摘要: |
Traffic congestion has become a major problem in many urban areas, and has related environmental, economic, and equity impacts. One potential method of reducing urban traffic congestion is developing tools that plan multi-modal trips to encourage more people to ride public transportation and to provide better driving alternatives for less affluent citizens. Traffic state prediction is the key component to planning multi-modal trips in a complex transportation network. This research attempts to address transportation system state prediction problems considering private vehicles, public transit, and bike share services within the context of a multimodal transportation system. For public transit service, the proposed effort focuses on developing real-time passenger demand prediction models using multiple data sources to enhance prediction accuracy. For bike share services, the proposed effort focuses on developing prediction models for the number and travel times of bikes. Finally, for private vehicles, this research develops a comprehensive traffic prediction tool by including different categories of prediction models. The proposed prediction algorithms and tools are evaluated by comparing their performance using the field data collected in multimodal transportation system to the performance of existing prediction methods using the same data. |