2A2-M06 Vision based Localization and Mapping for Indoor Robots using RGBD Sensor

Abstract

This paper presents a solution to the 3D-SLAM problem using low cost RGB-D sensors in indoor environments. For cluttered indoor environment it becomes important to safely navigate the robot from one position to another and 2D data such as scanning lasers do not provide enough environment information. We propose using a combined scanning laser and RGB-D sensor system using which rich 3D maps are built for robot to safely maneuver in a cluttered environment. A graph based SLAM approach is utilized to correct the robot poses and point cloud data are accumulated to generate the 3D map.