[CPCC] Talk by Syed Jafar (Stanford), May 19, 11 AM

[Ender Ayanoglu]

	      MIMO Capacity with Partial Channel State Information

				Syed Ali Jafar
   			     Stanford University

			     Monday May 19, 11 AM
				   ET 331
                   (www.eng.uci.edu/cpcc/?page=directions)

				  ABSTRACT

Although enormous capacity gains have been predicted for single user
multiple antenna wireless channels, these predictions are based on
somewhat unrealistic assumptions about the underlying time-varying channel
model and how well it can be tracked at the receiver as well as at the
transmitter. Also, relatively little is known about the multi-user
multiple-input multiple-output (MIMO) wireless links.  The focus of this
talk is on MIMO capacity results with partial channel state information at
the transmitter (CSIT) or receiver (CSIR) and extensions to multiple user
cellular systems.

The talk consists of the following two sections.

Part 1. Single User MIMO capacity results with channel partially known.
-----------------------------------------------------------------------
The linear growth of capacity with the number of antennas is one of the
most exciting promises of MIMO systems. However, the linear growth assumes
perfect channel knowledge at the receiver and spatially uncorrelated
fading.  In the absence of channel knowledge at the receiver it has been
shown that capacity may not increase at all with the number of antennas.
Channel knowledge at the transmitter can also have a strong impact on the
capacity of MIMO systems. Also, increasing the number of antennas leads to
spatially correlated fading as the inter-antenna spacing is reduced. The
first part of the talk addresses these concerns.  We will present our
results on the capacity of MIMO systems with partial channel knowledge at
the transmitter and/or the receiver, optimality of beamforming, and the
impact of spatial correlations on the channel capacity of a single-user
multiple antenna system.

Part 2. Capacity results for Cellular Multi-User Multi-Antenna Systems
--------------------------------------------------------------------------
The second part of the talk focuses on recent capacity results for a
multi-user MIMO cellular system. The multiple antenna downlink is an
instance of a non-degraded broadcast channel in information theory. The
capacity of such a channel has been an unsolved problem for around thirty
years. In the second part of the talk, we present our recent fundamental
contributions towards solving this long standing open problem. The best
known achievable rates on the MIMO downlink correspond to dirty paper
coding at the base station. We prove optimality of dirty paper coding on
the MIMO downlink with Gaussian inputs by developing an upper bound for
general broadcast channels that is shown to be tight for the multi-antenna
multi-user downlink. We demonstrate symmetries between the uplink optimal
strategy of successive interference cancellation (SIC) and the downlink
optimal strategy of dirty paper coding (DPC).  SIC and DPC are also shown
to be natural solutions to the problem of admission control in a cellular
system where the goal is to accommodate new users in such a way that
previously active users are not affected. Optimal resource allocation on
the MIMO downlink is a non-convex optimization problem. However, using the
concept of uplink-downlink duality we are able to convert it into an
equivalent convex optimization problem on the dual uplink. We develop fast
algorithms to solve this optimization problem. These algorithms are
currently in use at Bell Labs for system level simulations.

We conclude the talk by pointing out the similarities between multiple
antenna systems and CDMA systems and briefly summarizing our results on
adaptive multi-rate CDMA systems.


				BIOGRAPHY

Syed Ali Jafar received the B. Tech. degree in Electrical Engineering
from the Indian Institute of Technology, Delhi, India in 1997, and the
M.S. degree in Electrical Engineering from California Institute of Technology
(Caltech), Pasadena USA in 1999. He is currently working toward the Ph.D.
degree at Stanford University. He was a summer intern with the Wireless
Research group at Bell Laboratories, Lucent Technologies, Crawford Hill, NJ
in the summer of 2000-2001.  His research interests include multiple antenna
systems, spread-spectrum systems and multiuser information theory.


				DIRECTIONS

Directions to the Engineering Tower (ET) are available from
http://www.eng.uci.edu/cpcc/?page=directions