Computational and Applied Math Proseminar

Department of Mathematics and Statistics
Arizona State University

Tuesday, May 4, 2004 3:15 p.m. (NOTE LATER TIME!) in PSF 208

Gautam Pendse

Department of Mathematics and Statistics

Optimization Based Formulations using the Weyl Criterion for Input Signal Design in System Identification

Abstract System identification is the science of modelling the behavior of a dynamical system based on input-output data. Input signal design is a critical aspect of system identification. Highly interactive systems tend to have a preferred or a high gain direction. It is desirable to have an input signal that will promote output from the system in all directions.

We present a framework for input signal design that achieves this objective using cleverly designed constraints that force uniform distribution in the output state space of the plant. To test the new framework we perform three case studies. For each case study we formulate an optimization problem in accordance with this new framework and solve it using the NLP solver KNITRO 3.1 and the modelling language AMPL.

The results in all three case studies show the promising nature and generality of this approach which seems to be applicable to both linear and nonlinear models for the plant.

For further information please contact: mittelmann@asu.edu