Study on Travel Time Prediction Method Based on License Plate Data

In order to obtain more accurate travel time prediction results, this study proposes a travel time prediction method based on license plate data. The main steps are introduced as follows: firstly, it presents exponential smoothing of the historical travel time series based on the license plate data; and then the path detection method is used to query each intersection to meet the conditions; finally, all paths and the final prediction results are obtained from starting and ending points, and stacking the travel time of each path. Compared with other relative models, this prediction method reduces the sampling time span and enhances the correlation of the travel time series, ensuring the number of samples collected in order to increase the accuracy of the prediction results. The simulation results prove that the final prediction method is feasible and effective on the basis of less than 10% of travel time prediction errors of each section.