Development of an indoor localization and navigation system based on monocular SLAM for mobile robots

TL;DR

This paper presents an indoor localization and navigation system for mobile robots using monocular SLAM, integrating ROS, ORB_SLAM3, and path planning algorithms, tested successfully in simulation.

Contribution

It introduces a monocular SLAM-based localization and navigation system for differential-drive robots with integrated hardware and algorithms, tested in Gazebo.

Findings

System achieves accurate indoor localization.

Navigation system demonstrates efficiency in simulation.

Integration of SLAM with path planning enhances robot autonomy.

Abstract

Localization and navigation are two crucial issues for mobile robots. In this paper, we propose an approach for localization and navigation systems for a differential-drive robot based on monocular SLAM. The system is implemented on the Robot Operating System (ROS). The hardware includes a differential-drive robot with an embedded computing platform (Jetson Xavier AGX), a 2D camera, and a LiDAR sensor for collecting external environmental information. The A* algorithm and Dynamic Window Approach (DWA) are used for path planning based on a 2D grid map. The ORB_SLAM3 algorithm is utilized to extract environmental features, providing the robot's pose for the localization and navigation processes. Finally, the system is tested in the Gazebo simulation environment and visualized through Rviz, demonstrating the efficiency and potential of the system for indoor localization and navigation of mobile robots.