Jiangkun, L.; Ruixue, Z.; Ying, W.; Weiwen, D. Complexity Evaluation for Urban Intersection Scenarios in Autonomous Driving Tests: Method and Validation. Preprints2024, 2024100602. https://doi.org/10.20944/preprints202410.0602.v1
APA Style
Jiangkun, L., Ruixue, Z., Ying, W., & Weiwen, D. (2024). Complexity Evaluation for Urban Intersection Scenarios in Autonomous Driving Tests: Method and Validation. Preprints. https://doi.org/10.20944/preprints202410.0602.v1
Chicago/Turabian Style
Jiangkun, L., Wang Ying and Deng Weiwen. 2024 "Complexity Evaluation for Urban Intersection Scenarios in Autonomous Driving Tests: Method and Validation" Preprints. https://doi.org/10.20944/preprints202410.0602.v1
Abstract
Complexity is a crucial metric for evaluating urban intersection scenarios. However, existing complexity evaluations are often derived indirectly from the driving risk of autonomous vehicles, leading to subjective and incomplete evaluations. Therefore, this study undertakes a comprehensive evaluation of intersection scenarios from a systems perspective by quantifying the number of vehicles and the characteristics of vehicle interactions within the intersection.Specifically, we propose an objective framework for assessing complexity. In this framework, the number of vehicles serves as the foundation for assessment, while three interaction characteristics—interaction density, interaction disorder, and interaction risk—are modeled as scaling factors of complexity. In particular, by combining logistic and hyperbolic tangent functions, we accurately capture the nonlinear effects of vehicle numbers and interaction characteristics on complexity.The experimental results demonstrate that the evaluation outcomes of the proposed method are highly correlated with the risk and efficiency performance of autonomous vehicles at intersections, thereby validating the method. Additionally, this method is utilized to guide the generation of highly complex urban intersection scenarios, achieving double the generation efficiency compared to the random sampling method.
Copyright:
This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.