Testing Methods for Advanced Intelligent Driving Systems Based on Natural Driving Behavior Characteristics

The technology of advanced intelligent driving systems equipped in intelligent and connected vehicles evolves rapidly, while the corresponding testing and evaluation standards lag behind. To propose new testing methods for advanced intelligent driving systems, this study first identifies typical intelligent driving functions through extensive research. Based on the distribution characteristics of natural driving scenarios, a series of test cases for these typical functions are developed. By employing simulation methods and the "logical scenario-critical scenario-experimental scenario" mechanism, the generalized test cases are filtered, enabling the rapid extraction of high-value test cases for real-world field testing and validation. Finally, the typical intelligent driving functions of advanced driving systems are analyzed and evaluated from the perspectives of perception and decision-making, as well as control operations. The experimental results demonstrate that the proposed testing method, grounded in natural driving behavior characteristics, improves testing efficiency and enables the evaluation of the performance of advanced intelligent driving systems. This provides a theoretical foundation for establishing evaluation standards for these systems.