A conflict-based framework for proactive risk assessment of electric bicycle-motor vehicle interactions at signalized intersections.
Abstract
Objective
The proposed research uses a novel surrogate safety outcome called Maximum Residual Reaction Time (MRRT) to design a comprehensive risk assessment model for analyzing interactions between electric bikes and motor vehicles (EBs-MVs) with regards to signalized intersections. The overall aim is to enhance pro-active injury prevention by improving short-term risk detection and estimation for vulnerable road users.
Methods
Data were collected from video recordings at six signalized intersections, including 2,732 EB-MV interaction trajectories. The study proposes MRRT to measure maximum evasive time, using kinematic data from both vehicles. Enhancing this assessment, two derivative metrics are introduced: time-exposed MRRT (TE-MRRT) and time-integrated MRRT (TI-MRRT). Also, to facilitate autonomous assessment, a scheme has been proposed to identify four types of EB-MV conflicts. Risk thresholds were established via hierarchical clustering and validated by 400 manually observed conflict events.
Results
Conflicts with higher approach speeds and smaller minimum passing distances showed distinctly higher risk levels based on MRRT. The combined TE-MRRT and TI-MRRT model outperformed traditional indicators such as time-to-collision (TTC) and post-encroachment time (PET) in terms of 10-15% higher precision and up to 40% greater accuracy in identifying dangerous conflicts. Distinct thresholds for all four conflict types were successfully established, reflecting realistic variations in EB-MV behavior.
Conclusions
The MRRT-based framework offers a comprehensive and transferable scheme for estimating the risk of a near-crash event and mitigating prospective injuries for e-bicyclists in mixed-traffic situations. It makes safety management more effective by integrating both road-engineering and public health-oriented approaches.