Optimal selection of A2G and A2S links for UAM communications

Abstract

Urban Air Mobility (UAM) is a next-generation airborne transportation that has emerged as an effective alternative to the existing ground transportation system to better support today’s high transportation demand in large cities. In operating UAMs, two types of communication traffic must be provided seamlessly, which are UAM control function traffic and passenger data traffic. The former is used to ensure that UAMs in the airspace are under the control of the Air Traffic Center (ATC), whereas the latter is for passengers’ Internet connection. Since UAMs are flying in the three-dimensional space, there exist three kinds of wireless links that UAMs can utilize to establish the network connection: air-to-ground (A2G) / ground-to-air (G2A), air-to-air (A2A), and air-to-satellite (A2S) / satellite-to-air (S2A) links. While A2G/G2A communications relies on the conventional terrestrial network (TN) via ground base stations, A2A and A2S/S2A communications requires non-terrestrial networks (NTN) via nearby UAMs as a relay and via satellites, respectively. Therefore, UAMs should select an optimal link (among the three kinds) at each location on its trajectory considering different characteristics among them and their varying channel capacity and end-to-end latency. Moreover, such optimal link selection should also consider the overhead incurred when switching links between two consecutive selection epochs, which is caused by reconfiguration of the end-to-end wireless session due to the update in multi-hop routing path. To address the aforementioned issue, this thesis first sets up a complete system model for a UAM periodically choosing between A2G/G2A and A2S/S2A links from its takeoff to landing, and then formulates the scenario as a latency-constrained optimal link selection problem considering per-link channel capacity, end-to-end latency, and link switching overhead. Then, the thesis conducts a feasibility study in UAM uplink/downlink communication scenarios and discusses the results.