Evaluation of time headway maintenance of bus drivers and its impact on road crashes using ADAS alerts

Rear-end collisions accounted for 21.40% of road crashes in India in 2022, showing how strongly driver behaviour affects safety. Advanced Driver Assistance Systems (ADAS), particularly headway monitoring warnings (HMW), can reduce risks by alerting drivers in real-time when their following distances fall below safe thresholds. This study examines 87,356 HMW alerts from 190 public buses on NH-65, Telangana, triggered when time headway fell below 0.6 seconds. Findings revealed that most alerts occurred at moderate to high speeds, with 74.37% on midblock sections and 59.50% on 4-lane roads. Binary Logistic Regression models analysed fatal and injury crash risks using HMW alerts and average headway gaps. Model performance was tested using sensitivity checks, ROC-AUC analysis, and behaviour impact analysis. The fatal crash model performed very well (AUC = 0.83), while the injury crash model showed moderate accuracy (AUC = 0.79). Results indicate that unsafe headways significantly raise crash risk, particularly among drivers showing delayed deceleration or sudden acceleration. Overall, the study highlights that ADAS-based alerts can effectively mitigate crash severity and underscores the importance of adaptive ADAS calibration and driver training to enhance highway safety in India.