LiDAR-Based Autonomous Wheelchair Using ROS2 Navigation Stack

Abstract

Mobility impairments affect millions of individuals worldwide, significantly limiting their independence and access to essential services. This work presents an intelligent autonomous wheelchair system designed to enhance mobility through the integration of advanced motor control, sensor technology, and intelligent navigation algorithms. The system is powered by high-torque Rhino DC planetary encoder high-geared motors that provide precise speed control and smooth maneuverability. For environmental perception, a SLAM-based LiDAR sensor is utilized to generate real-time occupancy grid maps and detect obstacles. An MPU6050 inertial measurement unit is integrated to enhance motion stability by tracking orientation and tilt on uneven terrain. The wheelchair supports manual, semi-autonomous, and fully autonomous operation modes. The software architecture leverages the ROS2 Navigation Stack and SLAM Toolbox with A* and Dijkstra global planning algorithms to achieve optimal route selection and dynamic obstacle avoidance. Experimental evaluation demonstrates a localization update rate of 50 Hz and path-following error below 10 cm, validating the effectiveness of the proposed system for assistive indoor mobility.