China
Africa
Explore chapters and articles related to this topic
Ultra-Wideband Radar Receivers
Published in James D. Taylor, Introduction to Ultra-Wideband Radar Systems, 2020
James D. Taylor, Elizabeth C. Kisenwether
The higher order spectra, or polyspectra, defined in terms of cumulants, or higher order statistics, do contain signal phase relations which can distinguish one waveform from another and find nonlinearities which aid signal processing and identification. The motivations for using higher order signal processing include techniques to: (1) suppress additive colored Gaussian noise; (2) identify or reconstruct nonminimum phase signals; (3) find information due to deviations from Gaussianity; and (4) detect, characterize, and identify nonlinear properties in signals and systems. Nikias and Mendel provide a good introduction and overview in IEEE Signal Processing Magazine.40
Signal Modeling Using Spatial Filtering and Matching Wavelet Feature Extraction for Classification of Brain Activity Patterns
Published in Mridu Sahu, G. R. Sinha, Brain and Behavior Computing, 2021
Vrushali G. Raut, Sanjay R. Ganorkar, Supriya O. Rajankar, Omprakash S. Rajankar
Though the second-order statistics given earlier are significant for representing the features, they are not effective especially for indicating nonlinearities of the signals. For that, higher-order statistics (HoS) containing higher-order moments (m3, m4,…) and nonlinear combinations of these moments, known as cumulants, can be termed as a valuable descriptor. Skewness, a third-order cumulant, is a measure of symmetry or the lack of symmetry of the distribution given by equation (5.24). Kurtosis, a fourth-order cumulant, is a measure of the distribution in terms of heavy-tailed or light-tailed relative to a normal distribution and given by equation (5.25) [47,48]. These HoS features are suggested for representing the dynamics of the signal in this work. b=1n∑j=1N(xj−x¯σ)3b=1n∑j=1N(xj−x¯σ)4
Signal fingerprint feature extraction and recognition method for communication satellite
Published in Connection Science, 2022
Higher-order statistics are random variables whose statistical order is greater than or equal to the third order. As the need for signal analysis increases, some signal analysis challenges also arise. There are many non-stationary signals in some practical scenarios, and traditional low-order statistics can no longer meet their analysis needs. Therefore, higher-order statistics are applied to the field of signal analysis. Compared with ordinary low-order statistics, high-order statistics can better preserve the phase information of signals in the process of signal processing and have a certain role in suppressing white Gaussian noise. According to these characteristics, compared with ordinary low-order statistics, high-order statistics have advantages in non-stationary signal processing.