Multi-Stage Algorithm for Detection-Error Identification and Data Screening
-
2020-10-01
Details:
-
Creators:
-
Corporate Creators:
-
Corporate Contributors:
-
Subject/TRT Terms:
-
Publication/ Report Number:
-
Resource Type:
-
Geographical Coverage:
-
Edition:Final Report, Jan 2019 to Oct 2020
-
Corporate Publisher:
-
Abstract:During the past decades, roadside traffic detectors have been widely deployed in traffic management systems to help transportation agencies monitor and control traffic. However, detector data may contain some erroneous information caused by the malfunctioning of the detection system. Typically, these errors result from the lack of maintenance and the need for device recalibration, which can affect the decision-making processes that use detected data as their basis. To identify these errors in the database, a reliable screening algorithm is needed to examine the quality of recorded detector data, identify potential errors, and find the detection stations that require maintenance or recalibration. To fulfill such needs, this research project develops a multi-stage screening algorithm and fully considers the impacts of detector locations to the screening process as they would affect the data and the comparison source. First, the quality of data will be pre-evaluated and primary errors will be identified. Second, statistical analysis of data, according to their station locations, will be performed to confirm the pre-identified errors. Last, an in-depth review of stations will be carried out to certify the stations with potential errors. Notably, locating the detection stations with potential failures will help UDOT traffic engineers prioritize the detectors that need immediate attention, as timely actions on those detectors are essential to support the functioning of various traffic management tasks.
-
Format:
-
Funding:
-
Collection(s):
-
Main Document Checksum:
-
Download URL:
-
File Type: