Effects of incidents on networks with variable message signs

Abstract

Network modelling offers a controlled environment in which the potential effects of VMS can be studied in a variety of incident/VMS scenarios. The modelling results are derived using the 'single-day' version of the RGCONTRAM model based on the Southampton network. Two key performance measures have been used to assess the effectiveness of a VMS strategy: the journey time savings of network drivers and the journey time savings of incident drivers. In general, it was found that incident drivers benefit from diversion, even those incident drivers who do not divert, because of reduced congestion on the incident route. However, there can be substantial disbenefits to non-incident drivers, particularly for those travelling on diversion routes. The results describe the effects of VMS information, which vary significantly with incident severity, incident duration, traffic demand, VMS duration and the diversion routes available.

From August 1^st , 1999 to December 20, 2000, information on incidents occurring in Southampton, the VMS strategies used, and traffic data over the network have been collected and analysed. All major incident cases on the four main arterial in Southampton have been studied. It was found that drivers make a wide range of diversion decisions including both 'early' and 'late' diversion at VMS. Early diversions were defined as those which occurred shortly after the sign was seen. The results of incident case studies indicated that between 0-34% of relevant drivers made early diversions at VMS. The variations were dependent on the number and the quality of viable diversion routes, messages displayed and peak types. Such early diversion can be entirely the result of the information provided, because no abnormal queues would have been encountered by drivers before diversion. Also, late diversions were widely found in the incident cases studied. In most incident cases, the late diversion rates were found to be higher than the corresponding early diversion rates, which indicated that drivers were more likely to divert after their acquired VMS information was confirmed by the observed queues. An application of the implications of these post mortem results on the model findings has been made.

More information

Identifiers

Catalogue record

Export record

ASCII CitationAtomBibTeXData Cite XMLDublin CoreDublin CoreEP3 XMLEndNoteHTML CitationHTML CitationHTML ListJSONMETSMODSMPEG-21 DIDLOpenURL ContextObjectOpenURL ContextObject in SpanRDF+N-TriplesRDF+N3RDF+XMLRIOXX2 XMLReferReference ManagerSimple Metadata