Evolving Self-Organized Behavior for Homogeneous and Heterogeneous Uav or Ucav Swarms

This research designs an off-line evolutionary system to create multi-UAV behavior capable of searching for and attacking targets. The design for this behav- ior system assumes the UAVs have no apriori knowledge about undetected targets UAVs or the environment. In addition the system does not rely upon global com- munications. WIth regard to the behavior design and approach self-organization is a potential solution since exemplar systems relying upon it tend to be exceptionally robust scaleable and -exible. The UAV behavior evolved with a genetic algorithm relies upon a behavior archetype architecture. This design allows the system to evolve a small set of behaviors that are selected based upon particular sense inputs to the UAVs. The sense inputs summarize observable characteristics of each UAVs environmental representation such as the density of sensed UAVs and a simple target associated pheromone.