Reinforcement Learning for Age of Information Aware Transmission Policies in Slotted ALOHA Channels

We focus on remote monitoring applications, in which a large number of devices send time-stamped status updates over a wireless channel to a common receiver. An uncoordinated medium sharing policy based on ALOHA is considered, and the overall goal is to maintain an up-to-date perception at the receiver, captured via the average age of information (AoI) metric. In this setting, we propose and evaluate a simple reinforcement learning algorithm which is run independently at each node in a fully decentralized fashion. Leaning on a binary success/collision feedback distributed by the receiver, the solution adapts the access behavior of transmitters based on the current value of AoI. We compare the performance of the scheme to that of threshold ALOHA [1], a benchmark protocol that resorts to a central optimization of the access parameters. Interesting insights on the potential of reinforcement learning for AoI improvements in random access channels are derived.