This paper presents a general approach for recognition of driving maneuvers in advanced driver assistance systems (ADAS). Such systems often rely on the identification of driving maneuvers (overtaking, left turn at intersections, etc.) to improve the prediction of potential collisions or to trigger appropriate support for the driver. The proposed maneuver recognition approach combines a fuzzy rule base to model basic maneuver elements and probabilistic finite-state machines to capture all possible sequences of basic elements that constitute a driving maneuver. The proposed method is specifically tailored to ADAS requirements because of its low computational complexity, its flexibility and its straight-forward design based on easily comprehensible logical rules. In addition, we propose a suitable training method to optimize the fuzzy rule base. Our approach is evaluated on the recognition of turn maneuvers. Experiments on real data from a test vehicle demonstrate the feasibility of the proposed method.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Maneuver recognition using probabilistic finite-state machines and fuzzy logic


    Contributors:


    Publication date :

    2010-06-01


    Size :

    926025 byte





    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English



    Maneuver Recognition Using Probabilistic Finite-State Machines and Fuzzy Logic, pp. 65-70

    Hulnhagen, T. / Dengler, I. / Tamke, A. et al. | British Library Conference Proceedings | 2010


    Probabilistic Maneuver Recognition in Traffic Scenarios

    Firl, Jonas | TIBKAT | 2014

    Free access

    Probabilistic Maneuver Recognition in Traffic Scenarios

    Firl, Jonas | Catalogue medicine | 2014