Identification of traffic profiling scheme from operational data


Zheng, P., McDonald, M. and Jeffery, D. (2008) Identification of traffic profiling scheme from operational data. In, 87th Annual Meeting of the Transportation Research Board, Washington, USA,

Download

Full text not available from this repository.

Description/Abstract

Historic traffic profiles are widely used in many traffic applications as baseline traffic conditions for comparison purposes. They are usually constructed from traffic data obtained from a number of previous similar days. The most widely used profiling scheme is based on natural day groups such as weekdays, generally selected based on common sense and limited observations. The research reported here has increased knowledge of profiling schemes through a large scale investigation of traffic patterns using operational traffic flow and travel time data obtained over an extensive period. A new traffic profiling scheme has been developed using a data-driven approach. Operational traffic and travel time data collected by the NTCC (National Traffic Control Centre) on the strategic roads across England over a period of two years have been used to identify the most similar day groups. A general scheme with 10 day-groups and an exponential updating algorithm has been identified. Test results have revealed that the new scheme performed better than the traditional weekday based scheme. The research established a comprehensive understanding of traffic flow and travel time patterns in historical data. It is believed that the profiling scheme and methodologies developed in this research could be useful for traffic data profiling practice

Item Type: Conference or Workshop Item (Paper)
Related URLs:
Subjects: T Technology
T Technology > TA Engineering (General). Civil engineering (General)
Divisions: University Structure - Pre August 2011 > School of Civil Engineering and the Environment
ePrint ID: 52816
Date Deposited: 26 Aug 2008
Last Modified: 27 Mar 2014 18:36
URI: http://eprints.soton.ac.uk/id/eprint/52816

Actions (login required)

View Item View Item