Analysis of Discretionary Lane Changing Parameters on Freeways

Abstract

A lane changing event, at the instant when a vehicle crosses the lane marker, involves up to five vehicles: the subject vehicle, preceding and following vehicles in the original lane, and the preceding and following vehicles in the target lane. Understanding the interactions of the subject vehicle with the surrounding vehicles is fundamental to the study of the safety and modeling of lane changing behavior. This research studies the statistical properties of 10 lane changing parameters. These parameters describe the gaps, times to collision between vehicles and the subject vehicle's speed. The parameter values were extracted from the vehicle trajectory data in the Next Generation Simulation data sets. The results show that (i) all the parameters are positively correlated with each other; (ii) the gaps and distance are best described by the log-normal distribution; (ii) the times to collision are best described by the Laplace distribution; (iii) the speed may be described by the log-logistic distribution and the normal distribution. The results suggest that using one or few selected parameters may be sufficient to quantify the risk of a lane changing event.