Modeling Motorcyclists' Aggressive Driving Behavior Using Computational and Statistical Analysis of Real-Time Driving...

Area of Science:

• Road safety research
• Transportation engineering
• Human-computer interaction

Background:

• Aggressive driving is a primary cause of traffic accidents, leading to significant loss of life and property.
• Detecting dangerous driving behaviors is crucial for improving road safety.
• Utilizing driving data from mobile devices offers a novel approach to assessing driving styles.

Purpose of the Study:

• To detect dangerous and aggressive driving profiles among motorcyclists.
• To differentiate between non-aggressive and aggressive driving behaviors using collected data.
• To identify key characteristics indicative of aggressive, normal, or dangerous motorcycle riding.

Main Methods:

• A two-stage method was employed, first establishing driver profile thresholds (rules).
• Data were collected from motorcyclists using the Android application -Speedometer GPS-, including spatiotemporal information.
• Feature extraction and classification model development were performed on preprocessed data.

Main Results:

• Key characteristics were identified for categorizing driving behavior as aggressive, normal, or dangerous.
• The study successfully differentiated between non-aggressive and aggressive driving patterns.
• The findings revealed significant safety issues associated with motorcycle riding behavior.

Conclusions:

• The developed method effectively categorizes motorcycle driving behavior.
• The identified characteristics provide valuable insights for improving road safety.
• This research contributes to reducing motorcycle accidents through better understanding of driving behavior.