Estimating the Necessary Amount of Driving Data for Assessing Driving Behavior

Area of Science:

• Transportation Science
• Behavioral Analytics
• Data Science

Background:

• Assessing driver behavior requires adequate data collection.
• Establishing a standardized methodology for data collection is crucial for reliable analysis.

Purpose of the Study:

• To develop a framework for determining the necessary amount of driving data per driver.
• To investigate if driving behavior stabilizes after a specific duration or number of trips.

Main Methods:

• Utilized smartphone sensor data for detailed driving metrics.
• Employed mathematical and statistical methods to analyze data convergence.
• Examined harsh acceleration/braking, mobile usage, and speeding behaviors.

Main Results:

• No universal time point or trip count exists for driving behavior stabilization.
• Trip duration and driver aggressiveness significantly influence behavior convergence.
• Variability in metrics like harsh events and speeding was observed.

Conclusions:

• Individualized data collection strategies are necessary for accurate driving behavior assessment.
• Future research should consider trip context and driver-specific traits.
• The proposed framework aids in optimizing data collection for driver behavior studies.