A study of action control method of the humanoid robot using a pulsed neural network
| Project/Area Number |
25870842
|
|---|
| Research Category |
Grant-in-Aid for Young Scientists (B)
|
| Allocation Type | Multi-year Fund |
|---|
| Research Field |
Soft computing
Control engineering/System engineering
|
|---|
| Research Institution | Kanto Gakuin University |
|---|
Principal Investigator |
MOTOKI Makoto 関東学院大学, 理工学部, 准教授 (20440282)
|
|---|
| Research Collaborator |
KOYAMA Hiroaki
SHIMIZU Takumi
CHIBA Akira
LU Jun
|
|---|
| Project Period (FY) |
2013-04-01 – 2015-03-31
|
| Project Status |
Completed (Fiscal Year 2014)
|
| Budget Amount *help |
¥3,510,000 (Direct Cost: ¥2,700,000、Indirect Cost: ¥810,000)
Fiscal Year 2014: ¥650,000 (Direct Cost: ¥500,000、Indirect Cost: ¥150,000)
Fiscal Year 2013: ¥2,860,000 (Direct Cost: ¥2,200,000、Indirect Cost: ¥660,000)
|
|---|
| Keywords | パルスニューラルネットワ / 人型ロボット / 転倒回避 / パルスニューラルネットワ ーク / パルスニューラルネットワーク |
|---|
| Outline of Final Research Achievements |
I have obtained a kids-size biped walking robot: "RIC-90 (RT Co., Ltd) ". It is equipped with various sensors and a note PC so that the robot is able to act autonomously. In addition, using a small (the height is about 35cm) humanoid robot, I have improved the method that a biped walking robot avoids falling while walking by a neural network and a reinforcement learning. The neural network of the method is used a control of the attitude in real time. The reinforcement learning was used that build a rule in order to act in appropriate timings. Result of experiment elucidate that the success rate of the avoidance of falling improve than the conventional method. I will apply the improved method to kids-size biped walking robot in future. In addition, I will verify the validity of the method, using an actual machine with various ground including the sandy area and lawn area.
|
Report
(3 results)
Research Products
(8 results)
