Regression models to predict operating speeds of horizontal curves in mountainous terrain

Abstract

Majority of operating speed models in literature are mainly for flat and rolling terrain. However, mountainous terrain is not common and present different speeding environment to drivers due to steep grades and sharp curves. This study focused on developing models to predict the 85th percentile operating speeds for passenger cars and heavy vehicles on horizontal curves in mountainous terrains. The vehicle speed data used in the study was collected from 80 horizontal curves and approach tangent sections in one travel direction using MH Corbin Traffic Classifiers. Additional data was collected on geometric properties from as-built drawings and field measurements for each curve and tangents. Speed data was retrieved from the classifier and checked for tailgating using the HDM software, and further screened to remove error observations. Vehicles classification by length was done to identify the different types of vehicles. The 85th Operating speeds and mean speeds for passenger cars and heavy vehicles at mid-curve and approach tangent were determined using MS Excel. Operating speed models at mid-curve were developed, estimated and evaluated for passenger cars and heavy vehicles using a vector of exploratory variables. Curve Length (CL), Grade at Curve (GC) and Tangent length (TL) were the most significant variables for the operating speed model for passenger cars and Heavy vehicles with R2 – value of 63% and 69%, respectively. Both models were statistically significant at 95% confidence level. The models developed could be leveraged in geometric design by controlling say tangent length and its vertical grade to yield a consistent and safer alignment design in terms of expected speeds of vehicles.