城市快速路车辆检测器数据质量评价与控制方法

Quality evaluation and control method of vehicle detector data on urban expressway

为了提高车辆检测器数据的准确性,设计了城市快速路车辆检测器数据质量评价与控制方法,提出了一种综合考虑单一地点与流量守恒定律的数据质量评价与控制流程。通过分析路段上下游检测截面的流量累积曲线特性,设计了基于流量累积曲线的数据质量评价方法。以路网所有检测截面累积流量调整值与观测值差的平方和最小为目标函数,以流量累积曲线的特性为约束条件,设计了基于最优化技术的数据质量控制方法。采用某特大城市快速路实测数据对本文方法进行了验证和对比分析。验证结果表明:当引入误差分别为0.025、0.050、0.075、0.100时,采用评价与控制方法的计算结果的平均绝对百分比误差分别为2.40%、3.83%、5.70%、8.26%,评价正确率在97%以上,明显优于其他对比方法。

In order to improve the accuracy of vehicle detector data,a quality evaluation and control method of vehicle detector data on urban expressway was proposed.A data quality evaluation and control process was proposed by comprehensive considering a single location and the law of flow conservation.A data quality evaluation method based on flow cumulative curves was proposed by analyzing the characteristics of the flow cumulative curves of upstream and downstream detected sections.With the objective function that the square sum of the difference between the adjusted and the observed flow cumulative value of vehicles for all detected sections in the road network was minimum,and with the restriction condition of the characteristics of flow cumulative curves,a data quality control method based on the optimization techniques was proposed.The evaluation and control method was verified and compared by using the data collected from inductive loop detectors on a megacity urban expressway.Verified result indicates that when the introduced errors are 0.025,0.050,0.075,and 0.100,the mean absolute percentage errors computed by the method are 2.40%,3.83%,5.70% and 8.26% respectively,and its accuracy is more than 97%,so the method has better precision than comparison methods.