Section-wise modeling of traffic flow and its application in traffic state estimation

The section-wise macroscopic modeling of traffic flow is considered. A section is supposed to be bounded by detectors and several kilometers long. To compute a section model without input from the adjacent sections, feasible boundary conditions have to be specified and boundary variables have to be approximated by the detector measurements. These approximations lead to unknown disturbances in the section model, depending on the direction of propagation of information in the traffic flow. Various approximations of the boundary variables are described and a disturbance modeling approach is proposed. The application of section models in traffic state estimation and the effects on the estimation error are discussed. Simulation tests are showing the performance of the estimation, based on the various boundary approximations.