Study on Performance Testing and Evaluation of Autonomous Emergency Braking System Based on Self-Constructed...

Area of Science:

• Automotive Engineering
• Road Safety
• Artificial Intelligence in Vehicles

Background:

• Autonomous emergency braking (AEB) systems are crucial for vehicular safety and collision prevention.
• Advancements in assisted driving technology necessitate robust performance evaluation methods for AEB systems.
• Existing evaluation methods may not fully capture real-world driving complexities.

Purpose of the Study:

• To develop and validate a comprehensive performance evaluation framework for AEB systems.
• To establish novel performance indices for assessing AEB system reliability in diverse scenarios.
• To compare the AEB performance of specific vehicle models under controlled conditions.

Main Methods:

• Construction of four typical real-world driving test scenarios for AEB systems.
• Proposal and application of five comprehensive performance evaluation indices: braking parking distance, braking deceleration, collision warning time, speed variation, and accident collision avoidance rate.
• Development and implementation of the Comprehensive Performance Evaluation Index Model (CPEIM) and associated scoring rules for road testing data analysis.

Main Results:

• The Tesla Model Y achieved a score of 1.8857, and the Volvo S90 scored 2.0433 using the CPEIM.
• Both vehicles' AEB systems successfully averted collisions at speeds up to 35 km/h (Tesla Model Y) and 45 km/h (Volvo S90) on dry pavement.
• The proposed evaluation indices and model provided a reliable method for assessing AEB system performance.

Conclusions:

• The developed CPEIM offers an efficient and reliable approach for testing AEB system performance in real-world driving.
• The study highlights the varying capabilities of AEB systems in different vehicle models and speed thresholds.
• Further research can refine the CPEIM for broader application across various driving conditions and vehicle types.