Providing Travel-Time Predictions on the Internet 

Information technology is set to change road users’ perception of roads and traffic. The collection of real-time data from a combination of embedded sensors in roads, as well as cameras mounted at strategic locations, and the instantaneous transfer of that information back to traffic control centers mean that up-to-date network wide information is available for decision making. The dissemination of up-to-date traffic information to road users via a variety of means including radio broadcasts, mounted electronic displays on the roads, as well as through the Internet, means that road users will now have the right information to make their travel route choices.

The introduction of microscopic traffic simulation using real-time inputs forms the basis of predicting what traffic conditions will be in the very near future. Predictions of traffic conditions will be useful to a wide audience and for a variety of applications including travel time estimation for the general public, route guidance for motorists, route planning for logistics operations, electronic road pricing for regulatory agencies, emergency response planning etc. The analogy with weather prediction has prompted some observers who have coined the term ‘traffic weather’ prediction because of the close analogy to climate weather prediction. It is not inconceivable that, in the future, traffic and weather predictions will be presented together in travel advisories and news bulletins. Traffic simulation model is at the heart of a system for ‘traffic weather’ prediction.

The system consists of five modules (see Figure 1) - data-collection, data-preparation, simulation, database and web-server. Data in the form of vehicle counts at loop detectors and control data from a centralized signal control center are collected at regular time intervals and cached in the data-collection module. The data necessary for running the simulation are prepared in the required format by the data-preparation module. Presently, the input used for simulation is based on simulated data derived from historical data. The simulation module hosts the PARAMICS program along with the other routines used to control the simulation and information generation processes. The information is stored in the database module, from which the web-server module can dynamically extract the required information based on interactive queries based on a mapping interface. The web-server module hosts programs that can receive user requests, query the information from the database and present the information to the user in a readily understandable format on the client browser. The modular structure emphasizes the clean separation of functions in the system, and allows improvements to be made in individual modules without affecting the others. For example, in the future, the use of loop detectors for obtaining traffic counts can be replaced by the use of video cameras and image processing to obtain the same information.

Figure 1: System architecture.

As with weather reports and predictions, it is appealing to present traffic prediction information through the use of maps. The user interacts with the traffic-weather data hub through a map-based interface (see Figure 2). The interface handles common user view interactions like zoom, pan, feature selection and information query, as well as other more map specific interactions like geographical position determination, distance and area measurement, and corridor and radial area buffer selection. The mapping interface is implemented using a state-of-the-art web mapping software from Autodesk called Mapguide. Traffic report generation is implemented using ColdFusion.

Figure 2: Web interface showing travel time and route guidance information.

It is likely that ‘traffic weather’ information will be delivered to a variety of Internet-enabled devices and applications in ITS. The approach taken by the implementation of the traffic weather data hub is to use XML (eXtensible Markup Language) as the common information representation language between different applications. XSL (eXtensible Style Language) and XSLT (XSL Templates) are employed to perform on demand translation between XML and formats like WML (Wireless Markup Language), GML (Geographic Markup Language) and SVG (Scalable Vector Graphics). SVG is notable as it delivers crisp vector-based graphics at all ‘zoom’ levels unlike the present raster-based formats like GIF and JPEG which degrade into block-like patterns if the level of zoom is too high.

Research and development work to realise the concepts that make up the ‘traffic weather’ data hub was funded by a university research project grant under the auspices of the multi-disciplinary Laboratory for Concurrent Engineering and Logistics (LCEL).