Data-driven uncertainty quantification in macroscopic traffic flow models

We propose a Bayesian approach for parameter uncertainty quantification in macroscopic traffic flow models from cross-sectional data. consider both simple first order model consisting the mass conservation equation and its second version including speed evolution equation. A bias term is introduced modeled as Gaussian process to account limitations. validate results comparing error variables (flow, speed, density) models, showing that globally perform better reconstructing quantities of interest.