Mohamed BOULOUIRD | Université Cadi Ayyad (original) (raw)

Related Authors

Jorge Costa

Jorge Costa

ISCTE - University Institute of Lisbon (ISCTE-IUL)

Marc Champagne

David Seamon

Giuseppe Ceraudo

Levent Tasci

Jana Javornik

Gwen Robbins Schug

Gabriel Gutierrez-Alonso

John Sutton

Uploads

Papers by Mohamed BOULOUIRD

Research paper thumbnail of HOS-Based Algorithm for Blind Identification of MA Models Using BPSK Signals

Recent Patents on Computer Sciencee, 2009

Research paper thumbnail of A New Algorithm for Blind Identification of MA Models Using Cumulants

This paper addresses the problem of blind identifi- cation of Moving Average (MA) models. We prop... more This paper addresses the problem of blind identifi- cation of Moving Average (MA) models. We propose a new al- gorithm utilizing Higher-Order Statistics (HOS), namely fourth- order cumulants and autocorrelation functions of output signal of the MA model, to estimate the parameters of this model. A general relationship linking cumulants of the output signal of the model and the coefficients of this model is exploited to generate a Least-Squares (LS) solution. This new method is compared with C(q;k;0) algorithm, named also Giannakis algorithm. Input signal is considered like digital communications signal, non observable, but its statistical properties are known. Simulation results are presented demonstrating the performance of this algorithm.

Research paper thumbnail of HOS-Based Algorithm for Blind Identification of MA Models Using BPSK Signals

Recent Patents on Computer Sciencee, 2009

Research paper thumbnail of A New Algorithm for Blind Identification of MA Models Using Cumulants

This paper addresses the problem of blind identifi- cation of Moving Average (MA) models. We prop... more This paper addresses the problem of blind identifi- cation of Moving Average (MA) models. We propose a new al- gorithm utilizing Higher-Order Statistics (HOS), namely fourth- order cumulants and autocorrelation functions of output signal of the MA model, to estimate the parameters of this model. A general relationship linking cumulants of the output signal of the model and the coefficients of this model is exploited to generate a Least-Squares (LS) solution. This new method is compared with C(q;k;0) algorithm, named also Giannakis algorithm. Input signal is considered like digital communications signal, non observable, but its statistical properties are known. Simulation results are presented demonstrating the performance of this algorithm.

Log In