Acceleration and Deceleration Behavior in Departing and Approaching Sections of Curve Using Naturalistic Driving Data

Driver safety is one of the prime concerns of traffic engineers in the design and construction of highways. Highway collision causes both social and economic losses. So the primary focus is always there in understanding the driver behavior over various geometric sections for the development of safe geometric roads. So in this study, an attempt was made to understand the acceleration and deceleration behavior over the horizontal curved sections. To do so a four-lane divided highway was considered and continuous data of the vehicle was collected using GPS techniques with the help of 30 distinct drivers. Each of these drivers drove around a stretch of 40 km with the attached GPS device to the vehicle. Using this data six distinct acceleration and deceleration regression-based models are developed corresponding to the cases of maximum, minimum and average acceleration and deceleration. From these models, great insight on how the geometric parameters are influencing this driver’s acceleration and deceleration behavior is obtained. With these models, it is observed that radius and previous tangent length are the most influential parameters affecting driver behavior while approaching the curve. Radius and vertical gradient are found to be the highly correlated parameters with the acceleration behavior on the curve.