Adaptive MCMC Particle Filter for Nonlinear and Non-Gaussian State Estimation

Abstract

The particle filter is well known as a state estimation method for nonlinear and non-Gaussian system. However, particle filter has the inherent drawbacks such as samples less of diversity and the computational complexity depends on the number of samples used for state estimation process. In this paper, the adaptive Markov chain Monte Carlo (MCMC) particle filter is proposed in order to overcome these drawbacks. In the new algorithm, the KLD-sampling and MCMC sampling are simultaneously used to improve the performance of particle filter. The computer simulations are performed to compare the adaptive MCMC particle filter algorithm, the MCMC particle filter and particle filter in performance. The simulation results demonstrated that the adaptive MCMC particle filter is very efficient and smaller time consumption compared to MCMC particle filter and particle filter. Therefore, the MCMC adaptive particle is more suitable to the nonlinear and nonGaussian state estimation.