Research On Key Technology And Application Of Particle Filter

Particle filter (PF) algorithm is a technique for implementing a recursive Bayesian filter by Monte Carlo simulations which represents the required posterior density function by a set of random samples with associated weights. Therefore, PF can be typically used to describe any complex nonlinearities and non-Gaussian distributions and compute any statistical estimates, such as mean and variance, more accuracy. In recent years, PF has been paid more and more attention and become a new promising topic in applied statistics,signal processing, automatic control, artifical intelligence and machine learning communities. In spite of that, the study of PF is just in the primary stage, and there still exists many critical problems need to be investigated.The sample degeneracy is the critical problem existed in particle filter. In order to solve this problem, a new combined particle filter algorithm, based on the genetic simulated annealing algorithm and unscented Kalman filter algorithm, is presented in this paper. In the proposed algorithm, unscented Kalman filter algorithm is used to generate the importance proposal distribution which can match the true posterior distribution more closely, and genetic simulated annealing algorithm based upon the survival-of-the-fitness principle is applied to enhance the diversity of samples. The results show the effectiveness and feasibility of the proposed algorithm. In the end, the new algorithm is implemented in the target tracking.and image processing simulations, and the experimental results is satisfactory.