Internet of things (IoT) devices in the long term evolution-advanced (LTE-A) network require the random access (RA) for the data transmission. The base station in LTE-A requires a decision algorithm for the number of preambles and for the probability of devices to enter contention. This paper proposes an adaptive resource allocation and congestion control algorithm referring to the most recently observed contention results. Furthermore, this paper also proposes an estimation method which estimates the number of contending devices and the number of activated devices in the network based on the unused number of preambles. The performance evaluation using the RA simulator shows that the proposed algorithm can achieve the throughput which is close to the optimal throughput. We also address the limit of current LTE-A system to support massive number of devices based on the evaluation results.
