Stable and Robust Improvement of AMP for Supporting Massive Connectivity

Abstract

Compressive sensing techniques are widely leveraged to realize the active user detection of grant-free access in massive machine-type communications (mMTC). As a class of efficient data reconstruction methods, approximate message passing (AMP) algorithm and their varieties have attracted considerable attentions. However, as a multiple measurement vector (MMV) problem, AMP based active user detection in multi-antennas systems is difficult to converge, especially when the antenna number of base station (BS) and signal-to-noise ratio (SNR) grow to large value. In order to overcome this drawback of existed MMV-AMP algorithm, we develop an enhanced MMV-AMP algorithm that employs an adaptive iteration stopping criterion and a damping operation. Furthermore, deterministic sequences with low coherence are proposed to replace ordinary random preamble sequences, which could further improve the performance of enhanced MMV-AMP. Simulation results confirm that the proposed scheme efficiently improves robustness and stability of MMV-AMP method.