Incorporating the information and uncertainties of loop-detector and floating car data in freeway traffic state estimation

Traffic management measures may be implemented to make use of the road traffic network as efficient as possible. Therefore, it is an important field of research for traffic researchers. Potential traffic management measures first have to be proven theoretically. However, before the decision is made to implement these measures in real-life, their effectiveness has to be tested. Road traffic involves many different actors. For instance, different authorities are responsible for different parts of the road traffic network and traffic data is available at multiple, both public and private, organizations. A collaboration between these actors yields advantages. Not only for the actors, but also for the road users. In Amsterdam, the Netherlands, a unique project was setup in which different public and private organizations work together to test traffic management systems in a field operational test. The FOT Amsterdam consist of multiple phases of which the first has been successfully completed. The traffic management system tested in this first phase aims to prevent the congestion at the freeway as this negatively affects the maximum flow (capacity). In short, the traffic management system observes the traffic conditions on the roads, evaluates if breakdowns are likely to occur and if needed takes control action. This research focuses on a single (monitoring) component of the entire system, namely the freeway traffic state estimator. As the name states, this component estimates the traffic conditions on the freeway using the available data. In the first phase of the FOT Amsterdam the estimator only Loop-Detector Data (LDD) is utilized. LDD is gathered by sensors (loops) which have fixed road locations and is therefore denoted as Eulerian sensing data. At the location where loop-detectors are installed, for each individual lane, the one-minute aggregated speed and flow are available. In the second phase of the FOT Amsterdam a new data-type becomes available, namely Floating Car Data (FCD). Potential sources of FCD are mobile phones and navigation systems. In contrast to Eulerian sensors, which have a fixed road location, these sensors follow individual vehicles and FCD is therefore denoted as Lagrangian sensing data. The potential availability of this data is a consequence of the cooperative nature of the FOT Amsterdam. Potential FCD-providers are private companies. In a meeting with FCD-providers the characteristics of this data-type were discussed. It is expected that the individual vehicle speed and location may be available for up to 20 % of the road users. The …