Distributed Access by Multiple Sources for Age of Information Minimization Over a Finite Horizon

Abstract

Age of information (AoI) quantifies the freshness of updates in real time applications, such as vehicular networking or road traffic monitoring and control. This study explores the optimization of AoI over a finite horizon for multiple IoT devices independently tracking the same process of interest and reporting status updates. In this setting, the efficiency of distributed policies where sources probabilistically report their measurements is not adequate. Even assuming pre-defined rendez-vous transmission instants, if the choice about which one to utilize is left to the individual source, lack of coordination may arise, causing simultaneous transmissions (redundant and therefore inefficient) at times, and, consequently, other intervals where no node transmits. We investigate practical solutions to this problem inspired by random medium access techniques. Firstly, we introduce a protocol where no transmission instant is deserted thanks to carrier sensing. If no nodes choose to transmit, they all sense the channel as idle, and randomly revise their decision until at least one transmits. Subsequently, we explore an uneven spread of the transmission instants to balance the resulting scheduling. We measure the effectiveness of these improvements compared to full coordination. These techniques are shown to improve distributed policies by more than 20%, and in general offer valuable insights for future research on sensing in multi-source autonomic environments.