| Paper: | SPTM-P3.12 | ||
| Session: | System Modeling, Representation and Identification | ||
| Time: | Tuesday, May 16, 16:30 - 18:30 | ||
| Presentation: | Poster | ||
| Topic: | Signal Processing Theory and Methods: System Modeling, Representation, and Identification | ||
| Title: | A NEW TVAR MODELING IN CASCADE FORM FOR NONSTATIONARY SIGNALS | ||
| Authors: | Abdullah Zaman, Xiaoulin Luo, Mohammad Ahad, Mohammed Ferdjallah, University of Tennessee, United States | ||
| Abstract: | In the nonstationary process identification, time varying autoregressive (TVAR) model may possess temporal model instability in conventional direct form method. In this study, we are proposing a new TVAR cascaded form method, to overcome model instability. The model stability in TVAR cascaded form method is accomplished through the parameterization of time varying pole representations from pole tracking and monitoring methods. Our simulation on synthetic data demonstrates that the cascaded form model stability can easily be achieved, monitored, and controlled. The performance evaluation of TVAR cascade model has shown that the Cartesian coordinate with orthogonal representation performs better than any other pole representation. | ||
©2018 Conference Management Services, Inc. -||- email: webmaster@icassp2006.org -||- Last updated Friday, August 17, 2012