Capture effect-based optimal access class barring for IoT communications

For 5G and beyond-5G systems, a key requirement is the support for massive Internet-of-Things (IoTs) and Machine Type Communication (MTC) devices that are integrated in the network for the purpose of data collection and reporting in various applications. During the random-access protocol of LTE-A/5G, large number of IoT devices access the network through preambles (PA) resulting in congestion. This is due to large likelihood of selection of same PA by more than one IoT device, which is called collision. Access Class Barring (ACB) scheme is used to control the congestion by limiting access of the devices to the network thereby reducing chances of collision. However, success is still possible if the two or more devices select the same PA, and are decoded correctly by the gNodeB (gNB), which is called capture effect. In this paper, we incorporate the effect of capture probability in the system model of IoT devices, and derive the optimal transmission probability, called as optimal ACB factor. Using this optimal ACB factor, a modified success probability is computed, and its performance is compared with the success probability based on conventional ACB factor. A Bayesian strategy is also proposed to estimate the number of backlogged IoT devices based on the information of idle, success, and collided PAs. Moreover, we also derive important performance parameters like total service time (TST) and access delay, and conduct extensive simulations to verify the analysis. Simulation results show significant improvement in TST and access delay if capture effect probability is used, and this scheme can be used to achieve the latency requirements of 5G.