固定交通检测器布设对快速路交通流量预测影响的实证研究  

Empirical investigation of the effects of fixed traffic detector deployment on expressway traffic flow forecasting

作　　者：董春肖 张立涛[1] 孙洪运 张军[1] 张昕冉 DONG Chunxiao;ZHANG Litao;SUN Hongyun;ZHANG Jun;ZHANG Xinran(Management School,Shandong University of Technology,Zibo 255000,China)

机构地区：[1]山东理工大学管理学院,山东淄博255000

出　　处：《铁道科学与工程学报》2022年第2期359-366,共8页Journal of Railway Science and Engineering

基　　金：国家自然科学基金青年基金资助项目(71801145)。

摘　　要：交通检测器是智能交通系统的重要组成部分,交通检测器的布设策略对交通精细管控有重要影响,优化交通检测器的布设方案能够有效缓解交通拥堵问题。针对交通流量预测未充分考虑交通检测器布设方案的情况,从交通检测器的间距和数量2个方面对短时交通流量预测的影响进行实证研究。采用实证研究及BP神经网络算法,探究交通检测器的布设方案对短时交通流量预测的影响。采用20组交通检测器为研究载体,采集20组交通检测器的间距信息及每组交通检测器时长共30 d的短时交通流数据为研究数据,使用BP神经网络对当前组别的交通检测器及其上游相邻的交通检测器在高峰时段断面的交通流量进行预测。针对长路段,分别选择2~K组检测器数据预测交通流量。实验结果表明:相邻路段的交通流量具有空间关联性,交通检测器的最优布设间距处在330~450 m范围内,长路段选择2~3组检测器能极大地降低预测误差,优化布设策略与传统方法相比可降低误差20%以上。采用优化布设策略选择适当间距及数量要求的交通检测器可以显著提高短时交通流量预测的准确度。Traffic detector is part of the intelligent transportation system.The deployment strategy of traffic detectors has an important impact on the fine traffic control.Optimizing the layout of traffic detectors can effectively alleviate traffic congestion.However,traffic flow prediction has not fully considered traffic detector layout scenario,this study empirically investigated how spacing and number of traffic detectors affect short-term traffic flow prediction accuracy.The method of empirical research and BP neural network algorithm was used to explore the impact of the layout of traffic detectors on short-term traffic forecasts.The 20 sets of traffic detectors were used as the research carrierto collect the distance information of 20 sets of traffic detectors,30-day shortterm traffic flow data of each set of traffic detectors were used as the research data,and the BP neural network was used to predict the traffic flow of the current group of traffic detectors and their upstream neighboring traffic detectors during peak hours.For long road sections,2 to K groups of detector data were respectively selected to predict traffic flow.The experimental results show that the traffic flow of adjacent road sections is spatially correlated,and the optimal layout distance of traffic detectors is within the range of 330~450 m.Choosing 2 to 3 sets of detectors for long road sections can greatly reduce the prediction error.Compared with the traditional method,the proposedstrategy can reduce the error by more than 20%.Therefore,using an optimized layout strategy to select traffic detectors with appropriate spacing and quantity requirements can significantly improve the accuracy of short-term traffic flow prediction.

关 键 词：智能交通 交通流量预测 神经网络 交通检测器 检测器布设 

分 类 号：U491[交通运输工程—交通运输规划与管理]