Minimizing Wait Time and Age of Information in Mobility-Enabled Communication Systems

Recent advances in robotics present new paradigms in communication systems, particularly the use of autonomous vehicles to enhance communication capabilities in a number of settings. Polling systems, in which a single server services multiple queues and incurs some idle time when switching between these, provide useful models for a number of scenarios encountered in such systems. In this paper, we extend traditional polling systems to model these mobility-enabled communication scenarios by embedding the polling systems in 2D Euclidean space and defining service regions in this space where the queues' requests may be serviced. We show that this model can be used to capture the Age of Information (AoI) in the system as well. We pose the fundamental problem of minimizing the average wait time for a request, a key metric of the system's performance, and we provide lower and upper bounds on the optimal solution. Focusing on a cyclic visit order, we find provably optimal trajectories for the low- and high-traffic cases. Finally, we present numerical results from an IoT data collection scenario which illustrate the derived theory and show significant improvements in average wait time when compared to a move-stop-communicate baseline.