|
Activity Number:
|
383
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 4, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Physical and Engineering Sciences
|
| Abstract - #304650 |
|
Title:
|
Prediction Algorithm Using a Widely Lineal Processing
|
|
Author(s):
|
Rosa M. Fernandez-Alcala*+ and Javier Moreno-Kayser and Jesus Navarro-Moreno and Juan C. Ruiz-Molina
|
|
Companies:
|
University of Jaen and University of Jaen and University of Jaen and University of Jaen
|
|
Address:
|
Campus Las Lagunillas S/., Jaen, International, 23071 , Spain
|
|
Keywords:
|
complex-valued improper times series ; linear least mean-square error estimation problems ; second order stationary signals ; widely linear processing
|
|
Abstract:
|
The prediction problem is analyzed for second-order stationary complex-valued signals. This type of signal is characterized by having constant mean value and both the covariance and complementary covariance functions only depend on the difference between the involved times instants. Following a widely linear processing approach, i.e. under an improper treatment of the problem, a recursive algorithm for the computation of the predictor is devised. The main advantage of the proposed widely linear predictor with respect to the standard prediction techniques is basically the fact that the first one presents a better behavior in the sense of its mean square error. This work is supported in part by project MTM2007-66791 of the Plan Nacional de I+D+I, Ministerio de Educación y Ciencia, Spain, which is financed jointly by the FEDER.
|
