Education, Science, Technology, Innovation and Life
Open Access
Sign In
  1. Home
  2. Journals
  3. Automation and Machine Learning
  4. Vol 7, Issue 2, 2026

Construction of All-Weather Speed–Spacing Safety Benchmarks for Highways under Complex Weather Conditions Using FT-Transformer

Download as PDF

DOI: 10.23977/autml.2026.070203 | Downloads: 0 | Views: 16

Author(s)

Quan Yu 1, Huihui Guo 1, Jingfeng Zou 1, Feier Jia 1

Affiliation(s)

1 College of Electrical and Control Engineering, North China University of Technology, No.5 Jingyuanzhuang Road, Shijingshan District, Beijing, China

Corresponding Author

Huihui Guo

ABSTRACT

To address the limited adaptability of static speed and spacing regulations under complex weather conditions, this study develops a data-driven framework for constructing all-weather speed–spacing safety benchmarks for highways. A total of 54 composite weather–traffic scenarios were generated through driving simulation experiments by combining weather and road surface conditions, visibility levels, illumination conditions, and background traffic flow. Microscopic driving behavior data were extracted to characterize the probabilistic distribution of vehicle speed and car-following distance under different environmental constraints. A multi-task FT-Transformer model combined with conditional quantile regression was employed to estimate the 15th, 50th, and 85th percentile boundaries of the two-dimensional safety interval. The 85th percentile boundary was further integrated with physical headway constraints and statutory speed–spacing requirements to generate scenario-specific safety benchmarks. The results show that the proposed model achieves competitive prediction accuracy and provides smoother responses under continuous visibility variations compared with tree-based baselines. Benchmark analysis indicates that low-adhesion snowy conditions require greater speed reductions than those implied by visibility-based statutory limits, while extreme low-visibility and high-traffic scenarios require larger car-following distances than the statutory minimum. These findings provide quantitative support for dynamic speed control, spacing management, and cooperative vehicle–infrastructure safety applications under adverse weather conditions.

KEYWORDS

Traffic engineering, all-weather safety benchmark, complex weather, FT-Transformer, conditional quantile regression, speed–spacing control

CITE THIS PAPER

Quan Yu, Huihui Guo, Jingfeng Zou, Feier Jia. Construction of All-Weather Speed–Spacing Safety Benchmarks for Highways under Complex Weather Conditions Using FT-Transformer. Automation and Machine Learning (2026). Vol. 7, No. 2, 16-23. DOI: http://dx.doi.org/10.23977/autml.2026.070203.

REFERENCES

[1] Zhai, X., et al., "Evaluating efficiency and safety of mixed traffic with connected and autonomous vehicles in adverse weather," Sustainability 15(4), 3138 (2023).
[2] Hussain, Q., et al., "Assessing the impact of adverse weather on performance and safety of connected and autonomous vehicles," Civil Engineering Journal 9(12), 3040-3055 (2023).
[3] Martinelli, A., et al., "Explaining the road accident risk: Weather effects and traffic safety," Accident Analysis & Prevention (2023).
[4] Li, Y., et al., "Design and simulation of a variable speed limit system for freeway bottleneck areas," Sustainability 15(1), 162 (2023).
[5] Federal Highway Administration., "Traffic management systems managing the use of variable speed limits (VSLs) during adverse weather conditions," FHWA Reports, August (2024).
[6] Del Serrone, G., et al., "Dynamic signs and variable speed limits to enhance road safety and traffic efficiency," Sustainability 17(11), 4833 (2024).
[7] Lin, P., et al., "Exploring the effects of variable speed limit on the car-following process on bridges in fog with consideration of human factors," Journal of Transportation Engineering, Part A: Systems 151(4), (2025).
[8] Wu, Y., et al., "Variable speed limit control for mixed traffic flow on highways based on deep-reinforcement learning," Journal of Transportation Engineering, Part A: Systems 150(3), (2024).
[9] Wang, J., et al., "Deriving environment-friendly speed limit scheme using an integrated traffic simulation framework," Journal of the Air & Waste Management Association 71(5), 606-621 (2021).
[10] Zhu, M., et al., "IDM-Follower: A model-informed deep learning method for car-following trajectory prediction," IEEE Transactions on Intelligent Vehicles 9(6), 1-6 (2024).
[11] Wang, Z., et al., "Car-following models: A multidisciplinary review," arXiv preprint arXiv:2304.07143 (2024).
[12] Liu, H., et al., "A deep learning model for interpreting car following behaviors," IEEE Transactions on Intelligent Transportation Systems (2024).
[13] Zhang, Y., et al., "A survey of autonomous driving trajectory prediction: Methodologies, challenges, and future prospects," Sensors 13(9), 818 (2024).
[14] Farhadi, A., et al., "VT-Former: An exploratory study on vehicle trajectory prediction for highway surveillance through graph isomorphism and transformer," arXiv preprint arXiv:2311.06623 (2024).

Subscription
E-Mail Alert
Most popular papers
Call for Special Issue
Publication Ethics & OA Statement
Plagiarism Policy
Contact Us
Downloads: 5361
Visits: 273441

Sponsors, Associates, and Links

All published work is licensed under a Creative Commons Attribution 4.0 International License.

Copyright © 2016 - 2031 Clausius Scientific Press Inc. All Rights Reserved.