Aerial Swarm Search for GNSS-Denied Maritime Surveillance

The purpose of this article is to introduce an unmanned aerial swarm control strategy that has been developed specifically for the MBZIRC 2024 competition inspection task, which requires the UAV swarm to search within a global navigation satellite system-denied ocean space. The proposed approach utilizes control barrier functions (CBF) to encode the constraints and control each unmanned aerial vehicle (UAV) using an optimization-based method. We first establish the UAV model by taking energy consumption and the need for charging into consideration, which are then encoded using CBF. Following this, we propose a CBF that ensures each UAV maintains a certain distance from two selected UAVs or beacons on the coast, enabling the swarm to maintain a certain formation using localization algorithms by measuring relative distances. Finally, we present a CBF that integrates Lloyd's algorithm and variable density to enable the swarm to exhibit searching behavior. In numerical simulation with 14 searching UAVs and 3 beacon UAVs, the proposed algorithm shows great scalability and can successfully search the entire task area within 5 min, capable of handling all the constraints, i.e., localization distance, collision avoidance, energy consumption, and task performance, in the considered problem. In our physical experiment with 3 searching UAVs and 3 beacons, the proposed algorithm can accomplish the searching mission within 30 s.