DETECTION OF NON-GAUSSIAN SIGNALS IN NON-GAUSSIAN NOISE USING THE BISPECTRUM

Bispectrum analysis provides a means of testing a stochastic time series for non-Gaussianity (including nonlinearity). For cases of practical interest, the non-Gaussian time series may be partially masked by either Gaussian or non-Gaussian noise. In this paper, we cast the problem of detecting a non-Gaussian time series in the presence of additive Gaussian or non-Gaussian noise into a classical hypothesis testing framework, using the sample bispectrum as the test statistic. The power of the test is demonstrated as a function of signal-to-noise ratio, the degree of skewness of the signal, and processing parameters. The results are compared to the power of a classical energy detection test. © 1990 IEEE