On Achievable Multiuser Diversity Gain of Limited-Feedback-Based Block Diagonalization in Interference-Limited MIMO Broadcast Channels

The scaling law of sum spectral efficiency achieved by limited-feedback-based block diagonalization with semi-orthogonal user selection in an interference-limited multiple-input-multiple-output broadcast channel is investigated. An interference-limited regime is realized by assuming a large transmit power. An analysis of the distribution of the channel quality indicator of scheduling makes it possible to characterize the limiting behavior of the spectral efficiency. From the limiting behavior, it is shown that the achievable multiuser diversity gain can be significantly increased as the system approaches the interference-limited regime from large user regimes.

I. Introduction

With perfect channel state information (CSI) at the transmitter (CSIT), previous research has shown that the sum capacity in a multiple-input multiple-output (MIMO) broadcast channel scales like as increases, where is the number of transmit antennas and is the number of users; the factor is achieved by exploiting multiuser diversity from a sufficiently large [1]. In [2], a scheduling algorithm called semi-orthogonal user selection (SUS) was proposed, and with SUS, zero-forcing beamforming (ZFBF) was shown to achieve the optimal scaling law .