Visual Docking Guidance for Autonomous Boats using Distributed ESP32-CAM and AprilTag in ROS

Autonomous docking remains a critical challenge for unmanned surface vehicles due to GPS inaccuracies in close-quarters maneuvering. This thesis presents a scalable, low-cost visual docking guidance system utilizing distributed ESP32-CAM nodes and AprilTag markers within the ROS2 ecosystem. To prevent wireless bandwidth bottlenecks, image processing is distributed entirely to edge devices; highly optimized firmware enables €10 microcontrollers to track tags locally and transmit only lightweight localization data. A dedicated ROS2 bridge translates these packets into standardized messages, creating a plug-and-play, boat-agnostic perception sensor. Experimental results demonstrate robust relative localization, offering an efficient alternative for autonomous vessel navigation.

Visual Docking Guidance for Autonomous Boats using Distributed ESP32-CAM and AprilTag in ROS