### Tuesday, July 15

10:30 AM-12:30 PM

*Law School, Room 190*

## MS22

Structured Total Least Norm Approximation Methods and Applications

For many important applications in signal processing, parameter estimation, and system identification, it is necessary to compute a good approximate solution to an overdetermined system A(alpha)x ~ b, where the parameter vector alpha and the amplitude vector x are to be estimated. The data vector b may be in error, due to noise in the measured signal. For most applications the matrix A(alpha) has a structure (such as Toeplitz or Vandermonde) which should be preserved. The methods Separable Nonlinear Least Squares, Constrained Total Least Squares, Structured Total Least Squares, and Structured Nonlinear Least Norm have all been developed to solve problems of this type. These methods, and their relationship, will be described, and their use in specific applications will be summarized in this minisymposium.

**Organizers: Haesun Park, University of Minnesota, Minneapolis; and
J. B. Rosen, University of California, San Diego**

**10:30 Structured Total Least Squares: Algorithms and Applications**
- Philippe Lemmerling, Sabine Van Huffel, and Bart De Moor, Katholieke Universiteit Leuven, Belgium
**11:00 Structured Approximation Using the L1 Norm**
- Haesun Park, Organizer; John Glick, University of San Diego; and J. B. Rosen, Organizer
**11:30 Regularized Constrained Least Squares Image Restoration**
- N. P. Galatsanos, Illinois Institute of Technology; A. K. Katsaggelos, Northwestern University; and V. Z. Mesarovic, Illinois Institute of Technology
**12:00 Parsimonious Least Norm Approximation**
- P. S. Bradley and O. L. Mangasarian, University of Wisconsin; and J. B. Rosen, Organizer

AN97 Homepage | Program Updates|

Registration | Hotel and Dormitory Information | Transportation | Program-at-a-Glance | Program Overview

*MMD, 4/21/97*
*tjf, 5/28/97*