面向区域交通状态预测的时空集成模型  

Spatial-temporal integrated model for regional traffic state prediction

作　　者：吴家成 王洪钰 肖建力[1] WU Jiacheng;WANG Hongyu;XIAO Jianli(School of Optical-Electrical and Computer Engineering,University of Shanghai for Science and Technology,Shanghai 200093,China)

机构地区：[1]上海理工大学,光电信息与计算机工程学院,上海200093

出　　处：《上海理工大学学报》2021年第5期474-483,共10页Journal of University of Shanghai For Science and Technology

基　　金：国家自然科学基金资助项目(61603257)。

摘　　要：在智能交通系统中,交通状态预测发挥着至关重要的作用。针对现有的交通预测方法集中于中微观层面,且时间和空间维度单一的问题,提出了一种面向区域宏观交通状态预测的集成模型。该模型以交通指数为依据,在时间维度上采用时间序列预测方法获得时间预测结果,在空间维度上采用支持向量回归方法获得空间预测结果,并在集成模型中将两者的结果融合。在交通指数云图上的实验结果表明,与单一维度的时间或空间模型相比,该模型能显著提高预测精度。Traffic state prediction plays an important role in intelligent transportation systems.Aiming at the problem that the existing traffic forecasting models focus on the medium and micro level,and their temporal and spatial dimensions are single,an integrated model for macroscopic traffic state forecasting was proposed.Based on the traffic index,the time series prediction model was adopted to obtain the temporal predictive results in the time dimension,and the support vector regression model was adopted to obtain the spatial predictive results in the spatial dimension.The results of two models were fused in an integrated model.Through experiments on the traffic index cloud maps,the results show that significant improvement in the prediction accuracy can be achieved by the proposed integrated model compared with the single temporal or spatial model.

关 键 词：区域交通流 交通状态预测 时间预测模型 空间预测模型 模型集成 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]