G. Pai, A.K. Pradhan

Pages: 31-42

Abstract
Advanced Driver Assistance Systems (ADAS) aim to improve roadway safety by aiding drivers in various vehicle control tasks. However, the complexity of these systems may result in drivers’ poor or inaccurate mental models about these systems, which can adversely affect drivers’ understanding and use of the systems, potentially resulting in poor hazard avoidance behaviors. There are important gaps in the literature about the role of hazard avoidance in the ADAS context. Unlike traditional driving where hazards are almost entirely external, ADAS introduces hazards related to knowledge of system state and its changes. This study was conducted to understand drivers' hazard avoidance behaviors when driving with Adaptive Cruise Control (ACC). The study involved ten experienced ACC users, aged 25 to 55, divided equally into either the Eye Movements or Verbal Response groups. Each participant completed a simulator drive featuring six ‘edge-case’ and two ‘normal’ events. Eye movement observations suggest that drivers exhibited an increased tendency to scan the instrument panel when encountering different events. Verbal responses indicate that drivers possessed similar awareness about system status and in-vehicle changes, and external factors. These observations suggest that hazard avoidance behaviors may differ for ADAS when compared to traditional driving. Future work could use these insights to design better user interfaces for real-time assistance and to develop training and educational methods to improve hazard avoidance and mental models.
Keywords: Advanced Driver Assistance Systems; automation; roadway safety; eye-tracking; hazard avoidance; human factors