Stochastic modelling of pavement roughness

Abstract

Pavement roughness is usually characterized by a one figure statistic of pavement profile data. This approach discards a rich body of useful pavement information in the pavement profile data. In this study, informative parametric models for pavement roughness are developed where pavement roughness is treated as a stochastic signal. Based on this stochastic modeling of roughness, a new stochastic roughness index (SRI) is proposed. Case studies conducted on 18 FACE Dipstick profile data and 25 profile data collected using an Ames Profilograph shows that the new SRI, irrespective to the profiling devices, has a good correlation with the International Roughness Index.