Sie können Bookmarks mittels Listen verwalten, loggen Sie sich dafür bitte in Ihr SLUB Benutzerkonto ein.
Medientyp:
E-Artikel
Titel:
Using automatic number plate recognition data to investigate the regularity of vehicle arrivals
Beteiligte:
McLeod, Fraser N.;
Cherrett, Tom J.;
Box, Simon;
Waterson, Ben J.;
Pritchard, James A.
Erschienen:
TU Delft OPEN Publishing, 2017
Erschienen in:
European Journal of Transport and Infrastructure Research (2017)
Sprache:
Englisch
DOI:
10.18757/ejtir.2017.17.1.3181
ISSN:
1567-7141
Entstehung:
Anmerkungen:
Beschreibung:
This paper uses automatically-recorded vehicle number plate data from a network of 22 cameras in Dorset, UK, to investigate the extent to which regular trip making can be determined using the regularity of individual vehicle arrival times across the same sites and time intervals over extended periods of several months and illustrates how a cohort of recognised regular vehicles may provide indicative evidence of traffic delays. Regularity was defined based on minimum numbers of observations over a given period and with specified maximum values of standard deviation in arrival time, with sensitivity to different values being tested. It was found that around one-fifth of all vehicles were regular during the morning peak where the definition required at least 30 observations out of 210 working days and with a standard deviation in arrival time of no more than ten minutes; significantly fewer vehicles were found to be regular in the afternoon peak. The turnover, or churn, of regular vehicles was found to be considerable, with only one-tenth of defined regular vehicles being continuously regular throughout the period and with identified pools of regular drivers halving in size every three months, as vehicles ceased to be regular and where the pool was not updated. This suggests that any database of regular drivers should be updated at least quarterly to ensure that new regular vehicles are included and that old ones are discarded. These findings may have inferences for traffic information systems tailored for different driver groups according to assumed levels of network knowledge.