Simultaneous Localization and Mapping for Autonomous Mobile Robots Using Binocular Stereo Vision System

Abstract

This paper proposes a method of simultaneous localization and mapping (SLAM) for an autonomous mobile robot in an indoor environment using binocular stereo vision systems. Due to sensor model for binocular stereo vision sensor, parallax adjustment and non-local maximum suppression algorithm are used to extract certain two-dimensional horizontal environmental features and vertical edges respectively. This paper also presented an approach to complete Kalman filter (KF) localization and metric map building simultaneously based on the result of lines merging and feature fusion. The results of experiments with a Pioneer robot and a Videre Design stereo vision system demonstrated that robot accurate robot locations can be obtained using the proposed method.