Large-Scale Antenna-Assisted Grant-free Non-Orthogonal Multiple Access via Compressed Sensing

TL;DR

This paper proposes a massive MIMO-based grant-free non-orthogonal multiple access scheme using compressed sensing, enabling efficient activity detection and data decoding for massive IoT connectivity in 5G.

Contribution

It introduces a novel MMV model leveraging covariance matrices for activity detection in massive MIMO grant-free access, addressing a gap in existing one-shot communication research.

Findings

High activity detection rate achieved

Effective use of covariance matrix in MMV model

Suitable for massive MTC in 5G systems

Abstract

Support massive connectivity is an important requirement in 5G wireless communication system. For massive Machine Type Communication (MTC) scenario, since the network is expected to accommodate a massive number of MTC devices with sparse short message, the multiple access scheme like current LTE uplink would not be suitable. In order to reduce the signaling overhead, we consider an grant-free multiple access system, which requires the receiver facilitate activity detection, channel estimation, and data decoding in "one shot" and without the knowledge of active user's pilots. However, most of the "one shot" communication research haven't considered the massive MIMO scenario. In this work we propose a Multiple Measurement Model (MMV) model based Massive MIMO and exploit the covariance matrix of the measurements to confirm a high activity detection rate.