The OpenSwarm project develops an automated system that monitors and counts boats in a protected marine area (PMA), and tracks their speed to help manage underwater noise pollution. The proposed system uses smart buoys (the devices) equipped with hydrophones. The buoys can communication with one another using OpenSwarm’s mesh networking technology. One of the buoys is the gateway (the edge), and is connected to the OpenSwarm cloud. The buoys listen and detect boats from the sounds recorded by their hydrophones. They send the timestamped sound signature to the edge, which compares them, computes the location and speed of the boats, before forwarding that information to the OpenSwarm cloud running on a server on the Internet. The project demonstrates swarm communication, constrained AI framework and energy-aware swarm operation. It takes advantage of the OpenSwarm swarm compiler to allow a user to efficiently program and control the behavior of the devices. The AI model can assess the presence and well-being of wildlife and thus contribute to managing the traffic so that damage to nature is minimized. This system can be applied to all 3,150 European PMAs and to approximately 5,000 marinas in Europe.

We are conducting a comprehensive series of experiments in the Etang de Thau in the South of France. We use buoys equipped with hydrophones. Once installed, the buoys record the underwater noise as boats are passing by. The data we collect is used in a simulator where underwater sound propagation is recreated, allowing us to generate synthetic sound samples as training data for an embedded Neural Network solution.