Emergent Mind

LOS-based Conjugate Beamforming and Power-Scaling Law in Massive-MIMO Systems

(1404.1654)
Published Apr 7, 2014 in cs.IT and math.IT

Abstract

This paper is concerned with massive-MIMO systems over Rician flat fading channels. In order to reduce the overhead to obtain full channel state information and to avoid the pilot contamination problem, by treating the scattered component as interference, we investigate a transmit and receive conjugate beamforming (BF) transmission scheme only based on the line-of-sight (LOS) component. Under Rank-1 model, we first consider a single-user system with N transmit and M receive antennas, and focus on the problem of power-scaling law when the transmit power is scaled down proportionally to 1/MN. It can be shown that as MN grows large, the scattered interference vanishes, and the ergodic achievable rate is higher than that of the corresponding BF scheme based fast fading and minimum mean-square error (MMSE) channel estimation. Then we further consider uplink and downlink single-cell scenarios where the base station (BS) has M antennas and each of K users has N antennas. When the transmit power for each user is scaled down proportionally to 1/MN, it can be shown for finite users that as M grows without bound, each user obtains finally the same rate performance as in the single-user case. Even when N grows without bound, however, there still remains inter-user LOS interference that can not be cancelled. Regarding infinite users, there exists such a power scaling law that when K and the b-th power of M go to infinity with a fixed and finite ratio for a given b in (0, 1), not only inter-user LOS interference but also fast fading effect can be cancelled, while fast fading effect can not be cancelled if b=1. Extension to multi-cells and frequency-selective channels are also discussed shortly. Moreover, numerical results indicate that spacial antenna correlation does not have serious influence on the rate performance, and the BS antennas may be allowed to be placed compactly when M is very large.

We're not able to analyze this paper right now due to high demand.

Please check back later (sorry!).

Generate a summary of this paper on our Pro plan:

We ran into a problem analyzing this paper.

Newsletter

Get summaries of trending comp sci papers delivered straight to your inbox:

Unsubscribe anytime.