Examining two-wheelers' overtaking behavior and lateral distance choices at a shared roadway facility

Abstract

This article investigates the lateral distance that overtaking two-wheelers (bicycles, e-bikes, and e-scooters) keep from automobiles at shared traffic streets. A video-based computer vision technique is used to track road users, collect their trajectories, and measure the lateral distance. A full Bayesian logit model is developed to examine the factors that affect the likelihood of two-wheelers accepting the critical lateral distance that is defined as the 10th percentile lateral distance. The results show that (a) the average lateral distance between overtaking two-wheelers and automobiles is 1.54 m, (b) the lateral distance for bicycles is significantly larger than that for e-bikes and e-scooters, (c) the lateral distance follows a best-fitted Gamma distribution. Further results from the full Bayesian logit model show that (a) two-wheelers type, evasive action manner, occurrence of a platoon of moving two-wheelers, and two-wheelers' yaw rate ratio are significantly positively related to the probability of two-wheelers accepting the critical lateral distance and (b) the presence of heavy vehicles and the speed difference between two-wheelers and interacting automobiles are negatively associated with the above probability.

Keywords:

• shared roadway
• lateral distance
• logit model
• full Bayesian
• computer vision