Research On Particle Filtering Based On GA-PSO

The Particle Filter is a filtering method and it effectively combined Monte Carlofiltering method with the recursive Bayesian theory. It is simpler than the traditionalmethod of filtering operation, and better filtering, so it is widely used in the treatmentof non-linear and non-Gauss problems.Through in-depth analysis and research of the traditional theory of particle filter,we know that some of its problems. The particle degeneracy phenomenon is thetypical problems to be solved and scholars generally use re-sampling methods toimprove it. However, the re-sampling process will bring new problems of samplediversity reduction. Increasing the number of sampling particles will also improvethe calculation and affect accuracy of algorithm. So, to solve the sample degradationof particle filter has become a research hotspot in recent years.In view of the above problems and combined with the characteristics ofintelligent optimization, Particle Swarm Optimization (PSO) is used to improve thesampling process of the particle filter. Individuals of PSO have strong global searchand optimization capability; the particles can update the optimal solution by itselfand companions. So in the sampling process, particle filter can reduce the particledegeneration problem. But PSO is easy to fall into local optimum and the searchefficiency in later period is low and optimization model of genetic diversity,crossover and mutation in genetic algorithm can increase the understanding betweenthe particles. So this paper presents a method of improved particle filter based onGA-PSO.Through the simulation experiment proves the feasibility of the improvementidea and compared to the improved particle filter algorithm. After the introduction ofthe PSO in traditional particle filter, the tracking improving can be seen from thesimulation results. GA-PSO improved particle filter and the standard PSO improvedparticle filtering simulation proved the superiority of GA-PSO particle filter.