A decentralized decision-making algorithm of UAV swarm with information fusion strategy

Centralized and decentralized unmanned aerial vehicle (UAV) swarms rely on networked communication to exchange information among individuals and generate cooperative behaviors. This study proposes an Information-Fusion based Decentralized Swarm Decision Algorithm (IFDSDA) to coordinate UAV swarms when communication is interfered with or fails. Each UAV has the ability to perceive the area ahead using a monocular camera. An information fusion strategy is presented to unify the two types of information obtained from communication and visual perception in a standardized manner. This mechanism enables UAVs to fully utilize different information in cases of communication failure. Each UAV is controlled by a decentralized swarm decision module that takes the fused information as inputs and generates the heading orientation by combining basic action rules of UAVs. The weight parameters of the combination are optimized using the heuristic genetic algorithm in an offline manner. The collision/obstacle avoidance and area search missions are used as test scenarios for the proposed IFDSDA. Simulations with high confidence are conducted, which demonstrate the effectiveness, inherent scalability, and robustness of the proposed method, comparing it with the state-of-the-art ISOA method and its variant. This study reduces the dependence on network communication within the swarm and improves the adaptability of UAV swarms in complex battlefield environments.