基于统计分析的交通检测设备性能诊断方法

A Novel Approach of Performance Diagnostics for Traffic Monitoring Equipment Based on Statistical Analysis

针对智能交通(ITS)应用系统中检测设备长期性能的健康监测问题,在海量实测交通数据的分析挖掘基础上,提出了一套定量化的诊断方法.该方法以稳定性良好的实测交通流参数——自由流车速为主要指标,采用参数估计、假设检验与置信区间分析相结合的方式,完成针对检测数据的量化分析和检测设备性能状态判断.利用浮动观测车,在上海城市快速道路交通监控系统所积累的历史和实时数据中验证了该方法的有效性.

Identifying the ＂health problem＂ of traffic monitoring equipments is important to long-term performance of intelligent transportation system （ITS） application system. Based on statistical distribution analysis of plentiful traffic data archives, a quantitative method to seek the potential problem of traffic observations is proposed. Employing free-flow speed, a steady parameter of traffic flow, as a main indicator, parameters estimation, hypothesis test and confidence interval analysis are jointly used to determine the health condition of traffic detectors. The methodology is demonstrated through application to the historical and real-time data in the Advanced Traffic Management System of Shanghai urban expressway system, with validation by floating car measurement.