Development of Dynamic Hybrid Traffic Simulation Model and Its Validation Using Real Highway Dataset

• Published in: Transaction of the Japan Society for Simulation Technology Vol. 13; no. 1; pp. 37 - 47
• Main Authors: Takahashi, Toru, Abe, Kazuki, Fujii, Hideki, Ikada, Satoshi, Matsudaira, Masaki
• Format: Journal Article
• Language: Japanese
• Published: Japan Society for Simulation Technology 2021
• Subjects: Hybrid Model; Macroscopic Simulation; Microscopic Simulation; Traffic Simulation; Validation with Real Data
• ISSN: 1883-5031; 1883-5058
• DOI: 10.11308/tjsst.13.37

Various traffic simulation models have been proposed according to the phenomena of interest, the required reproducibility, and the size of the target area. One of them is a hybrid traffic simulation approach that employs two or more models with different granularity depending on the traffic conditions and the properties of road networks. Among them, a dynamic hybrid traffic simulation, which dynamically switches the application area of each model during the simulation running, seems to have the potential to be applied to a wide area while keeping high reproducibility. However, because there are few studies on this subject, the decision criteria for model switching and its effects have not been sufficiently investigated. In this study, the authors proposed a novel dynamic hybrid traffic simulation method, especially a model switching criterion, and evaluated its performance. In addition, the proposed model is validated using data obtained from real highways. The results show that the proposed model is applicable for highway simulations with high accuracy and low computational cost.