PTV Group - Logistics Forum

Input Data / Floating Car Data (aka FCD)

TomTom has been making real-time traffic information based on FCD in United States since 2010. Over the past 12 years, our systems evolved further, and the volume of input data has increased significantly. A quick overview of the data sources we use to produce traffic data in the United States:

• Probe data: The most important data stream is GPS data received from various vehicle fleets, including our own TomTom navigation, Automotive systems (including brands named on our website) and smartphone applications (including TomTom and other customers mentioned on our website).
• Through various sources, TomTom also collects information about events, such as roadworks and road closures. For many States in the US, we source data from various state DOT’s and commercial organizations running Traffic centers on behalf of the State DOT. We also collect data via our own tool, the Road Event Reporter. Certified users and TomTom moderators can submit events around accidents, closed roads and for example roadworks.
• Through the TomTom AmiGO smartphone app, we collect crowd sourced data. Via the app users can report accidents, hazards, roadworks, and other types of messages.
• Via Sensor Derived Observation (received via cameras on cars) we receive data from roadwork signs that are observed along the road. After Map matching and validation, we create and adjust roadwork locations based on this data stream.
• The last source is historical Speed Profiles. This data contains information about the freeflow speed and the expected congestion at any time of the day in the week based on a 2-year rolling archive. This historical data is used to validate current data and fill gaps when no real-time data is available.

The TomTom GPS fleet is extremely stable, mainly from our contracts in the automotive space. Contracts with Automotive partners are often for the life of a model and there for several years. As a result, we can offer a very stable and growing fleet as a base. TomTom can therefore rely on a broad base of GPS data providers and is not dependent on a limited number of apps.

Data Fusion / Data fusion and process steps

Next to a module that constantly receives all input data we have data fusion process module. In the data fusion, hundreds of validation and data fusion steps are repeated every 30 seconds. This module is, like the other modules, fully automated. Below is an overview of the most important steps in the data fusion process.
1. The map-matching of GPS data is done by linking the GPS-data to the internal TomTom map (DSEG) and where outliers are filtered out. (e.g., extreme speeds or data generated by errors in the GPS signal or GPS-chip);
2. Real-time speed per segment. In this process, all input sources are used to calculate the current real-time speed on the road-element. Speed values are matched with each other and with historical speeds in order to determine the current actual speed. In addition to the speed value, a quality indication value is also calculated. This value is based on the number of measurements and the quality of the measurements.
3. The determination of traffic jam locations, with their delay time and the precise head and tail position of the traffic jam.
4. The prediction of the life of each traffic jam. This is done based on congestion build-up, historical profiles and, for example, congestion (e.g., work).
5. The validation of the road closures based on GPS data. Road segments that are reported as closed but still contain (large) volumes of GPS data are adjusted or removed, so TomTom only shows road closures if the road stretch is closed.
6. Road closure detection. Road segments without GPS data, while we normally expect high volumes of GPS data, are considered as closed, until GPS-positions are received again.

Along with the above steps there are hundreds of other steps that are repeated every 30 seconds to generate our real time traffic information. In Figure 1, the data fusion is explained.