A traffic flow prediction method fusing one-dimensional convolution and an attribute enhancement unit comprises the following steps: 1) carrying out cleaning and preliminary preprocessing on traffic road information, traffic flow information and external factor information; 2) sending the traffic flow matrix to a traffic flow feature extractor composed of a one-dimensional residual convolutional neural network for feature extraction, and outputting a traffic flow feature matrix; 3) carrying out adaptive weight splicing through an attribute enhancement unit to obtain a traffic flow characteristic matrix after attribute enhancement; 4) inputting into a traffic flow prediction model for training; 5) calculating to obtain a loss value, ending the training when the loss value of the network model is converged to the minimum or the network model is trained to a specified number of iterations, and storing and outputting the optimal model of the training; and 6) testing the performance ability of the model by using the test set. The traffic flow of the road at the future moment is predicted through an artificial intelligence method, a basis is provided for road management, and timely and effective measures are taken.
一种融合一维卷积与属性增强单元的交通流量预测方法,包括以下步骤:1)交通道路信息、车流量信息以及外部因素信息,并将数据进行清洗及初步预处理;2)将车流量矩阵送到一个由一维残差卷积神经网络组成的车流量特征提取器中进行特征提取,输出车流量特征矩阵;3)通过属性增强单元进行自适应权重拼接得到属性增强后的车流量特征矩阵;4)输入到交通流量预测模型中进行训练;5)计算得到损失值,当网络模型的损失值收敛到最小或者训练到指定的迭代次数才结束训练,保存输出训练的最优模型;6)使用测试集进行检验模型的表现能力。本发明通过人工智能方法进行道路将来时刻车流量的预测,为道路管理提供依据,做出及时有效的措施。
Traffic flow prediction method fusing one-dimensional convolution and attribute enhancement unit
一种融合一维卷积与属性增强单元的交通流量预测方法
2023-12-22
Patent
Electronic Resource
Chinese
European Patent Office | 2024
|European Patent Office | 2021
|Short-term traffic flow parameter prediction method fusing traffic event influence
European Patent Office | 2022
|Prophet-DCRNN traffic flow prediction method fusing multi-modal information
European Patent Office | 2021
|European Patent Office | 2023
|