Empirical model of roughness effect on vehicle speed

Vehicle speed–roughness relationship has a significant research gap in life cycle assessment model. The current available models describing the roughness effect on vehicle speed are very limited and outdated. In this paper, 32 individual pavement sections, each of which has roughness data of up to 8 years, were selected to develop the model. The roughness data cover a wide range, and the selected pavement sections contain both flexible and rigid pavement types and various numbers of lanes. Involved regression variables include the following: vehicle speed, roughness, volume–capacity ratio, pavement type, number of lanes and speed limit. Analysis of variance was first performed, indicating that pavement type and speed limit are not significant factors influencing the average vehicle speed. Following, strict statistical technique was used to correct the unobserved heterogeneity during the regression using a one-way fixed random model. The obtained regression model reveals that the average vehicle speed decreases 0.0083 mph with every 1 in/mi increase of the roughness ( − 0.84 km/h per m/km).