Estimating International Roughness Index from noisy profilograph measurements

In this paper, we propose a method for estimating the international roughness index (IRI) of a pavement profile using data collected by a computerized profilograph making noisy measurements. We begin by modeling the pavement as a nonstationary stochastic process and show that, despite this nonstationarity, we can exploit the design of the profilograph to derive a Wiener-deconvolution filter for estimating the pavement profile. We demonstrate the performance of this approach via numerical simulation and analysis on actual datasets.