Effect of Combined Horizontal and Vertical Curve on Vehicle Longitudinal Acceleration of Mountainous Freeway

Longitudinal acceleration of a vehicle represents the change of driver’s speed, and the drastic change of speed indicates undesirable designs. However, for the adjacent combined sections, there is still no scientific and effective acceleration prediction method. To accommodate this limitation, a new modeling approach based on frequent and equidistant longitudinal acceleration measurements was proposed in this paper. The Tongji University driving simulator was used to conduct the experiment and collect free-flow longitudinal acceleration data. Longitudinal acceleration profile was derived and grade, curvature, etc., extracted from upstream and downstream segments were used to estimate a generous longitudinal acceleration prediction model. The model results show that the average slope, slope difference, and average curvature difference between the upper and lower reaches is positively correlated with the longitudinal acceleration, and the average curvature value is negatively correlated with the longitudinal acceleration.