Abstract
In this paper, we focus on the asymptotic cross layer analysis of multi-antenna systems with transmit MMSE (Tx- MMSE) beamforming, near orthogonal scheduling and outdated CSIT. To capture the effect of the potential packet outage, we introduce the average system goodput, which measures the average b/s/Hz delivered to the mobiles successfully, as the system performance objective. We derive closed-form expressions for the optimal power and rate allocations as well as a low complexity near orthogonal user scheduling (NEOUS) algorithm to solve the cross-layer optimization problem.We derive the asymptotic order of growth in system goodput for general CSIT error variance σ <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> and found that for sufficiently large nT (number of antennas at the base station) and K (number of users) where K = g <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">-1</sup> (nT) for some strictly increasing function g(x) = o(x), the the system goodput grows in the order of nT log[(1-σ <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> ) logK] when σ <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> < 1. This is the same order of growth as the optimal order of growth in broadcast channels with perfect CSIT and hence, the NEOUS is order-optimal. On the other hand, we need exponentially larger K to compensate for the penalty in multiuser diversity gain due to CSIT errors.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call