The invention discloses a travel time prediction method based on multi-feature selection and fusion, and the method comprises the steps: firstly obtaining traffic travel data including time, space and other data, and obtaining a traffic travel data vector; then, constructing a travel time prediction model; in the model, space, time and other data respectively pass through a feature extraction module to obtain a space feature vector, a time feature vector and other feature vectors; the space feature vector and the time feature vector respectively pass through a space attention unit and a time attention unit in a space-time attention module to obtain a space attention feature vector and a time attention feature vector; the space attention feature vectors and the time attention feature vectors are fused to obtain space-time features; other feature vectors pass through a self-attention mechanism to obtain other attention feature vectors; the spatial-temporal feature and other attention feature vectors are fused to obtain a travel fusion feature, and the travel fusion feature passes through a prediction module to obtain predicted travel time. According to the method, a new space-time attention mechanism is provided, the overall space-time correlation of the traffic travel data and the similarity between the traffic travel data and the historical travel data are considered at the same time, and the prediction precision is improved.

    本发明公开了一种基于多特征选择与融合的行程时间预测方法,首先获取包括时间、空间和其他数据在内的交通行程数据,得到交通行程数据向量;然后,构建行程时间预测模型;在模型中,空间、时间和其他数据分别经过特征提取模块,得到空间特征向量、时间特征向量和其他特征向量;空间特征向量和时间特征向量在时空注意力模块中,分别经过空间注意力单元和时间注意力单元,得到空间注意力特征向量和时间注意力特征向量;空间、时间注意力特征向量经过融合得到时空特征;其他特征向量经过自注意力机制,得到其他注意力特征向量;时空特征与其他注意力特征向量经过融合得到行程融合特征,行程融合特征再经过预测模块得到预测的行程时间。该方法提出了一种新的时空注意力机制,同时考虑了交通行程数据整体的时空相关性和交通行程数据与历史行程数据之间的相似性,提高了预测精度。


    Access

    Download


    Export, share and cite



    Title :

    Travel time prediction method based on multi-feature selection and fusion


    Additional title:

    基于多特征选择与融合的行程时间预测方法


    Contributors:
    GU JUNHUA (author) / LI WENYAO (author)

    Publication date :

    2024-03-26


    Type of media :

    Patent


    Type of material :

    Electronic Resource


    Language :

    Chinese


    Classification :

    IPC:    G06Q Datenverarbeitungssysteme oder -verfahren, besonders angepasst an verwaltungstechnische, geschäftliche, finanzielle oder betriebswirtschaftliche Zwecke, sowie an geschäftsbezogene Überwachungs- oder Voraussagezwecke , DATA PROCESSING SYSTEMS OR METHODS, SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTING PURPOSES / 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



    Travel Time Prediction of Multi-Source Historical Data Fusion

    Liu, Wen Ting ;Wang, Zhi Jian ;Yan, Qin | Trans Tech Publications | 2011


    Feature selection‐based approach for urban short‐term travel speed prediction

    Zheng, Liang / Zhu, Chuang / Zhu, Ning et al. | Wiley | 2018

    Free access

    Feature selection-based approach for urban short-term travel speed prediction

    Zheng, Liang / Zhu, Chuang / Zhu, Ning et al. | IET | 2018

    Free access

    Traffic flow prediction method based on multi-feature fusion

    LIU CONG'AI / QIN YUHUA / ZHANG XIAOJING | European Patent Office | 2024

    Free access

    TRAVEL TIME PREDICTION METHOD, TRAVEL TIME PREDICTION DEVICE, AND PROGRAM

    SHINAGAWA AKIO / TAMAI KYOHEI / IKEDA YUICHI | European Patent Office | 2016

    Free access