An Iterative Stochastic Approach to Constrained Drones' Communications

Abstract

The Internet of Drones paradigm is considered as a key enabler for several cutting edge verticals, including surveillance, planetary exploration, protection, loads transportation, and aerology. The main limitations to its wide-scale adoption arise from the constraints on the resources available onboard of drones: this concerns energy, computational and storage capabilities. Unfortunately, current literature mainly focuses on energy limitations, leaving unexplored the interplay with other constraints. To bridge this gap, the present contribution also encompasses the limitations on the memory onboard, which can be critical when drones have to acquire high resolution multimedia signals for ambient awareness services. In particular, an iterative stochastic approach is conceived hereby to tune data flows from/to drones subject to energy and memory constraints in order to fulfill an Out-of-Service probability below a given threshold. Stemming from the proposed approach, two algorithms have been also designed that seek a different complexity-performance tradeoff. The first one is less complex and more conservative, since it plans the mission once at the beginning. The second, instead, is slightly more complex and aggressive but it allows the drone to gather and upload a higher volume of data and shorten the gap with respect to the ideal case.