Age of Information Aware UAV Network Selection

For the command and control in unmanned aerial vehicles (UAVs), it is important to limit the latency of the real-time status updates. Previously proposed network selection schemes mainly select the closest or the strongest-signal base station (BS) for data rate maximization, thus neglecting the BSs' queueing and handover delays. In this paper, we aim to minimize the age of information (AoI) in both the network access and handover. Specifically, with the BS' load and UAVs' flight plan information, each UAV needs to choose between uncongested BSs for low-latency updates or BSs along its trajectory for less frequent handovers. As the UAVs' decisions are coupled towards the BSs' load, we formulate the UAVs' interactions as a noncooperative game, where each UAV aims to minimize its cost as the summation of the associated BSs' average AoI and the handover penalties. We show that it is a potential game by characterizing its exact potential function. It leads to the design of a distributed BS association (DBA) algorithm, whose output is guaranteed to converge to a Nash equilibrium within a finite number of iterations. Simulation results show that the DBA scheme's load-aware handover leads to a lower AoI cost than two benchmark schemes.