Stochastic Models for Parallel Road Tracks

Abstract

In this paper, we propose a method for constructing road roughness on parallel tracks. This method uses empirical power spectral densities and coherence of the left and the right-tracks estimated by non-parametric methods. The proposed method is based on a recently developed subspace-based spectrum estimation algorithm for multi-input/multi-output systems. Positivity of the estimated spectra are guaranteed by a scheme built-into this subspace algorithm. Numerical examples illustrating properties of the proposed method are presented and comparisons are made with isotropic road models.