Acceleration and Deceleration Calibration of Operating Speed Prediction Models for Two-Lane Mountain Highways

Abstract

In this study, to calibrate the acceleration/deceleration rates in operating speed prediction models, driving tests were performed on two-lane mountain highways using passenger cars, and continuous longitudinal acceleration data were collected. The acceleration/deceleration rates of large vehicles such as coaches and heavy trucks were also measured from the roadside. The collected data were used to determine the cumulative frequency profiles and statistical distribution characteristics of the peak acceleration for different types of vehicles. Based on the established correlations, models of acceleration/deceleration versus road geometry features were developed. The key findings of this study are as follows: First, for passenger cars, the deceleration rate is obviously greater than the acceleration rate. Therefore, it would be inappropriate to use the same value for both the rates in speed prediction models. Instead, acceleration and deceleration should be calibrated separately. Second, none of the breakpoints on the acceleration/deceleration cumulative frequency curves appears at the 85th percentile; rather, they appear closer to the 95th percentile. Therefore, the common practice of using the 85th percentile in highway design should probably be reconsidered. Third, for large coaches, the acceleration and deceleration rates are quite similar, whereas for heavy trucks, the deceleration rate is significantly larger than the acceleration rate. The acceleration and deceleration rates of heavy trucks are both lower than those of coaches. Finally, for passenger cars driving around a curve, a negative correlation exists between the acceleration/deceleration rate and curve radius and a positive correlation exists between the acceleration/deceleration rate and deflection angle.