基于自适应加权平均的路网MFD估测融合方法

路网宏观基本图(Macroscopic Fundamental Diagrams，MFD)的估测方法有基于固定检测器数据估测法和基于浮动车数据估测法，但很少有文献将两者结合起来，鉴于此，本文提出以车联网环境下联网车数据估测的交通参数为检验数据，引入动态误差，建立两个自适应加权平均数据融合模型，对两种估测法所得的路网加权交通流量和路网加权交通密度分别进行数据融合，从而更加准确地估测路网MFD.为验证模型的有效性，以广州天河区核心路网为研究区域，通过Vissim交通仿真建模分析，对比各种估测法所得路网MFD参数的平均绝对相对误差、路网MFD的状态比和差异值.结果表明，经数据融合后的路网MFD参数平均绝对相对误差和路网MFD差异值均最小，最接近标准路网MFD.

Macroscopic Fundamental Diagram Estimation Fusion Method of Road Networks Based on Adaptive Weighted Average

The macroscopic fundamental diagram (MFD) of road networks can be based on loop detector data and floating car data estimation methods. However, few studies have combined these methods. In view of this research gap, this paper uses the traffic parameters estimated from the 100% network car data of the Internet of Vehicles as the test data, and two adaptive weighted average (AWA) data fusion models are established by introducing the dynamic error to fuse the weighted traffic flow and weighted traffic density of the road network obtained by two estimation methods to estimate the road networks’ MFD accurately. The validity of this model is verified with the core road network in the Tianhe District of Guangzhou as the research area. Through Vissim traffic simulation modelling and analysis, the mean absolute percent error of MFD parameters of the road networks and the state ratio and the difference value of road networks’MFD are compared. The results show that after AWA data fusion, the mean absolute percent error of MFD parameters and the difference value of the road networks’ MFD are the smallest, which is closest to the standard road networks’ MFD.