Research On Multiple Targets Tracking And Particle Filter Algorithm

With the fast development of computer and control technology,particle filter has become an effective method for studying the optimal estimation problem of nonlinear and non-Gaussian system,it is also widely used to deal with the target tracking problems.In this paper,the related problems of multi-target tracking based on particle filter algorithm are studied,which are as follow:First,aiming at the problem of target tracking in the complex environment with glint noise,an improved interactive multiple model particle filter algorithm is proposed on the basis of traditional filtering technology,combined with particle filter and interactive multi-model.In this algorithm,particles are extracted from the observation likelihood function which is depended on the observation noises,and the observation noises are used to model so can be effectively tracked by fusing the latest observations.The bearings-only tracking of maneuvering target in a glint noise environment using the presented algorithm is realized by computer simulation,and compared with the original algorithm.It turns out that the proposed algorithm is more precise and closer to the real-time tracking requirements of high maneuvering targets.Secondly,aiming at the problem of multi-target tracking in clutters,a novel particle filter data association algorithm based on maximum entropy fuzzy clustering is introduced to achieve data association among multiple targets.In this method,the fuzzy membership matrix is obtained by maximum entropy fuzzy clustering algorithm,and the membership is used to represent the joint association probability between targets and measurements.Then the weight of particles in resampling particle filter is calculated,and the transmission between target states is realized by combining observation data.In simulation experiment,the proposed algorithm is compared with the original joint probability data association algorithm and the nearest neighbor algorithm.The results indicate that the presented algorithm has much better tracking performance than the other two algorithms in the interference environment,and it can track multiple targets more accurately in clutters.