Title :
Cumulant-based adaptive FIR filtering for frequency estimation
Author :
Dianat, Soheil ; Venkataraman, S.
Author_Institution :
Dept. of Electr. Eng., Rochester Inst. of Technol., NY, USA
Abstract :
Proposes an alternative technique for adaptive line enhancers (ALE). The new technique uses the fourth order cumulant, which is insensitive to white or colored Gaussian noise. The conventional ALE relies on measurements of the autocorrelation and/or power spectrum of the observations, or on the assumption that the additive noise is white. The autocorrelation is affected by additive Gaussian noise, and in many practical situations, the noise is spatially or temporally correlated. The proposed technique is insensitive to additive Gaussian noise. This is the main advantage of this new technique. Computer simulation has been performed.<>
Keywords :
adaptive filters; correlation theory; digital filters; filtering and prediction theory; parameter estimation; statistical analysis; adaptive line enhancers; additive Gaussian noise; autocorrelation; computer simulation; cumulant-based adaptive FIR filtering; fourth order cumulant; fourth order statistics; frequency estimation; Adaptive filters; Additive noise; Autocorrelation; Filtering; Finite impulse response filter; Frequency estimation; Gaussian noise; Line enhancers; Noise measurement; Power measurement;
Conference_Titel :
Spectrum Estimation and Modeling, 1990., Fifth ASSP Workshop on
Conference_Location :
Rochester, NY, USA
DOI :
10.1109/SPECT.1990.205583