Abstract
Providing continuous wireless connectivity for high-speed trains (HSTs) is challenging due to their high speeds, making installing numerous ground base stations (BSs) along the HST route an expensive solution, particularly in rural and wilderness areas. This paper proposes using multiple unmanned aerial vehicles (UAVs) to deliver high data rate wireless connectivity for HSTs, taking advantage of their ability to fly, hover, and maneuver at low altitudes. However, autonomously selecting the optimal UAV by the HST is challenging. The chosen UAV should maximize the HST's achievable data rate and provide an extended HST coverage period to minimize frequent UAV handovers constrained by the UAV's limited battery capacity. The optimization challenge arises from accurately estimating each UAV's expected coverage period for the HST, given both are moving at high speeds and the UAV's flying altitude is unknown to the HST. This paper utilizes the estimated HST-UAV channel parameters in the delay-doppler (DD) domain, employing orthogonal time frequency space (OTFS) modulation, to estimate the relative speeds between the HST and UAVs, as well as the UAVs' flying altitudes. Based on these estimates, HST can predict the maximum coverage period each UAV provides, allowing for selecting the best UAV while considering their remaining battery capacities. Numerical analysis demonstrates the effectiveness of the proposed approach compared to other benchmarks in various scenarios.