Traffic Conflict Techniques (TCTs) estimate safety proactively by identifying critical vehicle interactions that could have led to a crash. Existing TCTs constrained by ideal assumptions of interaction points and constant current speed often fail to consider the influence of neighboring vehicle density. The key challenge associated with incorporating the heterogeneity and disordered nature of traffic in estimating TCTs lies in varying vehicle dynamics and the collision scenarios. This study models a dynamic two-dimensional surrogate safety approach called the Vehicle Safety Envelope (VSE), which is the elliptical safety boundary around a vehicle that is required for safe and comfortable maneuvering through traffic. VSE is then utilized as a spatial filter to identify and classify vehicle-vehicle conflicts using trajectory data from five signalized intersections in India. Results showed that lateral and longitudinal safety clearance distance, the key parameters for defining VSE, were found to have a linear dependency on vehicle size and speed, indicating that at a higher speed, a larger VSE is required to maintain the same level of safety. The VSE approach was validated through comparison with two well-established surrogate safety measures, namely Modified Time to Collision and Deceleration to Avoid Crash, along with a Severity Index (SI) developed to quantify conflict risk. Smaller vehicles such as motorised two-wheelers and motorized three-wheelers were found responsible for a larger proportion of side-swipe conflicts. These results emphasize the dynamic nature of VSE, demonstrating its potential for proactive conflict estimation. The VSE could enhance safety perception for decision modules if implemented in an autonomous driving environment.
Avathkattil et al. (Wed,) studied this question.