The invention discloses a multi-graph hierarchical road network characterization method based on a graph neural network, and the method comprises the steps: carrying out the modeling of road network characterization at different levels through a structure perception graph neural network composed of spectral clustering and a graph attention network, and introducing two types of virtual nodes, namely a structure region and a function region, a multi-graph mechanism is adopted to guide a virtual node to correspond to a structural area and a functional area of a real world, structural similarity is established by utilizing road type attributes, functional similarity between road sections is defined by utilizing urban POI information, message sharing is executed at a high level, and then updated information is spread to low-level nodes, so that the real-time performance of the road is improved. Functional attributes of the road network are supplemented; by using the method, road network characterization containing structural features and functional features can be obtained; a road network structure and a function role can be conveniently determined; the method is helpful for revealing functional areas of a city, is helpful for route planning, arrival time estimation and position prediction, and is helpful for construction of an intelligent traffic system.

    本发明公开了一种基于图神经网络的多图分级路网表征方法,该方法中通过谱聚类和图注意力网络组成的结构感知图神经网络,对不同层次的路网表征进行建模,引入了两种虚拟节点,即结构区和功能区,采用了多图机制来引导虚拟节点与现实世界的结构区域和功能区域相对应,利用道路类型属性建立结构相似性,利用城市POI信息定义道路段之间的功能相似性,在高层次执行消息共享,然后将更新的信息传播到低层节点,并补充路网的功能属性;使用该方法可以得到包含结构特征和功能特征的路网表征;便于确定路网结构及功能角色;有助于揭示城市的功能区,有助于路线规划以及到达时间估计和位置预测,有利于智能交通系统的构建。


    Zugriff

    Download


    Exportieren, teilen und zitieren



    Titel :

    Multi-graph hierarchical road network characterization method based on graph neural network


    Weitere Titelangaben:

    一种基于图神经网络的多图分级路网表征方法


    Beteiligte:
    WANG JINGYUAN (Autor:in) / MA JINGTIAN (Autor:in) / LI CHAO (Autor:in)

    Erscheinungsdatum :

    2023-05-26


    Medientyp :

    Patent


    Format :

    Elektronische Ressource


    Sprache :

    Chinesisch


    Klassifikation :

    IPC:    G06F ELECTRIC DIGITAL DATA PROCESSING , Elektrische digitale Datenverarbeitung / G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen / G08G Anlagen zur Steuerung, Regelung oder Überwachung des Verkehrs , TRAFFIC CONTROL SYSTEMS




    Intelligent Graph Convolutional Neural Network for Road Crack Detection

    Djenouri, Youcef / Belhadi, Asma / Houssein, Essam H. et al. | IEEE | 2023


    MCAGCN: Multi‐component attention graph convolutional neural network for road travel time prediction

    Zhihua Zhao / Li Chao / Xue Zhang et al. | DOAJ | 2024

    Freier Zugriff

    Marine multi-ship encounter graph structured learning method based on graph convolutional neural network

    GAO MIAO / ZHANG ANMIN / HAN YU et al. | Europäisches Patentamt | 2024

    Freier Zugriff

    Urban road congestion space-time prediction method based on graph process neural network

    WANG JIANLONG / YI CHONGZHENG / ZHANG LI et al. | Europäisches Patentamt | 2023

    Freier Zugriff