Exploring the Potential for Incorporating Artificial Intelligence in Highway Resilience to Climate Change

Khalid Mohamed,Jia-Dzwan (Jerry) Shen

DOI: https://doi.org/10.1177/03611981241253610

IF: 1.7

2024-06-02

Transportation Research Record

Abstract:Transportation Research Record, Ahead of Print. Highway infrastructure needs to be able to withstand, recover quickly, and adapt to the extreme conditions and loads associated with climate change and other natural hazards. Investigation of new analysis methods of the different types of data for improving transportation resilience is critical. The highway owners are challenged with managing the highway geotechnical features and structures to achieve the adaptive capability. The optimization of the investment requires quantification of the changing hazards and correlation to infrastructure performance. The use of new and broader data sources and artificial intelligence (AI) and machine learning (ML) opens a door to more efficient development of necessary models with many potentially relevant variables and a more adaptive updating process. Climate tools developed by the Federal Highway Administration could be integrated with or enhanced using AI/ML modules trained using properly selected data. The improvement of infrastructure resilience also depends on adequate evaluation of the probability of various damage levels from extreme events. The application of AI/ML on a broader data source will enable data-driven probabilistic damage modeling approaches that are often not feasible using traditional regression processes. This paper discusses from a user's point of view several potential applications that could take advantage of AI/ML techniques and big data availability, including geotechnical asset management, geohazard programs, and probabilistic damage evaluation. Challenges in current practice are analyzed and connected to the need for a new approach. The anticipated applicable areas are offered to the AI/ML professionals to consider for further studies.