Combination Prediction of Traffic-Volume in Intersections Based on Wavelet Analysis and Kalman Filter

Based on the analysis of the distribution characteristics of urban traffic-volume, a combination prediction method based on wavelet analysis and Kalman filter was proposed to obtain more precise information on traffic-volume in intersections. With this method, some correlative data serials from an intersection are constructed, each data serial is treated respectivelywith Kalman filter and wavelet transfer, and the final result is formed fromthe predicted results of these serials. This method may be widely applied inthe fields of dynamic traffic prediction, such as the predictions of road-sect traffic-volume and traffic- volume distribution of intersections. Experimental results showthat the method can evidently decrease errors in the prediction and shows preferable robustness.