摘要: |
Although there are many tools and online services, such as Google Maps, that can show drivers the roadway traffic conditions in real-time, it’s often too late given that drivers may well be approaching the bottlenecks already. Being able to accurately predict traffic congestions in about a half-hour advance is very critical for advanced trip planning and traffic management. To address this problem, this study is to develop a model that can accurately forecast the traffic conditions at a signalized intersection up to a half-hour in advance. To achieve this goal, existing methods for intersection traffic flow forecasting will be reviewed and synthesized. Cycle by cycle traffic data will be collected from a real-world signalized intersection for model development and evaluation. New models for short term intersection traffic flow forecasting will be developed with different data mining methods. The performance of the developed models will be evaluated based on the collected traffic data, and the one with the best performance will be selected. The developed model can be used for advanced trip planning and traffic management. For example, it can help the freight and logistic companies to better plan their truck dispatching schedules and routes, thereby reduce their operation cost caused by traffic congestion. |