基于时空信息的交通流预测

交通流量预测是城市道路交通信号控制中的重要组成部分。为了提高预测的准确性,基于路口视频检测器数据,提出了一种交通流量预测的交通数据分析方法。随着逐年提高的计算能力,深度学习方法进行短时交通流预测越来越流行,经典的交通流量预测方法通常只能根据被预测道路自身的数据进行分析和决策,而往往较少考虑由同一区域不同道路之间的交通流量关联性。基于城市核心路网交通数据,提出一种基于时空信息的TS-LSTM模型,并与其他经典模型进行比较,所得出的结果验证了相比其他方法而言,该模型具有较高的预测精度。

Traffic Flow Prediction Based on Spatio-temporal Information

Traffic flow prediction is an important part of urban road traffic signal control.In order to improve the accuracy of prediction,a traffic data analysis method for traffic flow prediction is proposed based on the data of intersection video detectors in this paper.With the increasing computing power year by year,deep learning methods for short-term traffic flow prediction are becoming more and more popular.Classic traffic flow prediction methods can usually only analyze and make decisions based on the data of the predicted road itself.Correlation of traffic flow between different roads in the region.Based on the traffic data of the urban core road network,this paper proposes a TS-LSTM model based on spatio-temporal information and compares it with other classic models.The results obtaine verify that the model has higher prediction accuracy than other methods.