The zip file from above contains a series of Matlab m-files used to implement the FIR PU IGA proposed in the paper ``Iterative greedy algorithm for solving the FIR paraunitary approximation problem'' as well as in Chapter 4 of my Ph.D thesis ``Optimization Algorithms for Realizable Signal-Adapted Filter Banks.'' There are two main loop m-files that are listed below.
The first file is a main loop test file for the FIR PU approximation problem for the design of principal component filter bank (PCFB) type filter banks, whereas the second file is a main loop test file for the FIR PU interpolation problem. By adjusting the parameters in these m-files, a variety of different design problems can be solved.
The rest of the m-files in the above zip file are functions used to make the Matlab code more modular. Most of these are functions used to carry out operations on multiple-input multiple-output (MIMO) systems, both in the time and frequency domains.
I would like to send a special thanks to Peter J. Acklam for creating a code vectorized version of a multiplication routine that allowed me to carry out MIMO frequency domain multiplication without the use of for loops. This significantly cut down on the time required for implementing the IGA. The Matlab function, ndmtimes.m, as well as several other code vectorized m-files, can be found at the following link.
http://home.online.no/~pjacklam/matlab/software/util/fullindex.html