A Method for Identifying Characteristic Parameters of Expressway Traffic Flow and Conducting Macroscopic Modeling by Integrating Multi-source Heterogeneous Data

To accurately identify the characteristic parameters of expressway traffic flow, such as free flow velocity, congestion density, critical speed, and traffic capacity, and to address the estimation accuracy issues caused by a single data source and a single model as the optimization objective in basic graph modeling, thereby enhancing the accuracy of macroscopic traffic flow modeling. The multi-source data fused with millimeter-wave radar and binocular cameras is adopted. The density-velocity relationship is characterized based on the Van Aerde single-stage model, and the weighted least squares objective function of the fused density and flow error is constructed. The nonlinear least squares method is used to estimate the model parameters. Taking a certain lane of the G4 Beijing-Hong Kong-Macao Expressway in Hunan Province as the research object, traffic data for six days from September 12th to 17th, 2025 were collected. After screening and processing, valid samples were obtained. The final identified feature parameters The free flow velocity was 86.14 km/h, the critical velocity was 40.89 km/h, the congestion density was 75 pcu/km/lane, and the traffic capacity was 1480 pcu/h/lane. The basic graph fitting effect was excellent, verifying the effectiveness of multi-source data fusion and weighted least squares method in identifying the characteristic parameters of expressway traffic flow.