The Role of Feedback in AoI Optimization Under Limited Transmission Opportunities

Abstract

Scheduling updates from remote sensors is a key task for the internet of things (IoT). In this context, the mathematical concept of age of information is often used to capture the freshness of received data. This is, in turn, relevant to optimize the frequency of the exchanges, especially for resource constrained (e.g., energy-limited) sensors. Most investigations on the subject assume that the transmitter can leverage knowledge of the age of information at the receiver side to decide when to send data, even when the communication channel is unreliable. In reality, tracking the outcome of the updates would require additional consumption of resources to acquire a feedback. We investigate the optimal schedule of updates over a finite time horizon for a resource-constrained sensor that is allowed to perform a limited number of updates, as typical of IoT devices. We discuss the role of the feedback from the receiver, and whether it is convenient to ask for it whenever this causes additional energy consumption and consequently allows the transmission of a lower number of updates. We analytically identify regions for the feedback cost and the reliability of the channel where making use of feedback may or may not be beneficial.