation results show that compared with MMPF-TBD,MMOGRAPF-TBD has higher effective model utilization,detection and tracking accuracy,and more stable detection and tracking performance.(3)Research on Optimized Multi Model Genetic Resampling Auxiliary Particle Filter Track-Before-Detect(OMMOGRAPF-TBD)algorithm.Aiming at the problems of MMOGRAPF-TBD,such as the motion model can not cover the motion form of high maneuvering target,the conversion between models is complex and the probability of using effective model is low,this thesis proposes OMMOGRAPF-TBD algorithm.Taking changing the value of maneuvering acceleration as the core,a variety of motion model is used to simulate multiple motion models,which not only increases the types of motion models,but also avoids the conversion between multiple models.The simulation results show that compared with MMPF-TBD and MMOGRAPF-TBD,OMMOGRAPF-TBD has higher use efficiency of motion model,detection and tracking accuracy,and stronger applicability.