sarah-saeed / bda Goto Github PK

This repository contains MATLAB code that was used in the following paper:

  "Zero-forcing methods for downlink spatial multiplexing in multiuser MIMO
  channels", Quentin H. Spencer, A. Lee Swindlehurst and Martin Haardt, IEEE
  Transactions on Signal Processing, Feb. 2004, pp. 461-471.

  (http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=1261332)

I get periodic requests for the code used to generate the results in the
article, so I am putting this here. Note that the code was mostly written
in 2002 or earlier. Some of the code depends functions I had written at the
time that did not exist in MATLAB at the time (i.e. cdf, ccdf, random). I
have attempted to include my versions of these functions so that all of the
code runs correctly, because the functions now present in MATLAB may not
do exactly the same thing or have the same interface.

The main function for computing capacity of Gaussian multi-user MIMO
channels using the BD algorithm is bd_capacity.m. I have also included the
function bd_plots.m, which was used to produce many of the plots in the
paper. The plots were generated using GNU Octave rather than MATLAB, and
they contained a number of Octave-specific function calls that are no
longer supported in modern Octave or in MATLAB, so I have attempted to
remove all of those, which means some plot labels and annotations will
look different. All of the plots can be successfully produced in Octave
but have not been tested in MATLAB.

Several papers subsequent to my 2004 paper have found more efficient
ways of computing the same solutions.  If anyone is interested in implementing
faster solutions or updating the code in other ways to take advantage of
more modern MATLAB or Octave versions, I welcome any contributions.