Result representation of Rao-Blackwellized particle filtering for SLAM

Recently, particle filters have been applying to many robotic problems including the simultaneous localization and mapping (SLAM). Specifically, SLAM approaches employing Rao-Blackwellized particle filter (RBPF) have shown good results. However, no research is conducted to analyze representation of the results of particle filtering. After finishing the particle filtering, the results such as a map and a path are stored in the separate particles. In most cases, the result of the particle that has the highest importance weight is represented as the result. However, this approach does not give the best result all the time. Thus, We provide the analysis of final representation of particle filtering. In this paper, we compares several methods to derive the final representation of the result after finishing RBPF-SLAM. According to the result, combining data of each particle provides the better result with high probability than using just data of a particle such as the highest weighted particle representation.