Quantifying perceived risk in driving: A Monte Carlo approach for obstacle avoidance.

Area of Science:

• Traffic Safety
• Human Factors in Driving
• Risk Perception Modeling

Background:

• Limited research models driving risk from the driver's perspective.
• Understanding driver risk perception is crucial for enhancing road safety.

Purpose of the Study:

• Develop a quantitative model for perceived risk in obstacle avoidance.
• Investigate how drivers' risk perceptions influence their driving decisions.
• Incorporate fuzzy risk perceptions into a probabilistic driving model.

Main Methods:

• Utilized Monte Carlo methods to model driving uncertainties.
• Quantified perceived risk using drivers' fuzzy perceptions.
• Employed a logit model to link perceived risk with driving decisions.

Main Results:

• Experimental data showed variations in vehicle trajectories based on driver experience.
• Perceived Risk Indicator (PRI) values were higher for leftward bypasses.
• The PRI demonstrated strong predictive ability (AUC=0.820); logit model achieved 90% accuracy in predicting rightward bypasses.

Conclusions:

• The study provides a framework for understanding driver risk perception and decision-making.
• Findings aid traffic professionals in improving traffic safety.
• Highlights the importance of driver risk perception in driving behavior models.