Predictive Speed Models for Two-Lane Rural Roads Using GPS Equipment

This research is aimed at verifying the effectiveness of a low cost global positioning system (GPS) installed in a medium passenger car for modelling drivers speed behaviour and providing predictive models of operating speed and speed differential. This study proposed new speed models specific for Italian two-lane rural roads, based on the analysis of 36 speed profiles performed along a road segment characterised by 13 tangent-curve configurations. A predictive model for operating speed on tangent identified the length of the tangent and the local grade as the most significant explanatory variables. Moreover a predictive model of operating speed on curve revealed the length and the radius of the horizontal curve and the local grade as predictors. Finally tangent-curve speed variations were evaluated and a model for predicting 85MSR was developed. The length of the approaching tangent and the deflection angle were found the most significant predictors.