De-Noising Sequential Data Assimilation System Based on Empirical Mode Decomposition and Its Applications for Short-Term Traffic Flow Forecasting

摘要

This paper investigates a de-noising sequential data assimilation system and applies it into short-term traffic flow prediction. Models in traditional sequential data assimilation system are usually constructed using historical measurements. They always disturbed by local noises. Simultaneously, the accuracy of assimilation results will be affected. Developed de-noising sequential data assimilation system can separate measurements noises based on empirical mode decomposition to reduce their influences on model and assimilation results accuracy. And then applications into short-term traffic flow prediction using de-noising sequential data assimilation system and traditional sequential data assimilation system are presented. Experimental researches are based on the traffic flow measurements collected from a sub-area of highway between Liverpool and Manchester, England. Results indicate that de-noising sequential data assimilation system can successfully reduce effects of measurements noises on model construction and improve the accuracy of short-term traffic flow prediction when compared with traditional sequential data assimilation system.