Citation

Liao X, Wu G, Yang L, Barth MJ. Transp. Res. D Trans. Environ. 2023; 117: e103664.

Copyright

(Copyright © 2023, Elsevier Publishing)

DOI

10.1016/j.trd.2023.103664

PMID

unavailable

Abstract

Timely and reliable accident detection provides a foundation for traffic accident management (TIM), which is critical functionality for traffic management agencies. Effective TIM strategies mitigate negative impacts caused by non-recurrent events, improve quality of service and traveler satisfaction, and enhance transportation resilience. Most existing studies focus on traffic accident detection and system mobility. Very few systems attempted to quantify the environmental impacts of accidents. We examine a cloud-based data platform that fuses information from real-world traffic, probe vehicle data, and road weather. Moreover, we developed a data-driven approach to estimate the impacts of accidents using Otsu's method, morphological operation, time-series prediction, and emissions simulator model, allowing us to quantify the benefits of advanced accident detection. The proposed method was evaluated with a real-world scenario, showing that the studied accident may cause additional energy waste by 38% and CO emissions by 36%.

Language: en

Keywords

Accident impact; Energy and emissions; Machine learning; Morphological processing; State prediction