The AI Frontier in Transportation Engineering: Multi-Agent Systems and the Future of Autonomous Reasoning
MIE Department Seminar
March 10, 2026
11:00 AM - 12:00 PM America/Chicago
Presenter: Shan Bao, PhD, University of Michigan-Dearborn
Location: ERF 1043
Abstract: Traffic safety is entering a transformative era where the analysis of pre-crash behavior is no longer constrained by the limitations of traditional reconstruction. While human-led analysis has long been the gold standard, it is frequently hampered by subjective interpretation, data quality issues, and the sheer complexity of fragmented, multimodal data. This seminar explores how Artificial Intelligence (AI) is revolutionizing Transportation Engineering by bridging the gap between historical crash data and the future of Automated Vehicle (AV) decision-making.
Bao will share her vision for a new paradigm in crash causation reasoning, centered on a novel two-phase, multi-agent collaborative framework. This system leverages a hierarchy of AI agents to reconstruct pre-crash scenarios with unprecedented precision, outperforming traditional methods in identifying evasive maneuvers and Driver Hazardous Actions (DHA). By integrating Large Language Models (LLMs) to parse unstructured narratives and probabilistic reasoning to resolve evidence discrepancies, her research demonstrates how AI can achieve superior accuracy even when faced with incomplete or ambiguous data.
The implications of this work extend far beyond forensic analysis. As drivers transition from active operators to supervisory controllers, the ability of an AV to "understand" human error becomes the cornerstone of safety. Bao’s research provides the foundational logic for Human-AI Teaming, ensuring that future autonomous systems can navigate complex road-user interactions with the cognitive depth and real-time awareness necessary for global deployment.
Speaker Bio: Professor Shan Bao is a human factors researcher with 20 years of experience in conducting both theoretical and applied research. She holds the position of professor and department chair at the University of Michigan-Dearborn’s Department of Industrial and Manufacturing Systems Engineering, as well as that of research professor at the University of Michigan Transportation Research Institute (UMTRI) at Ann Arbor. Bao has led and conducted multiple large, simulator, and naturalistic driving studies for industry and government sponsors. Her expertise spans a wide array of critical areas, including Driving Automation Systems (DAS) evaluation, intricate statistical analysis of crash and naturalistic data, and safety of vulnerable road users. Bao has led or participated as a PI or Co-PI in a total of 64 funded grants and contracts reported for a total of over $17,000,000. She is the PI leading eight premier institutions on the NHTSA Human Factors Consortium team. She has published a total of 64 journal papers, 20 conference proceedings, and 2 patents. She has given multiple keynote speeches and served on expert panels at different conferences and meetings. Bao is an associate editor of the IEEE-Intelligent Transportation System and serves on the journal Accident Analysis and Prevention editorial board. She is also a member of the Human Factors Committee of the Transportation Research Board, National Academies of Sciences, Engineering, and Medicine.
Date posted
Feb 10, 2026
Date updated
Feb 10, 2026