Projects

I have participated in research projects with industries and public agencies. I mainly support data-driven decision making and develop learning-based operational strategies.

Deer Mobility, Mar 2021 - Aug 2021

Prediction of the Optimal Density of Shared E-scooter System

Research project finding the optimal fleet size of e-scooters in each service region

I contribute

Developed performance metric for evaluating travel regularity of service regions
Developed a daily usage prediction model
Developed indicator for the new entry region selection

Korean National Police Agency, Mar 2021 -Present

Development of Technology for Missing Traffic Information Generation and Cooperative Traffic Control Strategy

Research project generating missing traffic data generation using multi-source data and calibrate simulation parameters using reinforcement learning

I contribute

Conducted seminar on multi-modal deep learning
Selected research site and pre-processed sensor data
Applied graph neural networks to represent urban road topology

Graduate Course Project, Sep 2021 -Dec2021

Superpixel-based Graph Convolutional Network for Semantic Segmentation [pdf]

Research project conducted in "Machine learning for visual understanding" course.
Developed deep learning model for semantic segmentation
Applied graph convolutional networks (GCN) with superpixel penalty loss function and weighted node sampling

Korea Expressway Corporation, Mar 2022 -Present

Strategies for Strengthening the Competitiveness of Highway

Research project comparing competitiveness of highway and railroad
Establishing strategies for enhancing competitiveness of highway

I contribute

Developed indicator for comparing performance of transport by unifying different level of highway/railroad line
Analyzed the change of modal split with the sequential opening of high-speed railroad in Korea