Development of a Concept for Swarmed mini-UAVs for Automatic Target Recognition

Unmanned Aerial Vehicles (UAVs) have served to enhance and augment the surveillance and reconnaissance abilities of the military of the United States of America. In developing the next generation of UAVs, the application of novel methodologies to UAV control will become more important as the number and mission criticality of deployed UAVs continues to increase. One of these novel approaches, long considered to possess some ideal characteristics, is the idea of reactive agents and the associated emergent behavior. This technique possesses the potential to provide minimal user intervention, a high level of robustness, and largely autonomous operation provided that suitable algorithms and techniques can be developed. The proposed work should provide the basic mathematical framework for reactive agent based Automatic Target Recognition in which these reactive agents could be simulated and, based on this simulation, optimization strategies for simple emergent behaviors developed along with mission based parameters estimated. The proposed approach models sensing behavior from an information theory perspective and then utilizes this in generating mission based system parameters followed by optimization of those parameters. If successful, this should allow for the implementation of reactive agent based techniques for UAV control embodying the benefits as appropriate.