摘要
参考了国内外的研究成果,通过利用实时采集的高速公路联网收费数据,结合交通状态指标参数选取影响因素,综合考虑交通运行状态的影响因素,选取交通流量、平均行程速度以及道路占有率作为交通状态的识别指标;并且从输入参数选择,模型标定两方面详细说明了基于BP神经网络的路网交通状态判别模型的构建方法,并给出了模型的具体求解流程;最后以河南省高速公路基本路段收费站采集的数据为基础,分析了收费站数据的提取和处理过程,并且以MATLAB作为仿真媒介,完成基于BP神经网络的路网交通判别方法的仿真计算,使用抽样数据进行训练和测试并给出结果。
On the basis of research outcomes both domesticand aboard, the real -time networked toll collection data is used in this paper, then considering the influence factors choosing the traffic status index, the traffic volume, average traffic velocity, and lane occupancy are chosen as the index talked before; additionally, the BP neural network modeling for highway network status judgment is built and the specific computation steps is given; at last on the base of Henan toll collection plaza data from basic section, the data extraction and treatment process are analyzed, and via MATLAB simulation toll, the BP neural network simulation computation method is accomplished, plus the outcome of training and testing using sample data is given.
基金
河南省交通运输科技计划项目基金资助
项目编号:2012P28
教育部博士点基金资助
项目编号:20120092110044
关键词
高速公路
联网收费
交通拥挤识别
BP神经网络
highway
networked toll collection
trafficcongestion detection
BP neural network
