Comparative Analysis of SLAM Algorithms for Voice-Controlled Autonomous Wheelchair

Abstract

This paper compares various SLAM (Simultaneous Localization and Mapping) algorithms to determine the most suitable for autonomous navigation of robotic electric wheelchairs. The autonomous wheelchair can navigate, avoid obstacles, and respond to voice commands from the user, allowing it to reach its destination without needing joystick control. This feature is particularly beneficial for individuals with limited or no upper limb mobility who may struggle with joystick-operated wheelchairs. The system uses LiDAR technology to scan for nearby walls and obstacles and create a map of its surroundings. The study evaluates three leading SLAM algorithms-Hector, Gmapping, and RTAB-Map-in various scenarios with dynamic and static obstacles to identify the best algorithm for the project. The SLAM techniques utilize open-source codes from the ROS (Robot Operating System) to construct LiDAR-based maps and localization efficiently. The system is designed to be easily and safely integrated with existing electric wheelchairs.