Distributed feature based RBPF multi robot SLAM

Abstract

This paper presents a new approach to solve multi robot simultaneous localization and mapping (SLAM) problem by allowing fast and efficient features exchange among teammates. Each mate is able to build its own SLAM solution by using feature based Rao-Blackwellised Particle Filter algorithm. This scheme proposes that every feature extracted by individual mate should be shared with all team mates no matter that some features are viewed by more than one member. As the information exchange comprises only on small features set so the time taken to accommodate the information is small which makes it a good solution for distributed multi-robot system working with limited communication range and bandwidth. This paper presents experimental results for two different environments by using different platforms.