GLOBECOM 2005 Tutorial

Friday, 2 December
Majestic Ballroom F / Level Two / Renaissance Grand Hotel

Title: TU08 Capacity Limits of MIMO Systems

Duration: Half Day/9:00AM-12:00 Noon

Presenters:
Andrea Goldsmith, Stanford University, USA
Syed Ali Jafar, University of California, Irvine, USA
Nihar Jindal, University of Minnesota, USA
Sriram Vishwanath, University of Texas, Austin, USA

Abstract:
Multiple-input multiple-output communication systems have been the subject of great interest in the research community and in industry over the past decade because of their potential for extremely high data rates coupled with robustness to fading. However, the benefit of MIMO systems crucially depends on assumptions regarding the underlying time-varying channel model and the ability to track the channel. In this tutorial we provide an overview of information theoretic limits of point-to-point and multi-user MIMO systems under realistic assumptions regarding channel conditions. For time-varying channels, there are multiple definitions of capacity (i.e. ergodic and outage capacity) and many channel fading models (i.e. antenna correlation, line-of-sight propagation, etc.) to consider. We begin with a comprehensive summary of capacity results for point-to-point MIMO channels under different fading models and assumptions regarding the ability to track channel variation. These results indicate that the capacity gain of MIMO depends heavily on the amount of channel state information available at the receiver and transmitter, the channel fading model, and the system SNR. Next we consider multi-user MIMO systems - including MAC/broadcast channels, cellular systems, and ad-hoc networks - where we show that the potential benefit of MIMO can be even larger than for point-to-point systems. We give capacity results for different fading models and show that channel state information is the key to unlocking MIMO benefits in a multiuser setting. MAC/broadcast channel duality, multiplexing/diversity tradeoffs, user cooperation, and cross-layer design will also be discussed. Open research problems will be presented throughout the tutorial.

Motivation:
The goal of the tutorial is to provide attendees with an understanding of both the fundamental limits of MIMO and the settings in which the benefits of MIMO may be practically realized.