A Flexible Task Planning Method for Air–Ground Cross-Domain Uncrewed Swarm Inspired by Wolf Pack Hunting Behavior

To address the highly nonlinear and multimodal planning challenges encountered by the air-ground cross-domain unmanned swarm when executing time-sensitive missions in dynamic battlefield environments, this paper proposes a Flexible Task Planning (FTP) method inspired by the self-organized hunting behavior of wolf packs. By constructing models for heterogeneous unmanned platforms, time-constrained multi-task scenarios, and three-dimensional battlefield environments, a multi-objective effectiveness evaluation system, which integrates value gains, path costs, time costs, and damage costs, is established. Leveraging the wolf pack hunting mechanism, a rolling horizon strategy is adopted to achieve online real-time planning in dynamic settings: target priorities are dynamically ranked based on cost-effectiveness assessments, and a transition probability calculation method that integrates relative cost-effectiveness and responsiveness is designed to achieve flexible task allocation. Additionally, swarm collaboration effects are quantified with parallel path planning and a cooperative gain factor. Simulation results indicate that FTP performs exceptionally well in scenarios involving expanded swarm scales and increased task complexity, with a stable comprehensive score above 0.8 and a task completion rate over 94%. Moreover, its computational efficiency is significantly superior to that of traditional methods, which validates its remarkable advantages in terms of dynamic adaptability, collaborative efficiency, and scalability. This proposed approach offers an efficient solution for the cooperative operations of air-ground cross-domain unmanned swarm.