Within this paper it was given a novel approach on how quality of RTTI (real time traffic information) services can be assessed for use in future active safety applications. Our research is based on the fictional application TJA consuming the feature average travel speed coming from a RTTI Service; to detect downstream traffic situations that are not yet perceivable by vehicle sensors. To mitigate risk for the passengers while at the same time maintaining traveling comfort, TJA (Traffic Jam Assist) needs to grasp the quality of individual features. We show how quality can be measured in an online indirect qualification process by using the framework of MBB (model based benchmarking). Our work is based on simulated data that has been generated for a fictional test site. To measure quality, MBB performs an independent state estimation of the feature average travel speed with a Kalman Filter Bank. By applying spatial segmentation an top of a digital map database. we achieve a feature resolution of under 150 meters. Deviation analysis between the estimated state and ground truth information shows that error of state estimation decreases with increasing probe penetration rates within the observed test region. The import of simulated RTTI Services having different error characteristics into the framework allows MBB to measure value error and temporal shift between RTTI services and the internal state estimation. Based on theses quality indices, RTTI services can be differentiated which in return provides TJA additional value and allows for a more precise interpretation of features from the RTTI service. We are currently extending the algorithms of MBB for RTTI to deliver higher quality results, and preliminary results are encouraging. In future work we will defeat the requirement for simulation data by real life validation with probe data that is captured by a dedicated test site camera system that stretches multiple kilometers of a highway. Based on this data, we will continue to determine the influence of varying probe penetration rates on AIBB and research how historical data can be used to make state estimation and benchmarking more precise. Furthermore, we will extend MBB to deliver additional error dimensions like continuity and availability of RTTI services.


    Zugriff

    Zugriff über TIB

    Verfügbarkeit in meiner Bibliothek prüfen

    Bestellung bei Subito €


    Exportieren, teilen und zitieren



    Titel :

    Online qualification of real time traffic information using model based benchmarking


    Beteiligte:
    Klar, Tilman (Autor:in) / Sasse, AndreasP (Autor:in) / Hecker, Peter (Autor:in)


    Erscheinungsdatum :

    2012


    Format / Umfang :

    19 Seiten, 5 Bilder, 3 Tabellen, 14 Quellen



    Medientyp :

    Aufsatz (Konferenz)


    Format :

    Print


    Sprache :

    Englisch




    Online qualification of real time traffic information using model based benchmarking

    Klar,T. / Sasse,A. / Hecker,P. et al. | Kraftfahrwesen | 2013


    Online Stochastic Routing Incorporating Real-Time Traffic Information

    Du, Lili / Peeta, Srinivas / Kim, Yong Hoon | Transportation Research Record | 2013



    Real-time online traffic state detection method based on traffic monitoring video

    FU YUCHEN / ZHU FEI / ZHU HAIJUN et al. | Europäisches Patentamt | 2015

    Freier Zugriff

    A Real-Time Traffic Information Model Using GPS-Based Probe Car Data

    Chen, Wan-Jia / Chen, Chi-Hua / Lin, Bon-Yeh et al. | Tema Archiv | 2011