Development of a Mobile Robot Based on ROS2 Architecture

Abstract

The primary aims of this work were to create an autonomous path planner for an unknown environment and to implement Simultaneous Localisation and Mapping (SLAM), both achieved through the development and construction of a differential-drive mobile robot. Its approach is step-by-step, starting with kinematic and sensor validation of a Unified Robot Description Format (URDF) model in a Gazebo simulation environment, followed by physical hardware integration. The hardware is designed around the Raspberry Pi 4 Model B as the central processing unit, with the Arduino Nano connected via a serial bridge for low-level motor control. The sensor is the RPLIDAR A1, which perceives the environment. It is built on the ROS 2 Humble Hawksbill release, the SLAM Toolbox to create a 2D occupancy grid, and Navigation 2 (Nav2) to plan both global and local paths. The experimental results demonstrate that the system generates high-quality static maps and autonomously navigates while avoiding dynamic obstacles. Lastly, this low-cost, easily copiable platform offers a handy testbed and has a high likelihood of producing educational robotics and applied ROS 2 research.