Loose Formation Obstacle Avoidance Control for Multi-UAV Inspired by Birds Flocking Based on Power Diagrams

This paper presents a novel obstacle avoidance method for unmanned aerial vehicle (UAV) formation, focusing on six-degree-of-freedom low-speed UAVs and adaptable to scenarios with multiple obstacles. Prior to takeoff, a centralized path planning approach is employed for multiple UAVs, initially utilizing a segmentation method based on power diagrams to obtain a grid-based digital map. Subsequently, an A-star algorithm is applied to generate a sequence of waypoints. During flight, the UAVs employ a loose formation obstacle avoidance control method inspired by bird flocking behaviors. By simulating behaviors such as attraction, repulsion, alignment, regulation, foraging, and evasion observed in small bird species like pigeons and European starlings, a controller is constructed to dynamically adjust flight speed and direction, enabling simultaneous passage through waypoints while avoiding obstacles. This facilitates distributed formation obstacle avoidance control. Mathematical simulations demonstrate the effectiveness of the proposed approach in resolving multi-UAV obstacle avoidance challenges.