2002 Fiscal Year Final Research Report Summary
Study on a graspless Pick-and-Place Task Using a Palm-like End effector
| Project/Area Number |
13650273
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Intelligent mechanics/Mechanical systems
|
|---|
| Research Institution | The University of Electro-Communications |
|---|
Principal Investigator |
TAKASE Kunikatsu The University of Electro-Communications, Graduate School of Information Systems, Professor, 大学院・情報システム学研究科, 教授 (10270927)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
HADA Yoshiro The University of Electro-Communications, Graduate School of Information Systems, Research Associate, 大学院・情報システム学研究科, 教授 (90313428)
|
|---|
| Project Period (FY) |
2001 – 2002
|
| Keywords | graspless manipulation / dynamic manipulation / pick and place / skill / visual feedback / palm-like end effector |
|---|
| Research Abstract |
In a class of manipulation task, an object must be manipulated without grasping. In such case, human can handle the object skillfully by rolling over, sliding, and/or tossing up it on his/her hand. The ability is referred to as dynamic manipulation skill. For lack of this kind of skill, the function of conventional robot is very limited and consequently its application field is limited. To overcome the situation, we try to realize such skills for a robot.
We implemented a set of manipulation skills for a pick-and-place like operation without grasping an object. It is a dynamic manipulation on a palm. To make dynamic manipulation stable and reliable, it is necessary to keep error not to accumulated. We introduced a visual feedback to avoid the accumulation of error and realized stable and reliable dynamic manipulation.
The scheme makes it possible to manipulate a fragile and big-sized object which have been difficult to deal with by conventional robot system. Dynamic skill primitives for a pick-and-place task are implemented and experimented successfully, verifying the feasibility and effectiveness of the approach. Using a vision, current pose of the object is measured in each motion and next movement will be regulated by the information. In this way, the accumulation of error of the manipulation can be prevented.
|
Research Products
(17 results)
