Large Language Models in Transportation Systems Management and Operations: From Text Reasoning to Multi-modal Decision Support
2026-05-31 • Artificial Intelligence
Artificial Intelligence
AI summaryⓘ
The authors review how large language models (LLMs), including those that handle multiple types of data (MM-LLMs), are used in managing transportation systems. They look at applications in three areas: transportation operations, mobility and fleet services, and data-driven decision support. The paper highlights challenges like handling different kinds of data, making quick predictions, and explaining results. The authors also discuss opportunities for improving localized use, running models on smaller devices, and working across agencies. They find that LLMs mainly help support decision-making, especially when combining text, images, and sensor data.
Large Language Models (LLMs)Multi-Modal Large Language Models (MM-LLMs)Transportation Systems Management and Operations (TSMO)Data HeterogeneityReal-Time InferenceDecision Support SystemsMulti-Modal Data FusionEdge DeploymentPRISMA MethodologyCross-Agency Collaboration
Authors
Siyan Li, Zehao Wang, Jiachen Li, Kanok Boriboonsomsin, Matthew J. Barth, Guoyuan Wu
Abstract
Transportation systems management and operations (TSMO) increasingly depends on timely interpretation of heterogeneous data, from various sensor streams, incident reports, traveler feedback, and visual observations. Large language models (LLMs), including emerging multi-modal large language models (MM-LLMs), provide a new mechanism for integrating these structured and unstructured inputs into operator-facing decision support. This survey paper reviews LLM- and MM-LLM-based applications in TSMO across three domains: transportation operations & services (supply), mobility & fleet services (demand), and data, modeling & decision support. Using a PRISMA-guided screening process, we synthesize current studies while distinguishing operationally oriented applications from prototype and emerging concepts. We further identify recurring challenges in data heterogeneity, real-time inference, explainability, multi-modal fusion, and governance. Finally, we outline existing gaps and future directions in localized adaptation, edge deployment, benchmarking, and cross-agency collaboration. Overall, LLM-based systems appear most promising as a decision-support layer, with MM-LLMs offering particular value when heterogeneous text, visual, and sensor inputs must be integrated.