ROS based Autonomous Mobile Robot Navigation using 2D LiDAR and RGB-D Camera

Abstract

This paper presents an implementation of autonomous mobile robot with the robot operating system (ROS). The system utilizes 2D LiDAR and RGB-D camera with ROS 2D navigation stack, with low power consumption and inexpensive onboard computer. Safe to property and human is of priority. Regarding software, we use official ROS packages with minimal default parameter changes. For hardware, the limitation of equipment and system setting are among challenges. Our proposed systems can perform navigation with dynamic obstacle avoidance capability. The Contribution of this paper is two system setups of ROS navigation stack are proposed. The first system is implemented on Raspberry Pi 3 using 2D LiDAR only. The second system is implemented on Intel NUC using 2D LiDAR and RGB-D camera. To evaluate the performance, usability testing was performed in multiple experiments. Our experiment results show that the robot can avoid objects in their path, or stop in case of unavoidable. Discussion of problems and solutions are presented after the experiment results.