• DocumentCode
    2803933
  • Title

    Singularisation of parcels with a sensor-based robot-system

  • Author

    Baerveldt, Albert-Jan

  • Author_Institution
    Inst. of Robotics, Swiss Federal Inst. of Technol., Zurich, Switzerland
  • fYear
    1991
  • fDate
    3-5 Nov 1991
  • Firstpage
    523
  • Abstract
    Presents a highly robust robot-based system for the singularisation of parcels. Singularisation can be described as the removal of a single part from a stream or pile of parts. The system uses range imaging and optical short-range sensors, mounted in the fingertips of the gripper, to guide the robot. It was the intention that by using multiple sensors, i.e. sensor fusion, the image processing could be kept simple in order to achieve real-time performance. The resulting 3-D vision-system is able to deal with almost every pile configuration effectively as its shortcomings are corrected by the optical fingertip-sensors. A new approach of active perception is applied to the unsolvable, rather exceptional, pile-configurations: the robot is told to change the pile with a bulldozing-like operation. This results nearly always in a solvable pile-configuration. Through the use of the described sensor fusion and active perception an autonomous and very robust singularisation-system has been developed
  • Keywords
    computer vision; postal services; robots; 3-D vision-system; active perception; bulldozing-like operation; computer vision; gripper; optical short-range sensors; postal services; range imaging; sensor fusion; sensor-based robot-system; Grippers; Image processing; Image sensors; Optical imaging; Optical sensors; Robot sensing systems; Robustness; Sensor fusion; Sensor systems; Streaming media;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Intelligent Robots and Systems '91. 'Intelligence for Mechanical Systems, Proceedings IROS '91. IEEE/RSJ International Workshop on
  • Conference_Location
    Osaka
  • Print_ISBN
    0-7803-0067-X
  • Type

    conf

  • DOI
    10.1109/IROS.1991.174525
  • Filename
    174525