Path planning for large-area search by a collaborative Truck-UAVs system

Unmanned aerial vehicles (UAVs) are regarded as valuable tools for field search and rescue because of their wide field of view. In response to the frequent disappearance of tourists in the wild, UAVs can conduct comprehensive searches over vast wild areas to help locate missing tourists and improve rescue efficiency. In this paper, we investigate a collaborative truck-UAVs path planning problem in large-area search and develop a mixed-integer linear programming model with complex constraints such as area decomposition, optional parking-point selection, task allocation, and scan-path connectivity. To solve the problem, a multi-stage adaptive annealing framework with ϵ-greedy large neighborhood search (MAAGLNS) is proposed. MAAGLNS decomposes the problem into multiple stages, including area decomposition, task allocation, truck path planning and UAV path planning, thereby reducing the solution difficulty of the integrated problem. Gurobi is used to validate the proposed model and to provide exact or near-optimal reference solutions on small-scale instances, and extensive experimental results on instances of different scales demonstrate that the proposed method can obtain near-optimal solutions for small-scale instances and significantly outperforms the other five competitors for larger instances. In addition, ablation and sensitivity analyses are conducted to further examine the roles of different operators and the impacts of key parameters on the results. © 2026 Elsevier Ltd. All rights are reserved, including those for text and data mining, AI training, and similar technologies.