Robot Localization based on Modified Particle Filter with Intelligent Ensemble Kalman Filter and MCMC Step

The localization problem is a fundamental requirement for autonomous robots. Different methods are proposed for localization that particle filter based localization is most effective methods. However, this method has problems of inconsistency, degeneracy and dependence on statistics of noises. To solve these problems, in this paper, the modified particle filter based on localization with intelligent ensemble Kalman filter (IEnKF) and MCMC step is proposed. In this approach, the proposal distribution is created using IEnKF that an adaptive neuro-fuzzy supervised its performance. In addition, to increase diversity of particles after resampling, MCMC step is used.