Design of Scheduling Policies for Information Freshness in IRSA with Minimal Feedback

Massive machine-type communication (mMTC), en-abled by modern communication networks, supports push-based communication, which involves generating and transmitting a large number of short packets to a common destination. In such applications, information freshness, measured using performance metrics such as Age of Information (AoI), is becoming increasingly important. In push-based communication systems, optimal scheduling to minimize AoI requires feedback. However, the cost for feedback can be of the order of a forward transmission in mMTC as the data packets are small in size. Motivated by this problem, we consider the design of transmitter scheduling policies with minimal feedback requirements for information freshness. We first consider a point-to-point communication model with an unreliable link. We propose a judicious policy that uses age-beliefs, which requires minimal feedback, namely the link reliability. The average AoI -cost tradeoff performance of this policy is shown to be better than that achieved by an optimal policy for a system with costly feedback of the AoI at the receiver, when the link reliability is high. We then extend our minimal feedback policy to an mMTC scenario with irregular repetition slotted ALOHA and minimal feedback in terms of the decoding success probability. We compare the AoI performance of our policy for IRS A with that of an IRSA-based policy without feedback and show that the AoI is reduced substantially without a significant increase in feedback cost.