Cross-vehicle injury spillovers of ADAS functions in two-vehicle rear-end and same-direction sideswipe crashes
Parvez Anowar, Mohamed Abdel-Aty, Chenzhu Wang, Shaoyan Zhai
Abstract
Advanced driver assistance system (ADAS) functions are evaluated mainly by the crashes they prevent, and little is known about how they relate to the injuries sustained in the crashes that would still occur. Because two colliding vehicles share one impact, a function on one vehicle may also be associated with the other driver's injury. This study examines whether an ADAS function on either vehicle in such two-vehicle crashes is associated with either driver's injury. It focuses on rear-end (n = 2,273) and same-direction sideswipe crashes (n = 1,216) from Texas police reports across 2023 to 2025. Striking and struck roles were assigned within each crash, and the ADAS functions of each vehicle were identified from vehicle identification numbers. For each crash type, a bivariate probit model, with random parameters for rear-end crashes, jointly estimates whether each driver was injured. Significant cross-equation correlations (0.559 and 0.375) support the joint specification. Electronic stability control on the striking vehicle is associated with a lower struck-driver injury probability in both crash types, by 0.036 to 0.073 in rear-end and 0.102 in sideswipe crashes. Lane departure warning on the striking vehicle shows a similar association in sideswipe crashes (about 0.078). Traction control with anti-lock braking on the struck vehicle shows the largest association among the ADAS functions with its own driver's injury probability (0.097 and 0.103 lower). No struck-vehicle function is associated with the striking driver's injury, and forward-looking functions show no association with either driver's injury, consistent with their benefit lying mainly in crash prevention. The findings suggest that the protective benefits of stability and lane departure functions may reach beyond the equipped vehicle to the other driver, strengthening the case for their wider fleet adoption.
Create a lesson
Related papers
Quantifying and Predicting Pitch Sequence Effectiveness Using Visual Deception Metrics
Joseph DiGiovanni, Justin Kurland
Hierarchical Bayesian Joint Modeling of Athlete Performance Trajectories and Latent Peak Ages
K. Ken Peng, Tianyu Guan, Xiong Zhao
Reducing the Uncertainty in Determining the Origin of Natural Commodities via Earth Embeddings
Sebastián Jiménez, Willem Waegeman, Victor Deklerck et al.
Shape-Preserving Treatment Effects for Functional Outcomes
Agathe Chabassier, Joris Olivier, Julie Josse et al.
Dynamic conditional density network of wind vectors
Keigo Sato, Yasumasa Matsuda
Covariance spectra of random linear systems near instability
Pengbin Feng