Distributed Dynamic Pursuit of Multiple Evaders via Allocation Strategy in Obstacle Environments

In this paper, we propose a distributed dynamic pursuit framework for multiple evaders with the same mobility as pursuers in an environment with obstacles, which ensures safety and rational allocation of resources for a multi-pursuer system in complex environments. The framework comprises three main components: dynamic allocation of pursuers, computation of the Collision Avoidance Pursuit Voronoi Cell (CAPVC), and pursuit path control. In the dynamic allocation of pursuers, we first group the pursuers using an improved K-means algorithm and then design an auction algorithm within each group to pair pursuers with subtargets based on the grouping results. The CAPVC consists of three parts: reciprocal collision avoidance hyperplanes, obstacle avoidance hyperplanes, and a buffer term, ensuring collision-free navigation for pursuers in the environment. Finally, based on the allocation results, the nearest point on the CAPVC is identified, and the pursuit path is controlled to chase the evaders. Simulations validate the effectiveness of the proposed framework.