Effect of correlation between azimuth spread and Rician K-factor on MIMO performance

Abstract

In this paper, we study, for the first time, the effect on multiple-input multiple-output (MIMO) wireless communications system performance of recently measured statistical relationships among the azimuth spread at the base station (AS b ), the azimuth spread at the mobile station (AS m ), and the Rician K-factor. We construct a 4 × 4 MIMO channel model with lognormally-distributed and correlated AS b , AS m , and K and use this model to simulate the performance of the IEEE 802.11n standard. We find that that disregarding the distributions and correlations of these parameters may yield unrealistic MIMO performance expectations. For example, accounting for recently-measured distributions and correlations for AS b , AS m , and K yields more than 4 dB improvement at 10−4 average bit-error rate (BER) for 16-QAM and orthogonal deterministic component of the MIMO channel matrix. While in another situation whose deterministic component is rank-deficient, to achieve the expected BER, we may have to put more power in practice.