Combinatorial fractal Brownian motion model (Q5926402)

A concept of combinatorial fractional Brownian motion model is defined: The random process is the sum of two correlated fractional Brownian motions. A method to estimate the two Hölder exponents is proposed. The link is done between these estimators and the fractal parameter estimator of a single fractional Brownian motion. The authors declare satisfactoring results when the method is applied to actual data: The motivation was to determine the non-scaled interval when processing radar clutter. Finally, it is concluded that when combinatorial model is observed on different scales, one of the fractional structures is more obvious. About the same theme, look also at \textit{A. Benassi}, \textit{S. Cohen}, \textit{J. Istas} and \textit{S. Jaffard} [Stochastic Processes Appl. 75, No. 1, 31-49 (1998; Zbl 0932.60037)] in the case of the sum of two independent generalized rational Brownian motions, where the estimators are similar, but their limit laws are given.