Multi-UAV Collaborative Task Allocation Under Uncertain Demand in Complex Scenarios

Abstract

In complex scenarios such as emergency rescue, power inspection, and battlefield reconnaissance, multi-UAV collaborative task allocation is the core technology for improving task execution efficiency and guaranteeing task reliability. However, the multi-source uncertainty in real scenes severely limits the adaptability of traditional allocation schemes. To solve these issues, we construct a three-layer coupling relationship of "scenario-demand-resource" and establish a robust multi-UAV task allocation model. Furthermore, we propose the Multi-UAV Situation-Aware Collaborative Routing-Task Allocation Model (MSCAR-TAM) and design a grid-based simulation environment that integrates wind field and terrain perception. We categorize task environments into basic, fluctuating, and extreme scenarios and design multi-level fluctuating resource demand patterns to quantitatively characterize the coupled uncertain disturbances. Moreover, we introduce the "scenario weight factor" into the Improved Chaotic Particle Swarm-Grey Wolf Optimizer (ICPS-GWO) to prioritize adaptive optimization for high-uncertainty scenarios, thereby improving the overall environmental adaptability of the scheduling strategy.