A safety-evaluation-driven UAV swarm collision avoidance decision-making based on pigeon-inspired optimization

This research proposes an adaptive avoidance decision-making mechanism that combines safety assessment with pigeon-inspired optimization (PIO) to enhance the collision avoidance capabilities and coordination efficiency of unmanned aerial vehicle (UAV) swarms operating in complicated environments. A multidimensional safety evaluation index system is constructed by analyzing both local conflict risks and the overall coordination level of the swarm. Principal component analysis is employed to derive a comprehensive safety score, based on which a three-stage avoidance control strategy is formulated. An artificial potential field function with adaptively tuned parameters is designed according to the safety score, enabling responsive and autonomous individual-level control. Furthermore, to improve global cooperative avoidance under high-risk scenarios, the improved PIO algorithm is embedded to jointly optimize the velocity direction and avoidance parameters of swarm members. The effectiveness and resilience of the suggested strategy in dynamic contexts are validated by simulation findings, which show that it dramatically lowers the likelihood of local conflicts while preserving swarm cohesion.