WSEAS Transactions on Signal Processing
Print ISSN: 1790-5052, E-ISSN: 2224-3488
Volume 11, 2015
The Cramer-Rao Bound for 3-D Frequencies in a Colored Gaussian Noise
Authors: , , ,
Abstract: Estimation of model parameters (3-D frequencies), based on the high resolution spectral analysis methods known by their performances and their precision such as 3-D ESPRIT, remains a problem which is essential in the modeling of the signals by a sum of 3-D complexes exponential (3-D SCE model) embedded in an additive gaussian noise. Indeed, good results are obtained when the noise is white and by using the Second Order Statistics (autocorrelations), but if it becomes colored, the results are degraded which forces us to remedy this problem, to think about the Higher Order Statistics (cumulants). To verify the efficiency of estimators of 3-D frequency, we calculate the asymptotic Cramer-Rao Bound (CRB).
Search Articles
Keywords: Spectral Analysis, High Resolution, 3-D ESPRIT, Second Order Statistics, Higher Order Statistics, Fourth Order Cumulant, Cramer-Rao Bound
Pages: 227-234
WSEAS Transactions on Signal Processing, ISSN / E-ISSN: 1790-5052 / 2224-3488, Volume 11, 2015, Art. #27