Broken robots learn to adapt and recover like animals: study
-
해당 공공저작물은 외부사이트에서 보유하고 있는 저작물로써, 원문보기 버튼 클릭 시 외부사이트로 이동됩니다. 외부사이트의 문제로 인하여 공공저작물로 연결이 되지 않는 경우에는 사이트 바로가기 를 클릭하여 이동해주시기 바랍니다.
- 저작물명
- Broken robots learn to adapt and recover like animals: study
- 저작(권)자
-
저작자 미상 (저작물 2267374 건)
- 출처
-
- 이용조건
-
- 공표년도
-
- 창작년도
- 2015-06-08
- 분류(장르)
- 영상
UCI 로고
- 요약정보
-
When this robot breaks a leg‚ it's not out of commission -- it's up and walking again in 40 seconds thanks to a new "intelligent trial and error" algorithm.
It was created by French and American computer scientists who wanted robots to be able to adapt to and recover from damages‚ just like humans and animals.
So‚ when the six-legged robot loses a leg‚ it runs through its library of nearly 13-thousand walking patterns‚?? thousands of which only use five legs.
The patterns are calculated in advance using a computer model‚ which shortens the trial and error process for the robot itself.
Researchers also tested the algorithm on a robotic arm.
Until now‚ the only way for a robot to become adaptable was the installation of expensive self-diagnosis sensors‚ whereas the algorithm is much more cost-effective.
The team says its research‚ published in the journal Nature‚ could work for any kind of movement.
That means robots will be better able to perform tasks in treacherous conditions‚ from nuclear accidents to earthquakes.
Park Se-young‚ Arirang News. Reporter :
- 저작물 파일 유형
-
- 저작물 속성
-
1
차 저작물
- 공동저작자
-
1유형
- 수집연계 URL
-
https://www.kf.or.kr
- 분류(장르)
-
영상
- 원문제공
-
원문URL
저작물 정보 레이어 닫기
