From the Technologisches Gewerbemuseum (TGM) Institute of Technology, two departments join forces to compete in the Monaco Energy Boat Challenge: Information Technology and Mechanical Engineering. The IT department contributes three specialised project groups, each developing artificial intelligence models that tackle a distinct aspect of autonomous vessel operation for one of the races. The Mechanical Engineering department is responsible for the design and fabrication of the hull, with particular attention to hydrodynamic optimisation and energy efficiency, as well as the steering mechanism that supports accurate, reliable manoeuvring. Every subproject is carried out by students as part of their diploma theses, making the campaign both a competitive entry and a genuine academic milestone.
To find the right course through the buoys for the different races, the team integrates machine learning and neural networks into Unity via the Unity Inference Engine, initially for object detection from the camera feed. The boat itself is simulated in water using the Dynamic Water Physics package, allowing the team to test and refine behaviour before going on the water.
At this stage, the team knows it will need to train its own object-detection model specifically for buoys. The approach for the control AI, the component that turns detected objects into propulsion and steering commands, is still being decided: the team may train a dedicated neural network, or instead adopt a symbolic AI, an explicitly programmed decision logic rather than a learned model, if that proves better suited to the task.
