Detection of Invalid Loop Detector Data in Urban Areas

Abstract

Urban traffic data can be used for purposes besides actuated signal control. However, it is important for the data to be of sufficient quality. This paper discusses a procedure for detecting invalid traffic data produced by single-loop detectors at signalized intersections. Basic quality checks—based on minimum and maximum flow thresholds—are used to detect erroneous data, whereas quality checks based on the principle of conservation of vehicles are introduced to detect inaccurate data. In most cases, the principle of conservation of vehicles implies that flow measurements have to be consistent between upstream and monitoring detectors within one intersection. The basic quality checks are appropriate for detecting erroneous data. Moreover, quality checks based on the principle of conservation of vehicles are a useful addition to basic quality checks, because they are appropriate for detecting inaccurate traffic data not detected by the basic quality checks. However, because flows are inconsistent between upstream detectors mutually in some cases, it is not always clear whether the results of this quality check are reliable. These cases were analyzed further by field observations. Finally, because the designed control procedure detects more invalid data than the detector station, it can be concluded that these macroscopic quality checks are a useful addition to the microscopic quality checks.