The FHWA Office of Safety and Operations Research and Development plans to recruit a postdoctoral research associate to conduct research on advanced Artificial Intelligence (AI) and Machine Learning (ML) applications in transportation safety and operations.
While significant progress has been made in improving traffic safety and operations, recent advancements in communications, connected vehicle technologies, edge computing, and data analytics have dramatically expanded the availability of real-time traffic data to transportation engineers and operators. In addition, AI approaches, such as recurrent neural networks (RNNs), reinforcement learning, and various generative AI models, present unprecedented opportunities to transform traffic management and control by enabling real-time, data-driven decision-making that enhances system-wide safety and operations. This includes, but is not limited to, enhancements in road traffic prediction, adaptive signal control, and traffic speed limit control [1][2][3].
However, this big data surge combined with AI/ML approaches has outpaced the current capabilities of some Traffic Management Centers (TMCs) to process, interpret, and implement proactive traffic control strategies at system scale [3]. Many applications remain in their early stages, mainly using AI/ML as a Blackbox in a local infrastructure without deeply integrating domain knowledge. While general AI models exhibit potential, they are not optimized for traffic data analysis and applications. Customizing AI models with domain expertise can maximize their effectiveness in transportation applications. To unlock the full potential of AI/ML in traffic safety and operations, future advancements must integrate AI-driven optimization, real-time control mechanisms, and transportation domain expertise to develop innovative, adaptive, and scalable solutions [3].
A research associate is sought to conduct research to significantly contribute to one of the following priority objectives:
1. Develop AI-powered methods for network-wide traffic data acquisition to ensure sufficient spatiotemporal coverage for extracting actionable insights for system safety and operation applications.
2. Leverage AI technologies to improve traffic pattern prediction and adaptive traffic operations under both normal and abnormal conditions.
3. Develop transportation-specific AI models, and/or enhance existing AI models with transportation domain knowledge to improve their applicability to transportation challenges.
4. Develop AI-enhanced Cooperative Driving Automation algorithms. Given connected/or automated vehicles face complicated and real-time driving decision-making demands under dynamic environments, traditional optimization and control approaches face limitations. AI, such as reinforcement learning, offers more efficient approaches and alternatives for dynamic and adaptive driving control solutions.
The successful applicant will demonstrate a strong background and knowledge in AI models, traffic flow theories, control theories, and optimization modeling and solution algorithm design. The selected applicant will work in person at the Turner-Fairbank Highway Research Center (TFHRC), in McLean, Virginia and to collaborate with the onsite federal and contractor research staff.
Keywords: AI Applications in Traffic Operations, AI, network-wide traffic prediction and management, Cooperative Driving Automation
Reference:
[1]. Hanyi Yang, Du, Lili, Gaohu Zhang, Tianwei Ma (2023). A Traffic Flow Dependency and Dynamics based Deep Learning Aided Approach for Network-Wide Traffic Speed Propagation Prediction. Transportation Research Part B: Methodological, 167, 99-117.
[2]. Ouallane, Asma Ait, Ayoub Bahnasse, Assia Bakali, and Mohamed Talea. "Overview of road traffic management solutions based on IoT and AI." Procedia Computer Science 198 (2022): 518-523.
[3]. Liu, C., Pu, C., Du, L., & Wang, Y. (2024). Potentials and Challenges of AI-Empowered Solutions to Urban Transportation Infrastructure Systems: NSF AI-Transportation Workshop Phase I. Journal of Transportation Engineering, Part A: Systems, 150(9), 02524001.
AI Applications in Traffic Operations, AI, network-wide traffic prediction and management, Cooperative Driving Automation