Dynamic Random Access Without Observation Under Deadline-Constrained Periodic Traffic

This article focuses on random access in uplink systems under deadline-constrained periodic traffic, which is typical for many real-time Internet of Things scenarios. To achieve very low overhead in random access, we consider dynamic slotted ALOHA without observation where each active node adopts time-dependent but observation-independent transmission probabilities. Built on the theory of blind Markov decision processes, we develop an analytical framework of such dynamic access with a simplified version of information states, which leads to optimal time-dependent transmission probabilities. Further, based on this framework, we derive simple closed-form expressions for optimal time-dependent transmission probabilities and maximum long-run system throughput, which makes our scheme also enjoy very low complexity. Numerical results show that the proposed scheme outperforms other random access schemes without observation over a wide range of network configurations.

I. Introduction

To enable timely monitoring and control, real-time Internet of Things (IoT) systems that impose a strict deadline on packet delivery have become prevalent in industry, environment, and transportation domains [1], [2]. Out of several such systems, this article focuses on uplink IoT systems under deadline-constrained periodic traffic, which are typical for automation control loops and process control use cases [3]. For example, with a predefined delivery deadline, pressure sensors periodically send values to a machine PLC, so that the PLC can continuously monitor the status and take necessary actions.