Overview
This research develops sophisticated computational frameworks for analyzing multi-source traffic data in urban transportation networks, addressing the common case where incomplete data collection and anomalous traffic events co-occur — requiring analytical approaches that can handle both simultaneously.
Technical Approach
- Tensor-based representations capturing multi-dimensional relationships in traffic data across space, time, and measurement source.
- Large-scale optimization algorithms with convergence guarantees.
- Data-fusion techniques integrating volume sensors, occupancy detectors, and speed measurements.
- Pattern recognition to identify normal traffic patterns and flag anomalous events.
Expected Outcomes
Robust computational tools for traffic-data analysis that handle real-world data imperfections, improved incident detection and response, and practical frameworks for transportation authorities and smart-city initiatives.
