| Project/Area Number |
16K00767
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Research Field |
Home economics/Human life
|
|---|
| Research Institution | Shonan Institute of Technology |
|---|
Principal Investigator |
Ozaki Fumio 湘南工科大学, 工学部, 教授 (10712474)
|
|---|
| Project Period (FY) |
2016-04-01 – 2019-03-31
|
| Project Status |
Completed (Fiscal Year 2018)
|
| Budget Amount *help |
¥3,250,000 (Direct Cost: ¥2,500,000、Indirect Cost: ¥750,000)
Fiscal Year 2018: ¥910,000 (Direct Cost: ¥700,000、Indirect Cost: ¥210,000)
Fiscal Year 2017: ¥1,040,000 (Direct Cost: ¥800,000、Indirect Cost: ¥240,000)
Fiscal Year 2016: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000)
|
|---|
| Keywords | ロボット / 高齢者生活支援 / 歩容認証 / 人物認識 / 見守り / 見守りロボット / 車椅子認識 / 移動ロボット / 深層学習 / RGB-Dセンサ / 物体検出 / 転倒検出 / 生活支援ロボット / 高齢者 |
|---|
| Outline of Final Research Achievements |
We have been researching on elderly care robots, especially for watching over the elderly. The following technologies have been studied: recognizing the elderly to be watched over by image processing, moving a robot to follow the elderly, and recognizing the elderly by gait recognition.
To recognize the elderly by image processing, we have applied mobilenetSSD algorithm and have found that the algorithm could recognize human even at near distance, at which the skeleton based human recognition did not work. To move a robot to follow the elderly, we have used SLAM (Simultaneous Localization and Mapping) to make the map of a room, and also using the above image processing to recognize human, and to measure the distance between the human and the robot. To recognize the elderly by gait recognition, we have make use of an RGB-D sensor to get the skeleton data of the elderly. We have achieved 91 % accuracy of gaits not only for straight walk but also for free walk, using deep learning.
|
| Academic Significance and Societal Importance of the Research Achievements |
高齢者を見守る技術として人の認識機能の研究を行った。これにより人との距離を保って追従したり、また人の動きを認識することが可能となった。特に人の動きの認識機能を使って歩容認証機能を開発した。歩容認証は大部屋での入退出管理に使え、入退出した人がケアの必要な人（例えば認知症）であるかどうかを認識し、ケアが必要であれば知らせることが出来る。本研究ではRGB-Dセンサを用い、骨格データを進行方向基準に座標変換することで従来難しかった自由に歩いている人の認識においても認識率を91%とすることが出来た。
|
