2005 Fiscal Year Final Research Report Summary
A Research of Humanoid Vision System Improving Human-Robot Communication
| Project/Area Number |
16500105
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Perception information processing/Intelligent robotics
|
|---|
| Research Institution | Gifu University |
|---|
Principal Investigator |
YAMAMOTO Kazuhiko Gifu University, Faculty of Engineering, Professor, 工学部, 教授 (60273116)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
KATO Kunihito Gifu University, Faculty of Engineering, Research Associate, 工学部, 助手 (70283281)
|
|---|
| Project Period (FY) |
2004 – 2005
|
| Keywords | Humanoid vision / Humanoid robot / Gesture recognition / Person detection / Man-Machine communication / Eye movement analysis |
|---|
| Research Abstract |
In this research, we proposed "Humanoid Vision". Nodaway, humanoid robot's exterior and features are becoming as similar as human. Then we consider that "the humanoid robot has to have humanoid functions", and eyes for humanoid robot have to be "Humanoid Vision". Humanoid vision is the vision system which is focused on human actions of the robot, and emulation of human beings. We used a humanoid robot "YAMATO" which has two camera eyes to achieve this research goal.
The first year, we constructed an emulation system by using distance measuring, detection of the person and gesture recognition method. We developed their algorithms and introduced on our robot. Our method recognizes gestures of persons standing in front of the robot, and the robot can follow that actions.
The second year, we researched about user friendly interface between human and robot. Robot has to act more real human actions to feel the humanity in human and robot communication. Therefore, in the first step, we studied the human action of tracking an object by the eyes, and that actions are modeled to the robot actions. By using this model, our robot can track an object as similar as human beings.
We think that the human being is optimized for human frameworks, thus the humanoid vision is the best vision system for humanoid robot which has human like.
|
