Coordinated sea rescue system based on unmanned air vehicles and surface vessels

This paper presents a sea rescue system based on a coordinated team of a sensing/monitoring Unmanned Aerial Vehicle (UAV) and a rescuing Unmanned Surface Vessel (USV) that exploits the measurements provided by the UAV to estimate the castaways position. The system models the castaway location evolution using an Artificial Neural Network (ANN) that is trained before the rescue starts using a map of the sea wind and currents. The UAV predicts the position of the castaways with the prediction ANN and searches the castaways using a controller implemented with another ANN trained with searching behaviors. The USV, with a bigger computational power, incorporates a Particle Filter (PF) to estimate the castaway location. This PF uses the prediction ANN to predict the position of the particles and the measurements of the UAV and USV to update their weights. Finally, the paper presents some simulated experiments that show the whole system performance under different situations.