The invention discloses a traffic state prediction method and system fusing multi-period and spatial semantic association, and belongs to the field of traffic state prediction. Comprising the following steps: a training stage: acquiring original traffic state data X of each road section in an actual road network; constructing a training sample set by using the traffic state data X, and training to obtain a traffic state prediction model; wherein each training sample comprises recent related data Xr, daily period related data Xd, weekly period related data Xw and a training target Ytrue; and an application stage: processing historical data corresponding to traffic state data to be tested into recent related data, daily period related data and weekly period related data, inputting the data into the trained traffic state prediction model, and outputting a traffic state in a prediction time slice. According to the method, the influence of multi-periodicity and semantic information of the traffic state data on prediction is considered at the same time, and the traffic state prediction precision can be effectively improved.
本发明公开了一种融合多周期和空间语义关联的交通状态预测方法及系统,属于交通状态预测领域。包括:训练阶段:获取实际路网中各路段的原始交通状态数据X;利用所述交通状态数据X构建训练样本集,训练得到交通状态预测模型;其中,每个训练样本包括:近期相关数据Xr、日周期相关数据Xd、周周期相关数据Xw及训练目标Ytrue;应用阶段:将待测交通状态数据对应的历史数据处理为近期相关数据、日周期相关数据、周周期相关数据并输入至训练好的交通状态预测模型中,输出预测时间片内的交通状态。本发明同时考虑了交通状态数据的多周期性及语义信息对预测的影响,能有效地提高交通状态预测的精度。
Traffic state prediction method and system fusing multi-period and spatial semantic association
融合多周期和空间语义关联的交通状态预测方法及系统
2023-10-10
Patent
Electronic Resource
Chinese
ROAD TRAFFIC SPEED PREDICTION METHOD FUSING MULTI-FEATURE NEURAL NETWORK
European Patent Office | 2024
|Prophet-DCRNN traffic flow prediction method fusing multi-modal information
European Patent Office | 2021
|European Patent Office | 2025
|Trajectory prediction method fusing traffic state and driving style and related equipment
European Patent Office | 2024
|Fusing Visual Quantified Features for Heterogeneous Traffic Flow Prediction
DOAJ | 2024
|