1996 Fiscal Year Final Research Report Summary
Development of an Intelligent Humanoid Robot
| Project/Area Number |
06555069
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (A)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 試験 |
|---|
| Research Field |
Intelligent mechanics/Mechanical systems
|
|---|
| Research Institution | The University of Tokyo |
|---|
Principal Investigator |
INOUE Hirochika The University of Tokyo, Dept. of Mechano-Informatics, Professor, 大学院・工学系研究科, 教授 (50111464)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
武田 周 (株)小松製作所, 研究部, 研究員
INABA Masayuki The University of Tokyo, Dept. of Mechano-Informatics, Associate Professor, 大学院・工学系研究科, 助教授 (50184726)
TAKEDA Shyu KOMATSU Co. , Research Center, Researcher
|
|---|
| Project Period (FY) |
1994 – 1996
|
| Keywords | Humanoid / Robotics / Emergence of skill / Learning of movement / Vision system / Robot hand / Grasping unknown objects |
|---|
| Research Abstract |
The research has been done on : 1. design and development of the humanoid Saika, 2. design and development of 3-fingered robot hand with distributed touch sensors, 3. experiment on grasping unknown objects, 4. acquisition of inverse dynamic model of Saika arms, 5. experiment on catching a ball, 6. experiment on dribbling a ball. The results of the research are summarized as follows :
1. A light-weight (8 [kg]), low developing-cost (eight thousand US dollar) humanoid which has 2-DOF (degrees of freedom) neck, 5-DOF dual arm, 2-DOF binocular camera arrangement, two ears and a torso was developed.
2. A 8-DOF 3-fingered robot hand which is attachable to the developed humanoid is designed and developed. The hand has a controller inside the forearm and has eighty-seven touch sensors distributed on the surface of the palm and fingers.
3. A strategy for grasping unknown objects by groping is proposed and experimented. Experiments are performed using the developed robot hand. The results demonstrate the capability of the strategy.
4. Experiments on acquisition of inverse dynamic model are performed using "feedback error learning method."
5. Catching-a-bell behavior is implemented in the developed humanoid. The humanoid has a capability of catching a dropping ball and a thrown ball.
|
