Online Control of Preamble Groups with Priority in Cellular IoT Networks

Internet of Things (IoT) is the ongoing revolution that offers a truly connected society by integrating several heterogeneous services and applications. A major transformation is urgently needed when the volume of connected devices far exceeds people oriented connections. Due to diversity in applications and requirements, one important trend is that the connected devices could manifest different priorities. With the variety of requirements accordingly, in terms of latency, payload size, number of connections, update frequency, reliability, the Random access process (RAP) as the first step to establish connection between most devices and the base station would also require modification in cellular IoT. In order to prioritize the IoT devices in RAP, we propose a novel online control algorithm with dynamic preamble distribution over multiple priorities. The proposed algorithm recursively estimates the number of active devices in each priority based on Bayesian rule and controls the number of preambles for each priority accordingly. Subsequently, we extend our proposal to incorporate access class baring (ACB) to optimize the algorithm. Extensive simulations show the effectiveness of proposed algorithm over multiple priorities.