Sample Size Reduction Method Based on Data Fusion for Freeways with Fixed Detectors. Issue 1 (January 2015)

• Record Type: Journal Article
• Authors: Zhan, Fengping; Wan, Xia; Zhang, Jian; Li, Rui; Ran, Bin
• Is Part Of: Transportation research record. Volume 2528:Issue 1(2015)
• Page Start: 18
• Page End: 26
• Publication Date: 2015-01
• Subjects: Transportation -- Periodicals; Roads; Transport -- Périodiques; Routes -- Périodiques; Routes -- Conception et construction -- Périodiques; Roads; Transportation; 388.05
• DOI: 10.3141/2528-03
• Languages: English
• ISSNs: 0361-1981

Abstract

This research investigates the required minimum number of probe vehicles (PVs) necessary to report the real-time travel speed and travel time with the desired statistical accuracy for freeways equipped with fixed detectors. The field data used were collected from the Freeway Performance Measurement System from 7:00 to 10:00 a.m. on May 28, 2014, during a typical weekday. With corridor information and traffic volume data, this study uses VISSIM to create simulation scenarios to generate PV data and loop data for further analysis. Then the average relative errors of travel time by setting are compared with various PV percentages in traffic on different links. The least squares errors of the fused speed data are also compared with the PV and loop data. Through analysis, the study reaches the following results. When the probe sample size reaches a certain level (3% in the testing environment), the estimated errors are reduced gently with the increasing number of PVs. To ensure that the link is traversed by at least one PV in one sample interval, at least 5-min, 5- to 10-min, and 20- to 25-min sampling intervals are required when the PV percentages are 3%, 2%, and 1%, respectively. The fusion of PV data and loop data can increase the accuracy of estimation results. About 3% of PVs are sufficient to collect highly accurate data for general freeways; 2% is enough to obtain the same or higher accuracy for the same road with fixed detectors.