Collaborative Access With Waiting Window: Enhancing Age-of-Information in CSMA Networks

The next generation of industrial Internet-of-Things systems demands timely delivery of fresh data, quantified by the Age-of-Information (AoI) metric, from a vast number of sensors. In such systems, sensors usually adopt random access where the back-off mechanism is mostly crucial. This paper studies the back-off mechanism for Carrier Sense Multiple Access (CSMA) systems in collision-prone scenarios. Considering the fact that transmitting immediately after a prior successful transmission yields limited AoI reduction, we propose an enhanced waiting CSMA scheme by introducing a waiting window to the existing contention window. In the proposed scheme, each source waits after each successful transmission. It enables a collaborative channel access strategy, such that sources have recently transmitted grant greater opportunities for others to access the channel. To optimize the window sizes, we analyze the Normalized Average Peak AoI (NPAoI) and Normalized Average AoI (NAoI), which are normalized by the packet transmission duration. Our findings indicate that minimizing these metrics is feasible through maintaining a nearly constant collision probability. Specifically, NPAoI reaches its minimum by adjusting either the contention or the waiting window, while NAoI requires a larger waiting window combined with a smaller contention window. Given the network scale, we derive optimal window sizes in closed form. When the network scale is unknown or dynamic, we propose a data-driven approach to adjust window sizes based solely on the real-time measured collision probability. Simulation results confirm that the proposed waiting CSMA scheme significantly outperforms existing methods for both NPAoI and NAoI, particularly in large-scale networks.