Posterior Elimination Fast Look-Ahead Rao-Blackwellized Particle Filtering for Simultaneous Localization and Mapping

摘要

In this paper, a method for further reducing computation time for Fast Look-Ahead Rao-Blackwellized Particle filtering (Fast la-RBPF) for Simultaneous Localization and Mapping (SLAM) problem is presented. By using the pose posterior probability, states with low posteriors are excluded from the Kalman filtering updates to the computation time can be reduced. In conjunction with Fast la-RBPF, the proposed algorithm is then called the Posterior Elimination Fast la-RBPF SLAM algorithm. Simulation results show that the proposed method is more efficient in terms of computation time than Fast la-RBPF without any sacrifice in both localization and mapping error performances.