Persistent tracking of ocean sharks withcooperative air vehicles

Abstract

This thesis presents an architectural design for an acoustic tracking system to allow continuous fish tracking in the open ocean. The tracking system consists of a team of unmanned aerial vehicles (UAV). The application is about sharks. Each shark is tagged with an acoustic signal transmitter, placed on their dorsal fin, allowing receivers mounted on the UAS to measure the signal intensity and pin point the location. Therefore, each UAV should have a unique and predefined search area to cover a larger surface. Since they are aerial vehicles they have to land on the water, allowing their acoustic receivers to listen for the acoustic signal, transmitted by the tag placed on the targeted shark. Until they find the shark, they should constantly land in different positions of their search area. Once one of the UAV's receives the signal, it should inform the other vehicles about the shark's estimate location, allowing them to reduce their search area and successfully track it easier. After finding the shark, regular future position estimation should be used for two reasons: (1) the signal has dead zones so even if the vehicle is in range, there will be no signal to listen to and, (2) the shark can dive deeper than the maximum range of the receiver, losing the signal for good. With future estimation, this problems can be minimized since it will be possible to estimate where the shark would be going in the following minutes, making it possible to define new search areas for the tracking team. Real data comprised by recorded shark routes and depth variation will be used to validate the final design.