Multiuser Media-based Modulation for Massive MIMO Systems

TL;DR

This paper explores media-based modulation (MBM) in massive MIMO systems, demonstrating that MU-MBM can achieve comparable or better performance with fewer antennas than traditional schemes, and proposes a compressive sensing-based detection method.

Contribution

The paper introduces multiuser media-based modulation (MU-MBM) for massive MIMO, showing its superior performance and proposing an efficient sparse signal detection scheme.

Findings

MU-MBM outperforms conventional modulation with fewer antennas.

Using 128 antennas with MU-MBM matches the performance of 500 antennas with traditional schemes.

Sparse signal detection algorithms like OMP, CoSaMP, and SP are effective for MU-MBM detection.

Abstract

In this paper, we consider {\em media-based modulation (MBM)}, an attractive modulation scheme which is getting increased research attention recently, for the uplink of a massive MIMO system. Each user is equipped with one transmit antenna with multiple radio frequency (RF) mirrors (parasitic elements) placed near it. The base station (BS) is equipped with tens to hundreds of receive antennas. MBM with $m_{rf}$ RF mirrors and $n_r$ receive antennas over a multipath channel has been shown to asymptotically (as $m_{rf}\rightarrow \infty$) achieve the capacity of $n_r$ parallel AWGN channels. This suggests that MBM can be attractive for use in massive MIMO systems which typically employ a large number of receive antennas at the BS. In this paper, we investigate the potential performance advantage of multiuser MBM (MU-MBM) in a massive MIMO setting. Our results show that multiuser MBM (MU-MBM) can significantly outperform other modulation schemes. For example, a bit error performance achieved using 500 receive antennas at the BS in a massive MIMO system using conventional modulation can be achieved using just 128 antennas using MU-MBM. Even multiuser spatial modulation, and generalized spatial modulation in the same massive MIMO settings require more than 200 antennas to achieve the same bit error performance. Also, recognizing that the MU-MBM signal vectors are inherently sparse, we propose an efficient MU-MBM signal detection scheme that uses compressive sensing based reconstruction algorithms like orthogonal matching pursuit (OMP), compressive sampling matching pursuit (CoSaMP), and subspace pursuit (SP).