Energy-Aware Trajectory Planning and Deployment of Heterogeneous Swarm of UAV-BS

Abstract

This paper investigates energy-aware joint trajectory planning and service scheduling in a heterogeneous unmanned aerial vehicle base-station (UAV-BS) network comprising conventional hovering UAV-BSs, referred to as Hovers, and graspingenabled UAV-BSs, referred to as Anchors. The service region is discretized into ground cells. Hover UAVs serve multiple adjacent cells from grid intersections, whereas Anchor UAVs directly serve hotspot cells from cell-center service points. Compared with conventional hovering service, grasping-enabled anchoring avoids sustained hovering and can therefore reduce service-phase energy consumption. Using an air-to-ground rate model and a unified time-energy accounting framework for horizontal flight, vertical maneuvering, grasping, and communication, we formulate a twostage mixed-integer linear program (MILP) that first maximizes collected data and then minimizes total energy. We also develop a gain-based greedy benchmark with mandatory-use initialization. Extensive numerical results show that mixed fleets consistently outperform homogeneous deployments. Averaged over map sizes from 4 × 4 to 7 × 7, the MILP transfers 6.1% more data while consuming 33.9% less energy than a randomized feasible baseline. Compared with the greedy heuristic, it achieves essentially the same transferred data while reducing energy consumption by 17.1%. These results highlight the benefit of jointly optimizing UAV trajectories and service scheduling in heterogeneous UAVBS networks under tight mission budgets.