Two-Level Estimation Enabled Online Congestion Control for Massive IoT Networks

In the massive Internet of Things (mIoT) scenario, characterized by a burst of access requests, the random access (RA) mechanism faces significant challenges in establishing radio resource control (RRC) connections. Access class barring (ACB) and Backoff are two typical control schemes. Devices first undergo the ACB check, and upon passing, transmit preambles and payloads in a contention-based manner. Failed attempts then enter the Backoff process for retransmission. Maximizing RA efficiency by collaborating these two control schemes is a critical challenge. This paper presents a performance analysis of the coexistence of ACB and Backoff and proposes an optimal control scheme. To enhance practical applicability, a Bayesian estimation-based approach is introduced. Simulation results validate the proposed algorithm’s substantial improvement in RA efficiency.