Domain-specific large models for transportation
Developing transportation-oriented large models for traffic perception, understanding, prediction, and reasoning in complex expressway scenarios.
Academic homepage · Transportation research
Jingyi Mao
Ph.D. student in Intelligent Transportation
I focus on developing domain-specific large models for expressway traffic, integrating multimodal data such as images, traffic flow, and structured information to improve traffic perception, understanding, and prediction.
School of Civil Engineering and Transportation
South China University of Technology
Ph.D. Supervisor: Prof. Peiqun Lin
Division of Transport Planning
Department of Civil and Architectural Engineering
KTH Royal Institute of Technology
Visiting Ph.D. Student · Host Supervisor: Assoc. Prof. Zhenliang Ma
About
Multimodal Traffic Large Language Models for Intelligent Expressways
I am a Ph.D. student in Civil and Hydraulic Engineering, specializing in transportation, at South China University of Technology. I am also supported as a joint Ph.D. student through the Overseas Study Program of the Guangzhou Elite Project.
My research focuses on multimodal domain-specific large models for intelligent expressways. I study how visual, traffic flow, ETC, and structured traffic data can be jointly modeled to improve traffic perception, understanding, prediction, and decision-making in complex expressway environments.
Research focus
Multimodal traffic intelligence for smarter expressways.
My research lies at the intersection of transportation engineering, multimodal learning, and domain-specific large models.
Multimodal traffic data fusion
Integrating visual, traffic flow, ETC, temporal, and structured traffic data to build unified representations of expressway traffic conditions.
Intelligent expressway perception and understanding
Building reliable methods for traffic state perception, congestion recognition, cause understanding, and future traffic forecasting under complex real-world conditions.
Selected work
Publications
-
2026
TRB 105th Annual Meeting
Low-Light Adaptive Image Recognition for Expressway Congestion using Feature Refinement
J. Y. Mao, J. H. Peng, T. Zhou, Z. H. Wang, T. C. Luo, R. J. Xu, P. Q. Lin*
-
2025
The Journal of Supercomputing
RSW-YOLOv8: An Improved Model for In-vehicle Vision Multi-object Detection in Complex Traffic Scenes
P. Q. Lin*, Q. T. Li, J. Y. Mao
DOI 10.1007/s11227-025-07540-z -
2024
Computers, Materials & Continua
Attention-enhanced Voice Portrait Model using Generative Adversarial Network
J. Y. Mao, Y. C. Zhou, Y. F. Wang, J. Y. Li, F. L. Bu*
DOI 10.32604/cmc.2024.048703 -
2024
Scientific Reports
Co-embedding of edges and nodes with deep graph convolutional neural networks
Y. C. Zhou, Z. W. Hou, J. Y. Mao, F. L. Bu*, et al.
DOI 10.1038/s41598-023-44224-1
Background
Education & distinctions
Education
-
2024—Present
Ph.D. in Civil and Hydraulic Engineering
South China University of Technology · Transportation focus
-
2021—2024
M.Eng. in Electronic Information
People’s Public Security University of China
-
2017—2021
B.Eng. in Traffic Management Engineering
People’s Public Security University of China
Recent awards
-
2025
National Second Place, Transportation Foundation Model Agent Application Innovation Competition
-
2025
Provincial Third Place, Guangdong “AI + Transportation” Innovation Application Competition
-
2024
Third Place, China Graduate Contest on Smart-city Technology and Creative Design
-
2020
National Second Prize and Provincial First Prize, Chinese Mathematics Competitions
Contact
Let’s connect around transportation intelligence.
For research conversations and collaboration opportunities: