Particle Filter Algorithm Research And Application

Nowadays, Kalman filter has been widely used in the field of information fusion.It is suitable for linear, Gauss systems, but it can not be used in the nonlinear, non-Gaussian models .In such circumstances, particle filter is widely used because of its wider application.Particle filter is a method based on recursive Bayesian filter and Monte Carle simulation. The method is suitable for any non-linear, non-Gaussian system that could be represented with state model. It is more practical than conventional Kalman filter and its precision could approach optimal estimation. Particle filter has a parallel structure and it is flexible and easy to be implemented.(1) Research on particle filter to resolve the stochastic systems optimal control problems with noise of the non-Gaussian.(2) Research on particle filters for dynamic state-space models handling unknown static parameters are researched. The filters are tested on several different models, and the results are quite promising.(3) A number of different particle filter methods are researched. In order to improve Tracking results, the particle filter methods of tracking are researched deeply. We prove that the effect of tracking of PF-EKF, PF-EKF-MCMC, PF-UKF, PF-UKF-MCMC, under the nonlinear non-Gaussian environment, is better than that of EKF and UKF. The selection of important function will influence the performance of particle filter directly. The results indicate that PF-UKF, PF-UKF-MCMC algorithms are better than extended kalman filter and PF-EKF, PF-EKF-MCMC filter in performance. Specially, the PF-UKF filter with Markov Chains achieves better estimation than PF-UKF filter through a simulation.(4) Finally, the algorithm of IMM for the maneuvering target is researched, and the algorithms of IMM Particle Filter (IMM-PF) and IMM Kalman Filter (IMM-EKF) are researched. The IMM-EKF and IMM-PF are compared in this paper.It shows that IMM-PF filter achieves better estimation than IMM-EKF filter when dealing with non-gauss and non-linear systems.