基于SVM和Kalman滤波的BRT行程时间预测模型研究

摘要

本文提出了一个支持向量机进行初始行程时间预测并结合卡尔曼滤波算法进行动态调整的快速公交车行程时间综合预测模型.以快速公交车运行的GPS数据为基础,对北京市朝阳区快速公交2号线进行行程时间预测案例研究.利用该模型对其早高峰和上午平峰的两个不同时段的公交行程时间分别进行预测和对比分析,并通过与单一的卡尔曼滤波方法所得的预测结果进行比较.结果表明,该模型应用于快速公交行程时间预测具有更好的适用性,并且预测平峰时段的精度要高于高峰时段.%An integrated BRT vehicle travel time prediction model is proposed. It uses the support vector machine ( SVM) to predict the initial travel time and apply the Kalman filter algorithm to dynamically adjust the results of predicted travel time. Based on the global pasitioning system (GPS) data, a case study of the BRT line 2 in Chaoyang district of Beijing is conducted with the proposed model. Bus travel time during the morning peak hours and off-peak hours are predicted by both the proposed model and the Kalman filter model. The results show that the proposed model is more suitable to predict the bus travel time with high prediction accuracy, and the accuracy for off-peak hours is higher than the one for peak hours.