The automation of handling tasks has been an important scientific topic since the development of the first industrial robots. The first step in the chain of scientific challenges to be solved is the automatic grasping of objects. One of the most famous examples in this context is the well known ”bin-picking” problem. To pick up objects, scrambled in a box is an easy task for humans, but its automation is very complex. Besides the localization of the object, meaning the estimation of the object’s pose (orientation and position), it has to be ensured that a collision free path can be found to safely grasp the objects. For over 50 years, researchers have published approaches towards generic solutions to this problem, but unfortunately no industry applicable, generic system has been developed yet. In this thesis, three different approaches to solve the bin-picking problem are described. More precisely, different solutions to the pose estimation problem are introduced, each paired with additional functionalities to complete it for application in a bin-picking station. It is described, how modern sensors can be used for efficient bin-picking as well as how classic sensor concepts can be applied for novel bin-picking techniques. Three complete systems are described and compared. First, 3D point clouds, generated using a laser scanner, are used as basis. Employing the known Random Sample Matching algorithm and modifications of it, paired with a very efficient depth map based collision avoidance mechanism results in a very robust bin-picking approach. In the second approach, all computations are done on depth maps. This allows the use of 2D image analysis techniques to fulfill the tasks and results in real time data analysis. Combined with force/torque and acceleration sensors, a near time optimal bin-picking system emerges. As a third option, surface normal maps are employed as a basis for pose estimation. In contrast to known approaches, the normal maps are not used for 3D data computation but directly for the object ...


    Zugriff

    Download


    Exportieren, teilen und zitieren



    Titel :

    Bin-Picking - New Approaches for a Classical Problem ; Griff-in-die-Kiste - Neue Ansätze für ein klassisches Problem


    Beteiligte:

    Erscheinungsdatum :

    2015-07-10



    Medientyp :

    Hochschulschrift


    Format :

    Elektronische Ressource


    Sprache :

    Englisch



    Klassifikation :





    Gekuehlte Kiste - Batteriekuehlung

    Wolf,A. | Kraftfahrwesen | 2011


    Beziehungs Kiste. Opel Agila

    Priemer,B. / Opel,Ruesselsheim,DE | Kraftfahrwesen | 1999