Modeling traffic flow as exchange interactions in an "atmosphere" field shared with drivers

2014 Fiscal Year Final Research Report

• Project/Area Number: 25820247
• Research Category: Grant-in-Aid for Young Scientists (B)
• Allocation Type: Multi-year Fund
• Research Field: Civil engineering project/Traffic engineering
• Research Institution: Tokyo University of Science
• Principal Investigator: KASAI Makoto 東京理科大学, 理工学部, 助教 (20579792)
• Project Period (FY): 2013-04-01 – 2015-03-31
• Keywords: 交通流 / 交通容量 / 容量上の隘路現象 / 交換相互作用 / ドライビングシミュレータ / 追従積重ね試験 / 階層ベイズ

Outline of Final Research Achievements

This study discusses an effect of "atmosphere" shared with neighboring drivers in traffic flow to model and mitigate bottleneck phenomena in access-controlled section. First, traffic flow is modeled as exchange interaction in time headway instead of modeling car-following behavior based on homogenous traffic flow data collected by a driving simulator. Second, Hierarchical Bayesian Estimation is applied to estimation of robustness of exchange interaction in time headway. Markov Chain Monte Carlo methods and Type-II maximum likelihood estimation method are implemented for the estimation. As a result, the interaction in time headways in chronological order dominates in other interactions such as those of successive cars. Finally, a measure for improvement of discharge flow rate is tested. In the experiment, an effect of acceleration of neighboring cars is measured by a driving simulator. Some drivers steadily follow the leader in case of high acceleration rate of neighboring cars.

Free Research Field

交通工学