Artificial intelligence automated solution for hazard annotation and eye tracking in a simulated environment.

Area of Science:

• Human-computer interaction
• Artificial intelligence
• Transportation engineering

Background:

• High-fidelity driving simulators generate large datasets requiring manual annotation.
• Manual annotation of roadway elements like hazards is time-consuming and labor-intensive.
• Previous research focused solely on hazard detection in driving simulations.

Purpose of the Study:

• To propose an AI-driven solution for automating data annotation in driving simulations.
• To enhance existing systems by integrating hazard annotation with gaze-tracking data.
• To provide researchers with a streamlined approach for analyzing driving behavior.

Main Methods:

• Utilizing a high-fidelity full-cab driving simulator equipped with gaze-tracking cameras.
• Developing an AI system to automatically annotate roadway elements and integrate with gaze data.
• Combining vehicle handling parameters with driver visual attention data for analysis.

Main Results:

• The proposed AI solution automates the previously manual and time-consuming annotation process.
• The integrated system provides a unified view of driving behavior by combining vehicle dynamics and visual attention.
• Accelerated data analysis and research timelines are achieved.

Conclusions:

• AI-driven automation significantly enhances efficiency in driving simulation research.
• The integration of hazard annotation and gaze tracking offers deeper insights into driver behavior.
• This approach accelerates the understanding and advancement of driving behavior studies.