Long-Term Evaluation of Drivers' Behavioral Adaptation to an Adaptive Collision Avoidance System

Area of Science:

• Human-Computer Interaction
• Automotive Engineering
• Cognitive Psychology

Background:

• Adaptive automation systems adjust automation levels based on situational demands and user capacity, proving effective in dynamic environments.
• However, drivers struggle to adapt to dynamically changing automation configurations, highlighting a need for better user support.
• Existing research has limited focus on enhancing driver comprehension and behavioral adaptation to adaptive automation.

Purpose of the Study:

• To investigate the impact of human factors and user interaction on the performance and safety of driver support systems.
• To explore methods for improving driver understanding and acceptance of adaptive automation.
• To enhance the long-term effectiveness of human-automation interaction in driving.

Main Methods:

• A within-subjects design was employed with 42 participants in a 1-month driving simulation experiment.
• Participants gradually interacted with an adaptive collision avoidance system under hazardous lane-change scenarios.
• The experiment involved four stages to assess the effects of system interaction and training.

Main Results:

• Initial use of the adaptive collision avoidance system significantly reduced collisions compared to unsupported driving.
• Increased driver-system interaction and dedicated training led to significant improvements in driver trust and system understanding.
• Driving behavior and system adaptation improved over the course of the experiment with sustained interaction and training.

Conclusions:

• Designing driver support systems with human factors in mind is essential but insufficient on its own.
• Ensuring appropriate user understanding and acceptance is critical for maximizing system safety and usability.
• Findings underscore the importance of user-centered design and training for advanced driver-assistance systems and automated driving.