A framework and analysis for cooperative search using UAV swarms

We design and analyze the performance of cooperative search strategies for unmanned aerial vehicles (UAVs) searching for moving, possibly evading, targets in a hazardous environment. Rather than engaging in independent sensing missions, the sensing agents (UAVs with sensors) "work together" by arranging themselves into a flight configuration that optimizes their integrated sensing capability. If a UAV is shot down by enemy fire, the team adapts by reconfiguring its topology to optimally continue the mission with the surviving assets. We presetn a cooperative search methodology that integrates the multiple agents into an advantageous formation that distinctively enhances the sensing and detection operations of the system while minimizing the transmission of excessive control information for adaptation of the team's topology. After analyzing our strategy to determine the performance tradeoff between search time and number of UAVs employed, we present an algorithm that selects the minimum number of UAVs to deploy in order to meet a targeted search time within probabilistic guarantees.