Holistic Vehicle Instrumentation for Assessing Driver Driving Styles

Area of Science:

• Automotive Engineering
• Control Systems
• Human-Machine Interaction

Background:

• The increasing prevalence of autonomous vehicles necessitates seamless integration with human-driven vehicles.
• Understanding and replicating diverse human driving styles is crucial for natural and safe autonomous driving.

Purpose of the Study:

• To design a braking system controller capable of imitating various human braking maneuvers.
• To develop a method for dynamically classifying driver styles based on braking behavior.

Main Methods:

• Experimental tests were conducted on a sensor-instrumented vehicle performing maintained, progressive, and emergency braking maneuvers.
• Braking data, including hydraulic pressure, pedal force, deceleration, and power, were analyzed at different speeds.
• A methodology was developed to characterize braking types and classify driving styles.

Main Results:

• Three distinct braking maneuvers (maintained, progressive, emergency) were identified and characterized.
• The system successfully classified driving styles into cautious, normal, and aggressive based on braking dynamics.
• Vehicle speed, hydraulic brake circuit pressure, brake pedal force, longitudinal deceleration, and braking power were key indicators.

Conclusions:

• The study presents a novel approach to characterizing braking types and classifying driver styles using on-board sensor data.
• This classification enables autonomous vehicles to dynamically adapt to human driving behaviors.
• The findings contribute to developing more natural and human-like autonomous driving systems.