基于知识与数据驱动的高速路车辆速度计算方法模型研究 postprint

Abstract: Due to the current high accuracy of positioning via cellular base stations and the uncertainty of signaling data, estimating the accuracy of traffic information parameters using cell phone signaling data remains a significant challenge. Highway speed parameters are one of the most commonly studied parameters for evaluating road service quality. At present, the basic steps of traffic parameter estimation based on cell phone signaling data are as follows: Data collection and preprocessing, map matching, user route matching determination and traffic information parameter estimation. On this basis, according to the knowledge of physical kinematics, aiming at the uncertainty of handset signaling data and the influence of parallel highway noise on user data, an improved vehicle speed calculation method based on weighted distance fusion is proposed and further proposed on this method A Method of Vehicle Speed Calculation to Eliminate Parallel Highway Interference Data and Improve Calculation Precision. Using a platform provided by a mobile company for instance verification, and using the average absolute error MAE, root mean square error RMSE, average absolute percentage error MAPE as a performance indicator, in this paper, the performance of the method model is better than the traditional method model, and the average absolute percentage error is about 10% lower than the traditional method, which has important application value.