Deployment issues for massive MIMO systems

Neil, Callum T.; Shafi, Mansoor; Smith, Peter J.; Dmochowski, Pawel A.

doi:10.1109/iccw.2015.7247357

Cited by 3 publications

(1 citation statement)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Massive MIMO, which scales up the number of TX and RX antennas by at least an order of magnitude relative to conventional MU-MIMO systems, leverages the spatial dimension providing significant increases in data rate, link reliability, energy efficiency, and multiplexing gains while reducing inter-user interference. Consequently, the investigation of massive MIMO technology as an emerging technology is well underway [6][7][8][43][44][45][46][47][48][49][50][51][52][53][54][55][56][57][58][59][60][61][62].…”

Section: Massive Mimomentioning

confidence: 99%

On the Performance and Analysis of Massive MIMO for 5G Wireless Systems

Neil¹

View full text Add to dashboard Cite

<p>A novel technical solution, and paradigm shift, envisioned to achieve the significant spectral efficiency enhancements required for Fifth Generation (5G) wireless systems is massive multiple-input-multiple-output (MIMO). Massive MIMO systems scale up the number of transmit (TX) and receive (RX) antennas by at least an order of magnitude relative to conventional multi-user MIMO systems, which have been a key feature in current wireless standards, such as Long Term Evolution. Thus, massive MIMO leverages the spatial dimension by providing significant increases in all the virtues of conventional MIMO systems but on a much larger scale. Namely, data rate, link reliability, energy efficiency, and multiplexing gains can all be increased with massive MIMO systems, while simultaneously reducing inter-user interference through digital processing techniques. Further motivating the surge in research of massive MIMO systems are the additional channel properties which occur when operating with large dimensions. These properties arise as a result of random matrix theory asymptotics and under these conditions random variables become deterministic, simplifying analysis and allowing simple processing techniques to become (near) optimal. These idealistic properties, however, are based on the assumptions of an independent and identically distributed channel matrix with an infinite number of TX antennas. Physical contraints typically prohibit the deployment of large numbers of TX antennas. It therefore seems natural to determine the number of TX antennas required for large MIMO systems to begin to exhibit these favourable asymptotic properties. Analytically deriving the first and second moments of the composite Wishart channel matrix and numerically defining three convergence metrics, the rate of channel convergence is examined. Limiting matched-filter (MF) and zero-forcing precoding signal-to-interference-plus-noise-ratio (SINR) performances are then analytically derived and rate of convergence shown. Coordinated distributed MIMO systems can mitigate the detrimental effects of spatial correlation relative to a colocated MIMO system. The instantaneous and limiting MF SINR performance of a distributed massive MIMO system is derived, allowing clear insights into the effects of imperfect channel state information, spatial correlation, link gains and number of antenna clusters. The wide bandwidths vacant at millimeter-wave (mmWave) frequency bands are suitable for 5G wireless systems since they occupy regions of uncongested spectrum which enable large contiguous bandwidth carriers. Spatial correlation of an arbitrary antenna array topology is analytically derived for a mmWave channel model. Numerically, the effects of mutual coupling amongst antenna elements is then shown on the effective spatial correlation, eigenvalue structure and user rate of different antenna topologies. Channel models and measurements across a wide range of candidate bands for 5G wireless systems are then considered, motivated by the different propagation and spatial characteristics between different bands and different channel models within the same band. Key channel modelling and spatial parameter differences are identified and, in turn, their impact on various antenna topologies investigated, in terms of system sum rate, channel eigenvalue structure, effective degrees of freedom and massive MIMO convergence properties.</p>

show abstract