基于卡口数据的城市道路交通空间相关性分析  

Spatial correlation analysis of urban road traffic based on bayonet data

作　　者：宋大创 马晓旦[1] 夏晓梅[1] 孙明洁[1] SONG Dachuang;MA Xiaodan;XIA Xiaomei;SUN Mingjie(School of Management,University of Shanghai for Science and Technology,Shanghai 200093,China)

机构地区：[1]上海理工大学管理学院,上海200093

出　　处：《智能计算机与应用》2021年第8期151-153,157,共4页Intelligent Computer and Applications

摘　　要：在短时交通流预测中,道路交通空间相关性是客观存在的。现有研究在度量道路交通空间相关性上,通常采用时间序列数据直接进行统计分析,或是假定一定距离内具有空间相关性等。但这些方式忽略了道路之间交通影响的空间异质性。在卡口数据中,由于车辆牌照的唯一性特性,不仅可以计算出研究路段的交通流时间序列,还能得到每辆车的行驶轨迹。本文通过车辆轨迹,得到流量转移权重矩阵和不同卡口数据量化的网络权重矩阵,构造一个新的网络权重矩阵,度量城市道路之间的空间相关性,然后进行短时交通流预测及效果比较,验证了方法的可行性和有效性。In short-term traffic flow forecasting,the spatial correlation of road traffic exists objectively.Existing studies usually use time series data for direct statistical spatial correlation within a certain distance.However,these methods ignore the spatial heterogeneity of traffic impacts between roads.In the bayonet data,due to the unique characteristic of the vehicle license plate,not only the time series of the traffic flow of the research section can be calculated,but the trajectory of each vehicle can also be obtained.this paper combines the traffic transfer weight matrix obtained by vehicle trajectory and the network weight matrix quantified by different bayonet data to construct a new network weight matrix to measure the spatial correlation between urban roads.And then carry out short-term traffic flow forecasting and effect comparison,verifying the feasibility and effectiveness of the method.

关 键 词：卡口数据 城市道路网络 空间相关性 车辆轨迹 

分 类 号：P208[天文地球—地图制图学与地理信息工程]