Implementation of Visual Odometry on Jetson Nano.

This paper presents the implementation of ORB-SLAM3 for visual odometry on a low-power ARM-based system, specifically the Jetson Nano, to track a robot's movement using RGB-D cameras. Key challenges addressed include the selection of compatible software libraries, camera calibration, and system optimization. The ORB-SLAM3 algorithm was adapted for the ARM architecture and tested using both the EuRoC dataset and real-world scenarios involving a mobile robot.