基于时空轨迹数据的交通流量的预警方法研究  

Research on Early Warning Method of Traffic Flow Based on Spatiotemporal Trajectory Data

作　　者：马勇 MA Yong(Qinghai Police Vocational College,Xining 810000 China)

机构地区：[1]青海警官职业学院,青海西宁810000

出　　处：《自动化技术与应用》2024年第4期138-141,共4页Techniques of Automation and Applications

摘　　要：为了完成交通流量的精准预警,提出一种基于时空轨迹数据的交通流量的预警方法。以路网约束为依据,提取时空轨迹数据特征。计算不同类型时空轨迹数据之间的时间和空间相似度,对其归一化处理,获取不同轨迹之间的时空距离,完成时空轨迹数据的聚类处理。将图卷积计算引入循环网络记忆单元,组建时空图卷积神经网络直接提取时空关联特征。通过图卷积神经网络构建交通流量预警模型。实验结果表明,所提方法可以以较快的速度和较高的精度完成交通流量预警。In order to complete the accurate early warning of traffic flow,a traffic flow early warning method based on spatiotemporal trajectory data is proposed.Based on the constraints of road network,the characteristics of spatiotemporal trajectory data are extracted.It calculates the time and space similarity between different types of spatiotemporal trajectory data,normalizes them,obtains the spatiotemporal distance between different trajectories,and completes the clustering processing of spatiotemporal trajectory data.The graph convolution calculation is introduced into the cyclic network memory unit,and the spatiotemporal graph convolution neural network is established to directly extract the spatiotemporal correlation features.The traffic flow early warning model is constructed by graph convolution neural network.The experimental results show that the proposed method can complete the traffic flow early warning with faster speed and higher accuracy.

关 键 词：时空轨迹数据 交通流量预警 聚类处理 图卷积计算 卷积神经网络 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]