Priority-Index Based Policies for Minimizing Average AoI in Multi-Hop Network Convergecast With Complete Interference and Link Losses
We consider convergecast in a multi-hop sensor network, in which sensors sample physical processes of interest and transmit packets containing these samples to a single sink node over a routing tree. The sink node remotely estimates these processes based on received packets. Our objective is to design centralized multi-hop sampling and scheduling policies that minimize the average age of information (AAoI) across sensors for networks operating under a complete interference model as well as random losses in the wireless network. We propose a policy that samples sources and schedules links according to an analytically characterized priority index. Through simulations, we compare the AAoI of the proposed policy with policies from prior work and show improved performance. The improvement in performance is observed to be larger as the number of sources increases or when the reliability of links decreases demonstrating the utility of the proposed policy. An extension to $K$ -hop interference model is also discussed.