Abstract
Machine-type communications are emerging as a new paradigm for enabling a broad range of applications from the massive deployment of sensor devices to mission-critical services. To support massive machine-to-machine (M2M) communications with delay constraints in cellular networks, we design an efficient random access and data transmission system known as distributed queueing random access-multiple input multiple-output (DQRA-MIMO) data transmission system. This system has the advantages of both efficient collision resolution of DQRA protocol and the efficient data transmission of MIMO technology. To obtain higher throughput under delay constraint and limited time-frequency resources, we match the ability of collision resolution with the capability of MIMO transmission by optimally configuring system parameters. The closed-form expression of throughput is derived, which is a function of the total user equipments' traffic arrival rate, average packet number of each arrival, number of base station antennas, and number of access request (AR) slots. An optimization problem is formulated to maximize the throughput to obtain the optimal number of AR slots given a certain delay constraint for M2M traffic. Numerical and simulation results reveal that, for a given requirement of average delay, the proposed optimized DQRA-MIMO system, which dynamically adjusts time-frequency resource division to maximize throughput, can provide a higher throughput than that of a baseline approach.