固定检测器和移动检测器的交通信息融合方法

Data Fusion Method for Traffic Information Detected by Fixed Detectors and Floating Cars

固定检测器和移动检测器在检测参数、数据精度、覆盖范围、采集成本等方面存在较大差异,具有很强的互补性。分析了固定检测器与移动检测器进行信息融合的必要性,提出了交通信息融合的总体框架。在分别阐述了基于固定检测器和基于移动检测器的区间平均速度估计方法基础上,采用BP神经网络对区间平均速度进行信息融合。以上海市南北高架道路为对象,利用Vissim仿真软件对基于BP神经网络的交通信息融合方法进行了实例分析,结果表明,该方法可以明显提高区间平均速度的精度。

Fixed detectors and mobile detectors are different at detecting parameters, data precision, coverage area and cost. They are complementary. The necessity of data fusion for fixed detectors and floating cars was analyzed. The architecture of traffic data fusion was proposed. On the basis of discussing average speed calculating models based on fixed detectors and floating cars, a data fusion method by Back Propagation Neural Network was presented. At last, a Vissim software was used to validate a data fusion model. The result shows that Back Propagation Neural Network model can greatly improve data precision.