The need for so called Track-Before-Detect (TBD) solutions arised from the field of radar aerial surveillance in military applications with stealthy targets. The amplitudes of these targets are so weak, that a detection can only be accomplished by recognizing local deviations from background clutter statistics over time. This is not the case in automotive applications where other road users do not intentionally hide. However, technical and economic limitations of automotive environment sensors lead to three types of uncertainty. They are the uncertainty of the dynamic state such as position and velocity, due to measurement errors caused i.e. by discretisation effects, the uncertainty of object existence induced by false and missing detections, and the uncertainty of data association between multiple real objects and multiple measurements. Using TBD concepts provides a clean mathematical foundation of handling these uncertainties simultaneously as there are non-negligible dependencies. This paper describes the basic concepts of JIPDA filtering, a very tractable TBD algorithm. In this contribution a novel automotive application of Joint Integrated Probabilistic Data Association (JIPDA) is presented for a sensor fusion setup with lidar and video. The system is intended for next generation smart Adaptive Cruise Control (ACC) systems. JIPDA combines the concepts of IPDA and JPDA. IPDA incorporates existence estimation into ordinary position and velocity tracking. JPDA avoids sub optimal measurement to track associations by performing weighted filter innovations with multiple ambiguous association pairs.


    Access

    Access via TIB

    Check availability in my library


    Export, share and cite



    Title :

    Simultaneous processing of multitarget state measurements and object individual sensory existence evidence with the joint integrated probabilistic data association filter


    Additional title:

    Simultane Bearbeitung von Mehrziel-Zustandsmessungen und individueller Objekterkennung mit Sensoren unter Anwendung eines gemeinsamen, integralen, wahrscheinlichkeitsgestützten Datenzuordnungsfilters


    Contributors:


    Publication date :

    2008


    Size :

    6 Seiten, 10 Bilder, 3 Tabellen, 10 Quellen


    Type of media :

    Conference paper


    Type of material :

    Print


    Language :

    English




    Iterative joint integrated probabilistic data association for multitarget tracking

    Taek Lyul Song, / Hyoung Won Kim, / Musicki, Darko | IEEE | 2015



    Multitarget Sensor Resolution Model and Joint Probabilistic Data Association

    Svensson, D. / Ulmke, M. / Hammarstrand, L. | IEEE | 2012



    De-cluttering with Integrated Probabilistic Data Association for Multisensor Multitarget ACC Vehicle Tracking

    Maehlisch, M. / Ritter, W. / Dietmayer, K. et al. | British Library Conference Proceedings | 2007