Novel Method of Mobile Robot Simultaneous Localization and Mapping

Abstract

This paper provides a novel method to realize the mobile robot indoor simultaneous localization and mapping(SLAM),which is a key prerequisite for a truly autonomous robot.This method applies Sequential Monte Carlo(SMC) method which has made much success in localization recently.The localization and mapping is divided as state estimation and parameter estimation respectively,and multiple particle filters are applied to estimate the robot position and the obstacle position simultaneously.The environment information is observed through sonar sensors mounted on a robot itself.The Hough Transform is used for extracting environment obstacle limit feature from the sonar observation.The feature is represented by line approximately,and the robust algorithm is proposed to implement feature matching.The experiment is performed with a Pioneer 2 mobile robot in a real-world indoor environment,and the feasibility of this method is proved.