Robot motion classification from the standpoint of learning control

Shaw Ji Shiah*, Kuu-Young Young

*Corresponding author for this work

Research output: Contribution to conferencePaper

Abstract

In robot learning control, the learning space for executing the general motions of multijoint robot manipulators is very complicated. Therefore, in spite of their ability to generalize, the learning controllers are usually used as subordinates to conventional controllers or the learning process needs to be repeated each time a new trajectory is encountered, because the motion variety requires them to consume excessive amount of memory when they are employed as major roles in motion governing. To simplify learning space complexity, we propose, from the standpoint of learning control, that robot motions be classified according to their similarities. The learning controller can then be designed to govern groups of robot motions with high degrees of similarity without consuming excessive memory resources. Motion classification based on using the PUMA 560 robot manipulator demonstrates the effectiveness of the proposed approach.

Original languageEnglish
StatePublished - 1 Dec 1999
EventProceedings of the 1999 IEEE International Fuzzy Systems Conference, FUZZ-IEEE'99 - Seoul, South Korea
Duration: 22 Aug 199925 Aug 1999

Conference

ConferenceProceedings of the 1999 IEEE International Fuzzy Systems Conference, FUZZ-IEEE'99
CitySeoul, South Korea
Period22/08/9925/08/99

Fingerprint Dive into the research topics of 'Robot motion classification from the standpoint of learning control'. Together they form a unique fingerprint.

  • Cite this

    Shiah, S. J., & Young, K-Y. (1999). Robot motion classification from the standpoint of learning control. Paper presented at Proceedings of the 1999 IEEE International Fuzzy Systems Conference, FUZZ-IEEE'99, Seoul, South Korea, .