A fast time-domain estimator of the bandwidth of a Gaussian process

A fast time-domain technique is proposed to estimate the root mean square (rms) bandwidth of a Gaussian random process. The method is based on averaging the absolute values of the derivative of a signal at zero crossings. For processes with Gaussian power spectral density, the performance of the proposed bandwidth estimator has been studied by using Monte Carlo simulations. The structure of the estimator makes it well suited to practical implementation with the use of a digital storage oscilloscope.