Title :
Recognition of non-Gaussian signals against a background of noise using higher order statistics
Author :
Kudriavtseva, N.V. ; Tykhonov, V.A. ; Netrebenko, K.V.
Author_Institution :
Dept. of Electr. Eng., Univ. of Pardubice, Pardubice, Czech Republic
Abstract :
The aim of the paper is to study the possibility of advancing noised non-Gaussian processes recognition using the features based on higher-order statistics. Several new recognition features based on the higher-order statistics, the basis of advancing recognition results in the presence of noise, decision rules and recognition system frameworks we proposed in the paper. The efficiency test of the proposed method is performed by statistical modeling. We proposed features of recognition based on the third-order statistics.
Keywords :
feature extraction; higher order statistics; signal denoising; statistical analysis; decision rules; higher order statistics; noise background; noised nonGaussian processes; nonGaussian signal recognition; statistical modeling; third-order statistics; Filter banks; Higher order statistics; Interference; Mathematical model; Noise; Stochastic processes; Training; Autoregressive models; Higher order statistics; Power spectrum density;
Conference_Titel :
ELMAR (ELMAR), 2014 56th International Symposium
Conference_Location :
Zadar
DOI :
10.1109/ELMAR.2014.6923351