Expert Guidance System for Unmanned Aerial Vehicles Based on Artifical Neural Networks

Abstract

This article proposes an expert guidance system for Unmanned Aerial Vehicles (UAVs) for marine rescue missions. The difficulty of the problem, due to the time constraints that the mission has to fulfil are lightened by the use of Artificial Neuronal Networks, taking advantage of their high adaptability, low memory requirements, real time response capability, and extrapolation properties. We use them to implement two different types of behaviours for the two main phases of the task: in prediction mode they are responsible of calculating the displacement that the castaways suffer due to the sea and wind currents and in sensing mode they are in charge of guiding the UAV while it tracks already found shipwrecked and search for new ones. To illustrate the successful behaviour of the expert system embedded in a simulator, some results are shown in the final section.