Validation of Origin–Destination Data from Bluetooth Reidentification and Aerial Observation. Issue 1 (January 2014)

• Record Type: Journal Article
• Authors: Chitturi, Madhav V.; Shaw, John W.; Campbell, John R.; Noyce, David A.
• Is Part Of: Transportation research record. Volume 2430:Issue 1(2014)
• Page Start: 116
• Page End: 123
• Publication Date: 2014-01
• Subjects: Transportation -- Periodicals; Roads; Transport -- Périodiques; Routes -- Périodiques; Routes -- Conception et construction -- Périodiques; Roads; Transportation; 388.05
• DOI: 10.3141/2430-12
• Languages: English
• ISSNs: 0361-1981

Abstract

Vehicle reidentification with Bluetooth signal data has emerged as an effective and economical means for collecting traffic data including origin–destination (O-D) information, which is crucial for transportation planning. Direct vehicle tracking based on time-lapse aerial photography (TLAP) is also increasingly used for O-D studies. Neither technique has previously been validated, so the objective of this study was to validate O-Ds generated by using both techniques against a ground-truth O-D. Traffic volume–, Bluetooth-, and TLAP-based O-D data collected at an interchange in Madison, Wisconsin, were used in this study. Significant variability (2.3% to 7.2%) in Bluetooth match rates was observed for the 12 O-D pairs of the interchange, which were calculated by using a heuristic measure of the difference between traffic volumes (the GEH formula). Uniproportional scaling of the sample Bluetooth O-D (using the average detection rate) resulted in poor fit to the true O-D, but biproportional factoring (Fratar or Furness procedure) resulted in about 85% of the movements' having a GEH statistic less than 5. Combining Bluetooth data from multiple days resulted in a better fit than the use of a single day's data. The analysis showed that Bluetooth technology could be used to obtain representative O-D information of an access-controlled network for planning applications. TLAP sample O-D data were scaled with an origin-specific factor and resulted in reasonable fit to the true O-D data. Biproportional factoring resulted in a better fit in which 100% of the movements had a GEH value less than 5. Therefore, biproportional factoring is also recommended for TLAP data.