About testing of mixed traffic hypothesis

This article proposes some test for presence in traffic two different independent components with a single Hurst parameter $H$. We use $\alpha$-stable Levy motion and fractal Brownian motion as models for $\alpha$- and $\beta$-components respectively. The test statistic is based on frequency-scale sum of logarithms of the wavelet-coefficients absolute values and asymptotically converge to a normal distribution under null ($\beta$-traffic) and alternative ($\alpha +\beta$-traffic) hypothesis.