On the identification of risky lane changes: 2D time-to-collision and connected vehicle data

Lane changes could either be the triggers or evasive actions of traffic conflicts. It is essential to accurately identify safety-critical lane changes to lay the groundwork for further studies. Traditional surrogate safety measures (SSMs) are not suitable for the lane change scenarios collected by connected vehicles (CVs). To bridge this gap, this study proposes two-dimensional time-to-collision (2D-TTC) for CV data to identify risky lane changes. The large-scale connected vehicle data from the Safety Pilot Model Deployment (SPMD) are used. The results show that the lateral conflicts captured by 2D-TTC are highly correlated to the recorded sideswipe crashes. The whole process of conflict can be reflected by 2D-TTC accurately. The proposed method demonstrates great potential to be used for traffic safety assessments and applied in collision avoidance systems.