Higher-Order Spectral Analysis Toolbox

A collection of algorithms for analyzing non-Gaussian signals

The Higher-Order Spectral Analysis Toolbox contains specialized tools for analyzing signals using the cumulants, or higher-order spectra, of a signal. The Toolbox features a wide range of higher-order spectral analysis techniques, providing access to algorithms at the forefront of signal processing technology.

Features

Higher-order spectrum estimation
Conventional or parametric approaches
Magnitude and phase retrieval
Adaptive linear prediction
Harmonic retrieval
Quadratic phase coupling
Time-delay estimation
Array signal processing

Highlights

Advantages of Higher-Order Spectra. The Higher-Order Spectral Analysis Toolbox allows you to analyze signals corrupted by non-Gaussian noise or signals arising from a nonlinear process. Higher-order spectra, which are defined in terms of the cumulants (higher-order moments) of a signal, contain additional information that is not conveyed by the signal's autocorrelation or power spectrum. Higher-order spectra are useful because they:

Suppress additive colored Gaussian noise of an unknown power spectrum
Identify nonminimum phase signals
Extract information due to deviations from Gaussianity
Detect and characterize nonlinear properties in signals

Applications of higher-order spectral analysis include acoustics, biomedicine, econometrics, exploration seismology, nondestructive testing, oceanography, plasma physics, radar, sonar, and speech.

Advanced Signal Processing Techniques. The principal functions in the Toolbox support higher-order spectra, cross-spectra, linear prediction models, and time-delay estimation. A comprehensive 130-page tutorial features 30 examples and more than 50 illustrations.

Stefan Steinhaus, webmaster@steinhaus-net.de