Brian Williams
Smith
System Engineering
Executive Summary
This report is concerned with sources of automatic traffic and travel data that may be used for transport modeling. The report considers the sources of such data, both current and potential, and how they might be used by existing and future transport models.
Traditionally computer based models have been used to look at the impact of operational decisions, such as changes in traffic management or additional public transport provision. The existence of potentially large quantities of automatically collected data may enable the development of transport maudlin techniques which are capable of being used to investigate the impact of other policy decisions, such as the increase in fuel taxation or the introduction of road pricing.
As a result, such models will need to be multi-modal in nature, placing high demands on the quantity and quality of transport data which will be required to calibrate such models. Traditional methods of data collection, eg travel diaries and roadside surveys, can be expensive to conduct, obtain limited amounts of data and can be inaccurate and open to subjective analysis.
What is required are alternative sources of data, which can be automatically collected and do not involve extensive manual intervention. Such sources should be viable, cost effective and produce large quantities of accurate, consistent data that would yield useful inputs to transport models involving both passengers and freight.
The study consisted of three phases. These were:
- identify contacts/data sources;
- for each contact/source of data, obtain the type of information available, its cost and quality and explore institutional issues such as access to the data;
- assess the suitability of the data for maudlin purposes and the data requirements of the transport maudlin community.
Contacts and potential data sources were identified from a comprehensive literature review, World Wide Web search, e-mail search and from Smith’s own extensive contacts within the transport industry and academia. Using these contacts, information was gathered about current and potential data collection systems and various members of the transport maudlin community were approached to discuss the potential uses of the data sources and general data requirements.
In all 365 abstracts were examined, 40 papers obtained and 75 people/organization were contacted by telephone/fax.
The report classifies data sources as point data (eg loops, infrared sensors), section data (eg number plate matching techniques, beacon systems), origin-destination and route data (eg vehicle tracking, automated surveys) and pedestrian and passenger data (eg in-flow and out-flow of passengers at stations).
For each classification, current and potential installations are discussed, both in the UK and abroad, and several cost examples are given, including installation and communications costs.
The study found that, in general, there are few, if any, currently available national sources of traffic and transport data that are both suitable for the transport modeller and that are automatically collected. However, there are several proposed systems that may lead to the implementation of several national systems.
The most widespread provision is in the area of point sources of data, concentrated on inter-urban motorways and trunk roads. Data collection in urban areas is widespread, but more fragmented, while on inter-urban non-trunk roads, it is very poor. The report reviews current installations and also the potential of the Highways Agency’s proposed national MIDAS system and the Department of Transport’s Urban Traffic Management and Control initiative.
Currently, there is little by way of sectional data, but the report reviews the potential of various technologies, including number plate matching, infra-red beacons and enhanced inductive loop techniques. Also discussed are vehicle tracking systems for public transport and freight vehicles. Costed examples of national implementations of several potential implementations are presented.
Traditionally, origin-destination data have been collected by roadside interviews and travel surveys. The report considers recent advances in computer-based survey techniques and the possibility of using vehicle tracking systems to provide origin-destination and route choice data. The potential of autonomous route-guidance systems for providing vehicle probe data is discussed as well as the likely impact of a national motorway tolling scheme.
Finally, data sources for pedestrian movements and passenger flows are considered. It was found that there are very few sources of data in this area and the report reviews the potential for new sources of data, considering the likely impact privatization of public transport service providers.
By considering cost examples of various current and proposed systems, it was found that the cost of the provision of wide spread automatic data collection cannot be justified on the basis of providing traffic and transport modeling data alone. However, transport authorities are increasingly justifying the cost of such systems in terms of the savings made to travelers as a result of reduced congestion and better incident management. There is also direct evidence within the UK that motorists are prepared to pay for traffic information.
It was also found that fleet managers, both in the public and private sector, are also prepared to pay for vehicle tracking systems, with the cost of such systems justified in the savings made through increased fleet efficiency. Again, these systems’ primary function will be to provide real-time data for fleet management and not to store data.
It is therefore highly likely that advanced traffic data collection systems will be installed in the UK within the next five to ten years. However, these systems’ primary function will be to provide real-time data for traffic management, not to provide long term data storage for modeling. Therefore it is important that those wishing to obtain data for modeling (and monitoring) purposes should approach those responsible for existing and potential systems at an early stage. This should ensure that data storage is made an integral part of the data collection system, that the correct data are collected and that any additional cost arising from data storage is minimize by the correct choice of hardware and communications infrastructure.
Looking at the uses of data for modeling, it was found that there is a wide range of requirements within the transport modeling community. Often model development has been limited by the amount and quality of data available. Generally, rather than concentrate on how the model might be developed if improved sources of data were available, research effort has been focused on making the most of existing data sources. The report identifies several areas where research is required into the potential uses of new sources of traffic and transport data, especially those from vehicle tracking systems.
The report also identifies three main areas where there is a common demand for traffic data:
- provision of detailed validation data for microscopic and assignment models, with an emphasis on long term data allowing the examination of daily and weekly variation;
- provision of detailed and accurate origin-destination data for all models;
- provision of journey-time/flow data for assignment models.
The report provides a summary of the main existing and potential sources of automated data and makes recommendations as to which areas should be considered in more detail.
The full text of this report is available for download at the following address:
Last Modified: March 20, 2007
Maintained by: traffic@nmsu.edu
[HOME] [ABOUT US] [TECH. ASSISTANCE] [EVENTS] [PRODUCT INFO.] [PUBLICATIONS] [CONTACTS] [LINKS]