中文摘要：

根据道路交叉口交通流量相关性,提出一种交叉口检测器均衡布设优化方法.在分析检测器数据相关性的基础上,建立交叉口所有检测器的多元逐步回归模型,并得到检测器相关性矩阵,采用极大无关组确定检测器优化布设规模;借鉴经济学中基尼系数的概念,以基尼系数最小作为检测器均衡布设的优化目标,从极大无关组中优选得到最佳检测器布设位置;采用福州五一五四路交叉口实际数据对模型有效性进行验证,并对预测误差进行详细分析.结果表明：对非关键检测器数据预测的平均绝对误差与平均相对误差分别为10.44辆车/h和8.13%;显著性检验中的F值与布设检测器的规模具有很强的线性相关性.

英文摘要：

This work proposed an optimal location method of traffic detectors based on the correlation of flow at urban intersections.Multiple stepwise regression models were used for each detector,and the correlation matrix of detector field data was obtained.The number of location detectors was determined by maximal linearly independent subset.By introducing the concept of Gini Coefficient in economics,an optimization model of equal distribution for the selection of key detectors was set under the minimum Gini Coefficient.The field intersections data from Wuyi-Wusi Road in Fuzhou were used to verify the validity of the proposed model and to analyze the trend of prediction error.Experimental results show that the mean absolute error and the mean relative error of volume were 10.44 vehicle/h and 8.13%,respectively.The F value for test of significance has a strong linear correlation with the optimal number of traffic detectors.