Room Categorization utilizing Convolutional Neural Network on 2D map obtained by LiDAR

Abstract

The robotics sector is experiencing unprecedented growth, driven by the increasing demand for household and assistive robots. These robots need to navigate autonomously between various rooms in a home. To achieve this, they must construct a map of their surroundings and accurately locate themselves within it. Identifying different rooms can enhance the robot's performance. In this study, Gmapping, a Simultaneous Localization and Mapping (SLAM) technique employing a LiDAR sensor, is utilized to generate an environmental map. This map serves as the training data for a Convolutional Neural Network (CNN) designed for room classification. Both simulation and real-world testing demonstrate the effectiveness of CNN in room classification tasks.