Graphical Techniques for Cooperative Control of Unmanned Assets

Toyon Research Corporation proposes to research and analyze innovative technologies for the dynamic management of Unmanned Air and Ground Vehicles (UAVs, UGVs) and Unattended Ground Sensors (UGS) with an emphasis on achieving synergy and collaboration between the assets. Algorithms for the effective use of dynamic Bayesian networks and decision theory for performing data fusion and inference in military scenarios will be researched. These graphical techniques allow for human oversight of the reasoning process. Several methods of constructing candidate routes for UAVs and UGVs to follow will be investigated and analyzed for their suitability in military intelligence, surveillance, and reconnaissance (ISR) missions. Algorithms for allocating unmanned assets to destinations will be judged on their ability to perform effective pairings with an acceptable computational cost. We will implement the most suitable approach in our software testbed SLAMEMT and demonstrate the feasibility of the resulting system for performing ISR on an example military mission. In Phase II, we would continue to develop prototype software featuring these algorithms, address issues specific to decentralized control, and characterize the performance against a variety of scenarios requiring cooperation between the assets.