In this paper, the development of two new CW/CA (Collision Warning/Collision Avoidance) algorithms is reported. We first presented the evaluation results of selected existing CW/CA algorithms by using a large-scale human driving database-the ICC FOT database. A scalable numerical optimization algorithm is then introduced to optimize these algorithms. Among the existing algorithms evaluated and optimized, the JHU-APL algorithm scored the highest but failed to achieve high performance in terms of True Positive (the rate of correctly-detected points among the Threatening set) and Precision (the rate of Threatening points among all warnings). A Grand Finale algorithm, which combines the basis functions of all the CW/CA algorithms produces a much higher performance index, in excess of 55 %. However, the algorithm requires 10-parameter tuning and is less likely to be practical. A simpler algorithm using the lead vehicle acceleration and time-to-impact achieves a similarly high performance despite of its much reduced complexity. The success of both algorithms heavily depends on the lead vehicle acceleration information, which is not readily available. Subsequently, the concept of Electronic Brake Light is discussed, which transmits lead vehicle acceleration information, among other useful information via inter-vehicle communication. We believe the realization of the Electronic Brake Light concept is a critical step for successful development of practical CW/CA systems,


    Access

    Access via TIB

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Design and evaluation of forward collision warning/collision avoidance algorithms


    Additional title:

    Design und Auswertung von Algorithmen von aktiven Warnsystemen zur Vermeidung von Auffahrunfällen


    Contributors:
    Lee, Kangwon (author) / Peng, Huei (author)

    Published in:

    Publication date :

    2005


    Size :

    9 Seiten, 7 Bilder, 11 Tabellen, 23 Quellen



    Type of media :

    Article (Journal)


    Type of material :

    Print


    Language :

    English