Kraft: An Autonomous Robot Manipulation System Based on Geometrical Modeling and Processing
Hirohisa Hirukawa*, Yuuki Inoue**, Toshihide Yoshimura**,
Shinzo Kitamura***, Satoshi Sakakibara****, Makoto Hitomi*****
and Kazutoshi Sumiya*****
*Electrotechnical Laboratory, 1-1-4, Umezono, Tsukuba, lbaraki 305, Japan
**Information Processing Center, Kobe University, Rokkaido, Kobe, Hyogo 657, Japan
***Faculty of Engineering, Kobe University, Rokkaido, Kobe, Hyogo 657, Japan
****Nippondenso Co., Ltd., 1-1, Shouwa-machi, Kariya, Aichi 448, Japan
*****Matsushita Electric Industrial Co., Ltd., 1006, Kadoma, Osaka 571, Japan
This paper describes an autonomous robot manipulation system based on geometrical modeling and processing. We call it Kraft (Kobe Robot Arm Free from Teaching). Kraft plans the manipulator motion by the following sequence. Firstly, the model of the working environment is obtained interactively by a range finder and the geometric model of it. Secondly, the motion is planned from the model by a collision avoidance algorithm and an automatic generation method of the grasping position. Thirdly, the planned motion is checked by an offline simulator using graphic animation and collision detection by numerical calculation. Finally, the obtained motion is executed by a manipulator with 7 d.o.f. The burden of the software implementation has been reduced, because Kraft has a solid modeler as its kernel. It is confirmed that Kraft could plan the “pick up” motion autonomously in the working environment with relatively simple obstacles, and that not the preciseness of the model of the working environment but the consistency of it is crucial for the following motion planning.
This article is published under a Creative Commons Attribution-NoDerivatives 4.0 Internationa License.