This website is no longer updated (Aug. 2023). Please visit my new personal page: https://choi-seongjin.github.io/.
I’m an (upcoming) Assitant Professor in the Department of Civil, Environmental, and Geo- Engineering at the University of Minnesota, starting in January 2024. My research interests are broad and interdisciplinary, encompassing Urban Mobility Data Analytics, Spatiotemporal Data Modeling, Deep Learning & Artificial Intelligence, and Connected Automated Vehicles (CAV) & Cooperative-ITS. I am particularly driven by the desire to optimize urban mobility and contribute to the development of sustainable and efficient urban transportation system. My work involves utilizing data analytics to draw valuable insights from urban mobility data and applying cutting-edge AI technologies in the field of transportation.
Currently, I’m a Postdoctoral Researcher in the Department of Civil Engineering at McGill University in Canada. I’m with Professor Lijun Sun’s group and I’m funded by “Bridging Data-Driven and Behavioral Models for Transportation” project from Institute of Data Valorization (IVADO).
My research topics and interests are as follows:
- Urban mobility data analytics
- Generative AI for transportation and mobility data
- Application of machine learning and deep learning methods in transportation domain
- Cooperative-ITS and Connected and Automated Vehicles
For Prospective Students/Postdocs
I am looking for 1-2 PhD students (and/or a Postdoc) for 2024 Spring/Fall (starts in January/September) who are excited about machine learning for urban transportation and mobility data. If you’re interested, please send me an email. Please use “Prospective PhD student [Your name]” as your email subject. Due to the large volume of emails, I apologize for not replying to individual inquries I have received.
Current Researches Subjects
Characterizing the distribution of prediction error and correlated residuals in spatiotemporal forecasting
Multi-Agent Reinforcement Learning for Cooperative Adaptive Cruise Control
Vehicle-Pedestrian Interaction Prediction through Probabilistic Trajectory Prediction
Previous Research Subjects
Development of Multi-level Traffic Simulation for Performance Evaluation of Cooperative Autonomous Public Transport System
Transformer-based Map Matching model with transfer learning approach
For more information : TMM_Arxiv
TrajGAIL : a generative adversarial imitation learning framework for urban vehicle trajectory generation
For more information : TrajGAIL_TRpartC
For implementations : TrajGAIL_Github
I’m working on developing a microscopic version of TrajGAIL (I call it TrajGAIL-micro). I’m trying to predict the microscopic motion of a vehicle based on the previous coordinates of this vehicle. This model can be used for predicted intersection collision warning system.
Cooperative-ITS Transportation Management Center in Sejong City
- Efficient data sampling method for safety performance indicator
- Advisory Speed Algorithm for Connected and Automated Public Transit (maybe in real-time)
- Routing Algorithm for Demand-responsive Transit