摘要
针对目前交通流检测设备布设中出现的一些问题,借助均匀设计方法,结合神经网络处理观测交通流数据,在设置数量有限制条件下,开展布设方案优化设计研究。方案设计力图在人力、物力投入最小的前提下,通过典型路段、交叉口等交通基础设施的组合,可以更多的获取交通流参数信息,全面、准确的反映城市交通流分布情况。选用了67的均匀设计表作为方案设计基础表,经过拟水平处理后设计优化方案,运用BP神经网络构建交通量推断模型,判定设计方案的优劣,进行了实地交通流检测设备优化布设研究。通过研究可知,采取优选布设方案采集的交通量对检测路段交通量进行预测,相对偏差可以达到0.006。
Union design and neural network are adopted to settle the puzzles in traffic detectors settlng.The optimizing design research are carried out in limited setting point.By combination of crossing and main road, it is found that the method could get overall information of area tragic with minimal invested human power and material resources.The traffic flow parameters could be reflect by the information . 67 uniform design table are adopted as basic table to design scheme and BP net work are adopted as the method to establish a tragic volume model, which could be used as a reforence for optimizing scheme. The result shows that the deviation of optimization scheme could not exceed 0.006.
出处
《公路交通科技》
CAS
CSCD
北大核心
2007年第11期100-104,共5页
Journal of Highway and Transportation Research and Development
基金
国家自然科学基金资助项目(50578056)
关键词
交通工程
相对偏差减小
均匀设计
交通流检测布设方法
traffic engineering
decrease of relative deviation
uniform design
collocation scheme of traffic detect
