Title :
Estimating the number of sinusoids in non-Gaussian noise using cumulants
Author :
Rao, Satyanarayan S. ; Vaidyanathan, Chandra
Author_Institution :
Dept. of Electr. Eng., Villanova Univ., PA, USA
Abstract :
The problem of determining the dominant singular values of a cumulant matrix is considered. A criterion for rank determination which is quite general and can be used for any matrix is presented. The problem addressed is important due to the surge of new techniques in the cumulant domain, most of which require knowledge of the rank of the matrix. The criterion performs reasonably well under high and low signal-to-noise ratio conditions
Keywords :
eigenvalues and eigenfunctions; matrix algebra; noise; spectral analysis; cumulant matrix; cumulants; nonGaussian noise; number of sinusoids; rank determination; singular value decomposition; spectral estimation; Additive noise; Autocorrelation; Degradation; Equations; Estimation error; Gaussian noise; Information analysis; Random variables; Signal resolution; Singular value decomposition;
Conference_Titel :
Acoustics, Speech, and Signal Processing, 1991. ICASSP-91., 1991 International Conference on
Conference_Location :
Toronto, Ont.
Print_ISBN :
0-7803-0003-3
DOI :
10.1109/ICASSP.1991.150201