摘要
为了能够更加准确地预测高速公路服务区车流量,从而提高服务区智能化管理的效率,针对服务区车流量数据周期性变化、影响因素复杂与时序特征相互作用的特点,提出了基于Attention机制与双向长短期记忆网络的服务区车流量时序优化预测模型。使用重庆市大观服务区车流量数据进行案例分析,结果表明所提模型与传统时序预测模型相比预测准确度更高,能够为服务区车辆管理与交通拥堵疏导提供可靠的支撑。
In order to better accurately predict the traffic flow of expressway service area, so as to improve the efficiency of intelligent management of service area, in view of the characteristics of periodic change of traffic flow data in service area, complex influencing factors and interaction of time series characteristics, a time series optimization prediction model of traffic flow in service area is proposed based on the Attention mechanism and bidirectional long-term and short-term memory network. The results show that the prediction accuracy of the proposed model is higher than that of the traditional time series prediction model, which can provide reliable support for vehicle management and traffic congestion management in the service area.
作者
陈力云
薛彦聪
黄宏程
王卫平
CHENLi-yun;XUE Yan-cong;HUANG Hong-cheng;WANG Wei-ping(Chongqing Expressway Road Network Management Company Limited,Chongqing 401120,China;Chongqing University of Posts and Telecommunications,Chongqing 400065,China)
出处
《公路》
北大核心
2022年第6期212-217,共6页
Highway
基金
重庆交通局科技项目“基于高速公路服务区视频的大数据引用分析研究及示范应用”。
关键词
高速公路服务区
车流量预测
双向长短期记忆神经网络
注意力机制
expressway service area
traffic flow prediction
bidirectional long and short-term memory neural network
attention mechanism
