Work

Harnessing floating car data, traffic camera observations, and network flow analysis for traffic volume estimation

Year: 2026

Type: article

Abstract: Floating Car Data (FCD) have been increasingly adopted by cities and transportation agencies to monitor traffic conditions. Despite the advantages of broad spatial coverage and high temporal resolution, such data provide only a partial representation of urban traffic, with systematic biases across corridors and time periods. In contrast, surveillance cameras can provide high-fidelity traffic information, but only at a limited set of locations, typically sparsely distributed across the road network. This study presents a hybrid modeling and calibration framework that leverages these complementary data sources to produce physically consistent, network-wide estimates and short-horizon forecasts of traffic volumes. The framework incorporates principled traffic dynamics derived from the Cell Transmission Model into the training of a Graph Neural Network, which learns the spatiotemporal evolution of traffic states across the road network. Model predictions are subsequently calibrated against surveillance camera observations using a progressive data assimilation scheme based on the Ensemble Square-Root Kalman Filter. A topology-informed flow-weighted transition matrix is further employed to propagate learned corrections to camera-unobserved locations, enabling network-wide traffic state and volume estimation. The approach is demonstrated using vehicle trajectories and municipal traffic cameras in Manhattan, New York City, where it delivers improvements in accuracy over FCD-based estimates. The proposed framework accommodates varying sensor availability and produces calibrated traffic volumes with uncertainty estimates, supporting operational monitoring and evaluation of transportation policies in data-constrained urban environments.

Source: Discover Civil Engineering

DOI: 10.1007/s44290-026-00657-x

Authors: Antonina Kosikova, Mehmet Kerem Turkcan, Ahmed Darrat, Andrew Smyth

Institutions: Columbia University, Bellevue College

Language: English

Topic: Traffic Prediction and Management Techniques

Subfield: Building and Construction

Field: Engineering

Domain: Physical Sciences

SDGs: Sustainable cities and communities, Industry, innovation and infrastructure

Open Access status: diamond

Funders: National Science Foundation, Nvidia