A SLAM algorithm for indoor mobile robot localization using an Extended Kalman filter and a segment based environment mapping

Abstract

The present paper faces the Simultaneous Localization And Mapping (SLAM) problem for a mobile robot in an unknown indoor environment. A set of segments is used to model the robot surrounding environment and segments' starting and ending points are used as SLAM landmarks. A segment based mapping algorithm is proposed and used along with an Extended Kalman filter driven by measurements taken by ultrasonic sensors located on the robot. The proposed SLAM algorithm has been tested in both simulated and real experiments yielding to encouraging estimation and mapping results.