论文部分内容阅读
旅行时间作为交通系统运行的关键参数,可以为交通诱导系统和出行者路径选择提供决策建议。利用多源数据进行旅行时间的估计是智能交通系统运行的重要支撑。利用基于同一路段的3种检测数据,提出相应的权重分配模型和神经网络模型来进行多源检测数据的融合以获得融合后的旅行时间。对比研究了基于多断面检测器的旅行时间的2种推算方法:速度累进和速度平均。利用北京市典型道路数据对这2种融合技术的融合效果进行了对比分析,结果显示,多源数据融合可以提高旅行时间估计的准确性。
Travel time, as a key parameter of traffic system operation, can provide decision-making suggestions for traffic guidance system and traveler’s route selection. The estimation of travel time using multi-source data is an important support for the operation of ITS. Using the three kinds of detection data based on the same road segment, the corresponding weight distribution model and neural network model are proposed to fuse the multi-source detection data to obtain the merged travel time. Two kinds of estimation methods of travel time based on multi-section detector are comparatively studied: speed progressive and speed average. By using typical urban road data in Beijing, the fusion effects of the two fusion technologies are compared and analyzed. The results show that the multi-source data fusion can improve the accuracy of the travel time estimation.