• DocumentCode
    1721159
  • Title

    Spatial and temporal coarse to fine structuring method of distributed tactile sensors based on changing velocity of the sensation

  • Author

    Yoshikai, Tomoaki ; Kobayashi, Kazuya ; Inaba, Masayuki

  • Author_Institution
    Dept. of Mechano-Inf., Univ. of Tokyo, Tokyo, Japan
  • fYear
    2011
  • Firstpage
    1427
  • Lastpage
    1432
  • Abstract
    In order to achieve robots working around humans, whole-body tactile sensation is one of important keys. One problem for achieving such whole-body tactile sensation is data processing for enormous amount of data from distributed sensor elements. Some kinds of sensor data structuring method is necessary for efficient processing of such tactile information. From a view point of whole system architecture of a robot, all of tactile sensor data is not always necessary in the higher level computation. Sensor data for the parts on which given stimuli are changing frequently should be acquired in detail and dense data should be sent to the higher level computation parts. On the other hand, sensor data for the parts on which given stimuli are almost same should be summarized temporally and spatially and sparse data should be sent to the higher level. So far, not many researches are focusing on the structuring method of tactile sensor data. Only a few groups start trials to acquire structured data for whole body tactile information. Basically correlating relationship among tactile sensors are acquired offline in their methods. It is difficult to construct those structured tactile data online for realtime operation of robot´s behavior. Therefore, a new method to conduct online structuring method for tactile sensor data based on the changing velocities of sensor elements is proposed in this paper. Proposed method consists of two stages. At first, tactile sensor data is temporally structured at the level of local processor where A/D conversion is done. There, update frequencies for each sensor elements are automatically changed based on the changing velocities of each sensors. Next, temporally structured tactile sensor data is structured spatially: sensor elements are clustered based on the positional information and update frequencies. Furthermore, experiments using prototype distributed sensor boards are done for verifying proposed method.
  • Keywords
    analogue-digital conversion; data structures; distributed sensors; human-robot interaction; prototypes; tactile sensors; A/D conversion; data processing; distributed sensor elements; distributed tactile sensors; prototype distributed sensor boards; realtime operation; robot behavior; sensation velocity; sensor data structuring method; tactile information; tactile sensor data; whole-body tactile sensation; Computers; Distributed databases; Force sensors; Prototypes; Tactile sensors;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Robotics and Biomimetics (ROBIO), 2011 IEEE International Conference on
  • Conference_Location
    Karon Beach, Phuket
  • Print_ISBN
    978-1-4577-2136-6
  • Type

    conf

  • DOI
    10.1109/ROBIO.2011.6181490
  • Filename
    6181490