Design and Implementation of Visual SLAM Delivery Robot

Abstract: This paper presents the design, development and implementation of a new autonomous delivery robot employing Simultaneous Localization and Mapping (SLAM) technology. By integrating depth cameras, 2D LiDAR, and Inertial Measurement Unit (IMU) sensors, the robot constructs a real-time map of its environment and identifies obstacles. Utilizing the ROS2 navigation stack, the robot plans optimal routes and avoids obstacles, thereby facilitating efficient and autonomous delivery. The implemented hardware was able to successfully navigate through a test terrain with instantaneous localization and map generation, when target points for the delivery robot are provided via desktop software.

Keywords: Visual SLAM, robot navigation, guidance, sensor fusion.