Proceedings of 1st international symposium on From digital footprints to social and community intelligence
| Hi-index | 0.00 |
At the end of 2008 TomTom finalized the roll-out of its real-time traffic information service HD TrafficTM in 4 additional countries as the UK, France, Germany and Switzerland as an enhancement to the Dutch roll-out in 2007. The traffic information service is the next step in TomTom's Pan-European roll-out strategy of HD Traffic. Together with the historic speed profiles generated from billions of statistical information gathered by TomTom portable device the reliability of the routing and behaviour and the Estimated Time of Arrival is greatly improved. The presentation outlines the basic technological background, coverage and examples of the traffic information service as currently available for TomTom users of a GO X40 and ONE XL LIVE device range in the countries mentioned above. Furthermore the concept gathering historic speed profiles out of billions of statistical information which are gathered from the TomTom navigation devices. For the generation of real time traffic information the data fusion concept is explained exploiting millions of probe measurements using either cell phones or GPS enabled devices. The core data sources are based on probe data from cell phone operators in the different countries as well as a large number of GPS probe data from commercial fleets of TomTom Work and the installed base of connected devices of TomTom. The already existing very large installed base of GPS probe vehicles and cell phone handsets of the cooperating telecommunication operators guarantees an enhanced coverage of the service (highways, secondary roads and arterial urban roads) and greatly improves the travel time calculation and the delay time along a planned route or in a traffic jam. In the talk insights about the statistical data analysis of collected feeds is explained. Furthermore there are a few challenges in data storage, data processing and data mining to be met. For the generation of reliable historic speed information per road stretch billions of data sets are used and so far needs to be stored and accessed in a very efficient manner. When talking about more the 700 billion data sets need to be screened during the analysis conventional relational data bases also with spatial engines are not giving the appropriate performance. In this context a special system design is required to gather data e.g. a bounding box query in a pretty fast manner. Another important aspect for the historic speed profile generation is a highly efficient storage of the content on the navigation device. Storage capacity is always limited and so far is it not efficient to store 7 speed profiles (one for each weekday) per road stretch on the device. Just to mention the number of stretches we are talking to provide a European road network is about 50 Million road segment. In order to overcome this storage issue a method for data aggregation and compression of speed profiles is explained where the additional storage need not exceeds 2-3% of the original digital map size (network graph). Finally the presentation outlines how to combine historic and real-time speed information along a planned journey in order to provide best routing with most precise travel time information.