Skip to main content

Creating a Data Mart for Floating Car Data

Final Report Summary - TRACK AND TRADE (Creating a Data Mart for Floating Car Data)

The TRACK&TRADE project aimed at exploiting the untapped resource of floating car data (FCD) as a cost-efficient means for accurate assessment of traffic conditions. To address the shortcomings of today's FCD-based systems, TRACK&TRADE advocated a web-based approach for the collection and exploitation of the large pool of existing FCD sources.

The project's main goals were:
- web-based platform that simplifies the collection of FCD;
- core technology for the processing and manipulation of FCD and derived travel time data;
- value-added services based on travel times;
- business model for the exploitation of the project results.

The TRACK&TRADE data model was developed based on data format surveys and requirements specifications towards the set of attributes required for a universal data collection format. The extensible XML-based model captures not only various types of FCD sources, but also traffic count information and related data such as weather information and TMC data. A sophisticated map-matching algorithm was developed that relates FCD vehicle tracking data to the road network. The algorithm tries to find a path in the road network matching most closely the vehicle trajectory. This curve matching algorithm uses the Fréchet distance function to determine the quality of a match.

The project has the following major achievements:
- TRACK&TRADE data model - based on an extensive survey of existing data formats and standards relevant to FCD, traffic information and related information, an extensible XML-based data model has been developed.
- Data collection and mediation framework - based on different scenarios, a flexible data collection framework based on web services was developed. This software should simplify the process of connecting new data sources to the data mart. A mediation framework was developed that allows for the mapping of any kinds of data source to the TNT data model.
- Map-matching algorithm - a software framework for map-matching has been developed. Using this algorithm, FCD is related to a road network and travel times are derived.
- Data management - a spatiotemporal data warehouse schema has been defined that allows for the aggregation of historic travel times to provide for future estimates. Main-memory data structures have been developed to buffer live FCD data streams.
- Data mart - Graphic user interface (GUI) that allows access to the date and respective services. Two versions have been implemented (i) a web-based solution based on open-source software and (ii) a proprietary, more powerful solutions as part of the Cruiser platform.
- Value-added services - five products have been implemented and tested. These products are a direct result of the business model and exploitation strategy.
- Business model - a business model that includes five specific products, along with licensing agreement and billings models has been developed.

The project was structured into the following work packages (WPs):

WP1: Data collection and integration
The data collection technology that was created is based on web services that can be invoked by a data provider to upload its data. As an alternative, to connect legacy data sources, a poller was created that actively downloads the data from a data repository, e.g. ftp sites, databases, etc., to the data mart. A TRACK&TRADE data model was defined. This extensible model covers not only FCD but a variety of traffic-related data sources that we might want to collect in the data mart. The data model is based on the analysis of a variety of existing standards and models. It is used within the data mart to communicate the collected data.

WP2: Data manipulation.
A Data Store has been developed to manage (i) collected data and (ii) derived travel times. The data store is implemented by means of an Oracle database. A specific data model and storage schema has been defined to provide for efficient manipulation of travel times and the computation of estimated travel times.

WP3: FCD data mart.
The data mart integrates the data collection and data management developed in WP1 and WP2, respectively. It further provides the glue to the value-added services as it will host the travel time maps (cf. business model TTA.maps TTA.prints). The data mart and with it the software integration was a major focus of work. The results were an application built on top of the Cruiser platform and also a prototype implementation of a web-based version that relies on open-source components.

WP4: Services
The services that have been implemented are:
- travel time assessment and prediction (TTA.raw and TTA.data) - services that give access to the collected travel times in its raw and aggregated form;
- travel time maps (TTA.maps and TTA.print) - services that visualise relative travel times and thus traffic conditions in a map;
- routing (TTA.route) - a routing service that relies on the travel time database as derived in the project.
All services have been implemented and tested.

WP5: Exploitation and business model
A viable business model, billing models and licensing agreements in connection with five products has been defined. The business model is centred around five core products, which in turn affected the development of the value-added services of WP4.

WP6: Project management
The corner points here were the following:
- communication and co-operation: six project meetings; frequent communication over email and phone between involved project partners; use of Groove project management software; use of subversion site for software management; creation of template documents for technical communication and deliverables.
- Dissemination: project website, found at: http://www.trackandtrade.org; project logo; project brochure and poster.
- Other: contact to other consortia and research groups regarding new proposals and collaborations.

The core contributions from a technical point of view were:
(i) 'travel-time derivation technology' that allows for simple collection of FCD from various resources; utilises an efficient, both in terms of accuracy and running time, map-matching algorithm that derives travel times from tracking data; introduces data management technology that allows for the simple derivation of speed profiles.
(ii) value-added services that include a novel shortest-path algorithm that utilises dynamic travel times to provide superior routing solutions; visualisation techniques to provide a traffic assessment based on collected travel times.
(iii) a business plan that includes five specific products; associated business models and licensing models.

No actual problems were encountered that would cause a project delay and were not resolved in time. The TRACK&TRADE project created technology that allows foremost the SMEs partners to utilise a dormant resource in various and flexible ways. Overall, the encouraging project results have created an intensive collaboration between the project partners that is continuing beyond the TRACK&TRADE project.

Related documents