Robust FastSLAM Algorithm for Mobile Robot with an Adaptive Rao-Blackwellized Particle Filter

Aimed at SLAM problem, an adaptive Rao-Blackwellized particle filter (RBPF) algorithm is presented to get the unite estimation of the pose of mobile robot and the position of the environmental landmarks. There are three parts in the RBPF algorithm to be studied. For one thing, the resampling of particle filter is combined with local map matching presented in advance to reduce the uncertainty influence. In addition, the pose estimation of mobile robot is mended by adapting the resampling process grounded on the effective sample size (ESS) and by adopting mixture Gaussian distribution to approximate proposal distribution so as to improve the sample weight computation in obtaining ESS. Moreover, the unscented Kalman filter with the adaptation estimation for the process noise is introduced into the position evaluation of the environmental landmarks. With mobile robot MORCS-1 as experimental platform, the validity of the proposed algorithm in this paper is proved.