Personalized Shared Control for Automated Vehicles Considering Driving Capability and Styles.

Area of Science:

• Automotive Engineering
• Human-Computer Interaction
• Control Systems

Background:

• Shared control systems aim to enhance automated vehicle safety and comfort by integrating human and automated driving capabilities.
• Understanding and adapting to individual human driver characteristics are crucial for the acceptance and effectiveness of shared control systems.
• Current systems often lack personalization, limiting their ability to fully leverage human driving expertise.

Purpose of the Study:

• To propose and evaluate a personalized shared control framework that considers individual drivers' capabilities and styles.
• To develop methods for simulating human factors and evaluating driving characteristics for personalized control.
• To assess the impact of personalized shared control on driving safety, comfort, and workload.

Main Methods:

• Developed a simulated scenario generation method incorporating human factors.
• Defined and evaluated drivers' driving capabilities for improved driving authority allocation.
• Analyzed and characterized drivers' driving styles through field tests for intention-aware subsystems.
• Proposed a personalized shared control framework based on driving capabilities and styles.
• Evaluated the system using human-in-the-loop simulations and field tests on an automated vehicle.

Main Results:

• The personalized shared control system demonstrated improved performance across diverse driving capabilities and styles compared to non-personalized systems.
• The framework effectively integrated human factors, enhancing the rationality of driving authority allocation.
• Evaluation criteria including safety, comfort, and workload showed significant benefits with the personalized approach.
• The system proved effective in complex scenarios, outperforming purely human-driven or automated systems.

Conclusions:

• Personalized shared control, considering individual driver factors, significantly enhances automated vehicle performance.
• The proposed framework offers a viable approach for developing adaptive and human-centered automated driving systems.
• Future research should focus on further refining personalization algorithms and expanding real-world testing.