Abstract:

UAVs are widely used in civilian applications， especially in the era of the low-altitude economy. These applications often involve multi-rounds and long durations. However， small or medium UAVs generally have limited endurance. Compared to recharge after returning to their home base， cooperation with mobile ground-based energy-support UGVs may improve efficiency. This cooperation， however， imposes higher demands on task allocation. To address this issue， this paper considers the endurance constraints of UAVs and proposes a dynamic environmental modeling method to describe the usage of multi-round. Specifically， a grid-based dynamic graph model is designed to update waypoint deployment between each round. Aerial-ground heterogeneous platforms plan routes between waypoints while accounting for no-fly and no-entry zones. Through this， an energy cost matrix is constructed. Additionally， a hierarchical fast-solving method for multi-round task allocation is developed， which can be divided into two different levels. At the first level， a clustering algorithm is used to achieve uniform partitioning of way-points into subregions. At the second level， an improved multi-objective genetic algorithm is developed to allocate tasks in two scenarios： within the subregions for UAVs and between the subregions for UGVs. Finally， the parameters affecting task duration and take-off/landing cycles are analyzed. Simulations and experiments validate the efficiency of this heterogeneous system in long-duration and large-area missions. This paper uses the concepts of endurance constraints and redeployment for heterogeneous aerial-ground task allocation and provides a comprehensive framework with real-time applicability， offering a valuable reference for the deployment of autonomous missions.

Key words: aerial-ground heterogeneous collaboration, task allocation, task optimization, trajectory planning, long-term autonomous mission