Evaluating UAV flock mission performance using Dudek's taxonomy

We use Dudek's taxonomy in order to investigate the performance of a group of autonomous UAVs cooperating in mission execution against a group of enemy agents acting in an unknown environment. We show that increasing the number of UAVs in the group proves to be beneficial as it allows the group to react to more enemy events. We also show that using communication helps creating better cooperation between the flock members; however, using infinite communication range or infinite communication bandwidth results in considerable computational complexity. We conclude that it may be sufficient to use finite-bandwidth communication, keeping the computational complexity constant with the number of UAVs in the group, thus allowing the group to be scalable to large numbers of UAVs. We use flocking behavior to control the movement of the UAVs when searching, as the flocking helps the UAVs to disperse in the environment and improve the detection probability of the flock. We show that using flocking improves the group performance only if the group is capable of mission task collaboration.