Visual Marker Tracking and Following for Autonomous Robots: A ROS2 Implementation

This thesis is part of a project began at FieldRobotics, a company specializing in robotic solutions for agricultural and on-field data collection. The project focused on extending the functionality of an autonomous rover, enabling it to follow a visual marker for improved navigation in agricultural environments. The system integrates several key technologies, including Docker for containerized development, ROS 2 Humble for robot control, the Nav2 stack for autonomous navigation, and OpenCV for image processing and marker detection. The thesis begins with a review of the state of the art, covering autonomous robots and the challenges associated with target-tracking algorithms. It then explores the theoretical foundations of autonomous navigation, path planning, and localization. Next, it introduces essential concepts related to Docker and ROS 2, which played a crucial role in the project's development. Finally, the work delves into the implementation details, discussing the technical challenges encountered and the solutions adopted throughout the project.