Unveiling driver workload dynamics and road safety risks in assisted driving systems

Area of Science:

• Human-computer interaction
• Automotive engineering
• Cognitive psychology

Background:

• Automated driving systems are increasingly integrated into vehicles, aiming to enhance safety and reduce driver workload.
• The impact of these systems on driver behavior, particularly during mode transitions, requires further investigation.
• Understanding driver adaptation to and from automated control is crucial for safe system deployment.

Purpose of the Study:

• To investigate the effects of manual, Level 2 assisted, and resumed manual driving modes on lateral vehicle control and visual behavior during curve negotiation.
• To analyze the transition dynamics between manual and assisted driving states.
• To identify distinct driver behavioral states using multivariate analysis.

Main Methods:

• A simulator-based experiment involving sequential driving in manual, Level 2 assisted, and resumed manual modes.
• Analysis of lateral trajectory (Lane Gap) and visual engagement indices during curve negotiation.
• Application of linear mixed-effects models and fuzzy c-means clustering to identify behavioral patterns and transitions.

Main Results:

• Driving mode significantly affected both vehicle trajectory and visual engagement.
• Level 2 assisted driving led to reduced lateral variability and decreased anticipatory and guidance-related visual attention.
• Post-assisted driving manual control showed partial but incomplete recovery to baseline manual behavior.

Conclusions:

• Automated driving systems alter driver behavior, impacting lateral control and visual attention.
• Transitions into and out of assisted driving result in persistent behavioral changes.
• Findings inform the design of human-machine interfaces and road safety strategies for automated vehicles.