Evolutionary Design of the Controller for the Search of Area with Obstacles using Multiple UAVs

Simultaneous operation of multiple UAVs enables to enhance the mission accomplishment efficiency. In order to achieve this, easily scalable control algorithms are required, and swarm intelligence having such characteristics as flexibility, robustness, decentralized control, and self-organization based on behavioral model comes into the spotlight as a practical alternative. Recently, evolutionary robotics is applied to the control of UAVs to overcome the weakness of difficulties in the logical design of behavioral rules. In this paper, the neural net controllers evolved by evolutionary robotics are applied to the control of multiple UAVs which have the mission of searching area with obstacles. Several numerical demonstrations show the proposed algorithm has superior results to those of behavior based neural net controllers designed by intuition.