Driver Characteristics Oriented Autonomous Longitudinal Driving System in Car-Following Situation

Area of Science:

• Automotive Engineering
• Human-Computer Interaction
• Control Systems

Background:

• Advanced driver assistance systems (ADAS) aim to reduce driving burden and enhance comfort.
• Current ADAS often lack personalization, leading to driver discomfort (heterogeneity).
• A need exists for autonomous systems that align with individual driving intentions.

Purpose of the Study:

• To propose a personalized longitudinal driving system for car-following scenarios.
• To develop a system that mimics individual driving behavior for improved driver acceptance.
• To ensure harmony between automated control and driver intent.

Main Methods:

• A multi-layer framework comprising a speed planner and a driver parameter manager was developed.
• The speed planner uses a parametric cost function and constraints reflecting driver characteristics.
• Driver parameters are identified from real driving data via the driver parameter manager.

Main Results:

• The proposed system successfully mimics the driving style of actual drivers.
• Driving simulation validated the algorithm's ability to personalize longitudinal control.
• The system maintained safety by preventing collisions with preceding vehicles.

Conclusions:

• Personalized speed planning algorithms can effectively address heterogeneity in ADAS.
• Mimicking individual driving behavior enhances driver comfort and system acceptance.
• The developed system offers a promising approach for personalized autonomous driving.