Autonomous driving systems need to align with human risk perception in near-miss and safety-critical scenarios to improve passenger trust.

Naturalistic driving/crash dashcam videos, passenger survey responses, and risk ratings.

Surveys, perceived risk modeling, regression models, and statistical comparison between AV and human-driven vehicle scenarios.

Completed PI project funded by National Research Foundation of Korea (73.5M KRW), yielding publication manuscripts (under 1st revision at IEEE T-ITS).

Proactive risk assessment frameworks are needed to evaluate autonomous vehicle safety in real time without relying solely on historical crash records.

Microscopic vehicle trajectories, driving simulator datasets, and sensor-based telemetry.

Risk field theory, surrogate safety measures (SSMs), reliability analysis, and probabilistic crash risk estimation.

Completed PI project funded by National Research Foundation of Korea (96.1M KRW), culminating in doctoral thesis and publications in AMAR and AAP.

State transportation agencies need automated ways to process high-resolution signal event records to evaluate corridors and identify safety opportunities.

High-resolution traffic signal controller event logs, detector logs, and signal plans.

Large-scale data processing pipelines, performance measure computations (ATSPM), and dashboard-facing safety analytics.

Completed FDOT-sponsored research project (2024–2025) and agency-facing analytical dashboard.

Integrating and managing diverse traffic safety facility data in real time is critical to support cooperative traffic control strategies and connected vehicle deployments.

Traffic safety facility status logs, real-time control metrics, and detector feeds.

System architecture design, database management, and information integration frameworks.

Completed research deliverables, guidelines, and technical reports delivered to the Korean National Police Agency.

Establishing cooperative traffic control frameworks for connected environments to safely manage mixed traffic of human-driven and automated vehicles.

Signal timing plans, simulator-based traffic logs, and cooperative vehicle telemetry.

Multi-agent coordination, signal optimization models, and microscopic traffic simulation.

Completed research deliverables, system models, and guidelines for cooperative traffic control.

High pedestrian crash rates at urban crosswalks necessitate proactive monitoring and warning systems to detect near-miss conflicts before crashes occur.

Video monitoring streams, pedestrian tracking logs, and vehicle approach speeds.

Video-based pedestrian detection, trajectory tracking, behavior pattern analysis, and conflict risk estimation.

Completed research project delivering proactive warning algorithms and pedestrian risk assessment guidelines.

Traditional crash hotspot identification methods are reactive, rely on sparse crash histories, and suffer from regression-to-the-mean biases.

Historical crash records, regional roadway geometry datasets, and simulated collision frequencies.

Ensemble Empirical Mode Decomposition (EEMD), spatial statistics, and traffic collision frequency simulation.

Completed research project yielding non-stationary spatial analysis algorithms for hotspot identification.

Urban transportation networks lack proactive indicators that forecast traffic states and systems performance in real time using sparse connected vehicle telemetry.

High-resolution connected vehicle trajectories, speed grids, and weather data.

Spatiotemporal transformers (e.g., MMCAformer), cross-attention fusion, and deep learning for traffic state forecasting.

Active research project yielding manuscripts under review at IEEE T-ITS and TRB.

Traditional choice models fail to capture complex non-linear behaviors, while black-box deep learning models lack the interpretability needed for transport policy decisions.

Household travel surveys, demographic datasets, and regional public transit statistics.

Deep neural networks, explainable AI techniques (SHAP, LIME), and activity-based travel choice modeling.

Developed a data-driven travel demand model with feature attribution diagnostics.

Generating fine-grained, lane-level traffic flow and speed information where physical traffic detectors are sparse or missing.

Standard surveillance video feeds and raw camera streams.

Computer vision, lane-detection models, and deep generative algorithms.

Completed research project yielding lane-level traffic flow generation algorithms and technical papers.

Inductive loop detectors provide limited spatial coverage, making it difficult to monitor and manage highway bottleneck congestion dynamically.

UAV aerial video, loop detector occupancy logs, and highway traffic speeds.

UAV video vehicle tracking, multi-source sensor integration, and congestion mitigation strategy design.

Completed project delivering sensor integration frameworks and active traffic management algorithms.

Evaluating the social benefits and projecting transit ridership for extending municipal rail links into transit-deficient campus zones.

Smart card transit transaction logs, commute surveys, and geographic GIS layers.

Four-step transportation demand forecasting, modal split estimation, and cost-benefit analysis.

Completed feasibility reports and demand projections delivered to transit planning groups.

Modeling macro-level freight and passenger railway demand under potential geopolitical and logistics integration.

International trade flow datasets, historical rail capacity logs, and geopolitical benchmark profiles.

Logistics network assignment, cargo demand forecasting, and bottleneck identification.

Completed strategic rail network planning roadmap and investment priority blueprints.

Formulating a national-level framework and standardized metrics to systematically assess and improve pedestrian safety nationwide.

Nationwide pedestrian crash records, pedestrian facility inventories, and safety perception surveys.

Spatial safety risk modeling, policy benchmarking, Delphi surveys, and multi-criteria decision analysis.

Delivered framework guidelines and official indicators for Korea's first national pedestrian safety plan.

Conducting early-stage economic and engineering feasibility analyses for major transit lines to optimize national capital allocation.

National census datasets, regional travel logs, and GIS infrastructure maps.

Four-step planning demand model, utility-based mode choice, and economic feasibility evaluation (B/C ratio).

Delivered preliminary feasibility study report to the Ministry of Land, Infrastructure and Transport.