Freeway traffic incident reconstruction – A bi-parameter approach

• We parametrize incidents for traffic flow model.
• We characterize abnormality in traffic conditions by two incident parameters.
• We propose a method for state and incident parameters reconstruction.
• We evaluate the proposed model with incident corrupted real traffic data.

The paper suggests a novel alternative to generalized traffic incident descriptions within the macroscopic traffic model framework. The contribution of the paper is twofold. First, by extending already existing second order macroscopic conservation laws to characterize off-nominal traffic conditions, we define two main incident parameters such as direct and indirect ones. Physical interpretations of this incident parametrization is provided. These incident indicators are relative in view of the nominal traffic flow model parameters and carries physically meaningful macroscopic content. Second, the paper proposes to use a constrained and nonlinear, joint traffic state- and incident parameter reconstruction method and validates the suggested modeling idea via real traffic measurements fitting. Evaluation of the numerical results demonstrate the effectiveness of the methodology.