Co-Existing Preamble and Data Transmissions in Random Access for MTC With Massive MIMO

In this article, a Co-existing Preamble and Data (CoPD) random access approach for machine-type communications (MTC) is proposed and studied in massive multiple input multiple output (MIMO) systems. Unlike conventional grant-free random access in which preambles are time-multiplexed with data, the proposed approach enables CoPD transmissions by different groups of devices, which leverages favourable propagation of massive MIMO and successive interference cancellation (SIC) technique. To fully understand the performance behavior of the proposed approach, the error characteristics of channel estimation (CE) due to SIC error propagation are investigated. To avoid the CE error variance growing arbitrarily with time that leads to data transmission failure, key conditions, with respect to preamble length
$L$
, traffic intensity
$\lambda $
, and average number of data packets per active device
$\bar b$
, are derived. Under the conditions, we demonstrate that the throughput of the proposed CoPD approach cannot be arbitrarily high by increasing
$\bar b$
and
$\lambda $
even though the antenna array size, denoted by
$M$
, grows arbitrarily in massive MIMO. Accordingly, its maximum throughput and success probability are theoretically obtained. Simulation results verify our analysis results and confirm that the maximum throughput of the proposed CoPD approach can be at least 4 times than that of the conventional one. Furthermore, its success probability approaches 1 when
$M \to \infty $
as long as the derived conditions are met.