RESEARCH OF TRANSPORTATION NETWORK TRAFFIC VOLUME FORECAST BASED ON REAL TIME DETECTOR DATA

摘要

Traffic volume forecast is an important task for an intelligent transportation system. Accurate forecast results can assist traffic management department with making traffic control strategies in advance. Improved nonparametric and neural network methods are used in traffic volume forecast. Application of Grey system theory is also studied. A fusion forecast method is developed to improve forecast accuracy on the basis of analyzing other methods. Traffic volume data from Beijing's 2nd ring road are used to test each of the methods mentioned above. The test results indicate that forecast method based on data fusion performs better and has practical value.