Vol.6, Special Issue, 2007 pp. 39-52
UDC 007.52+004.89(045)=111
LEARNING OF GRASPS FOR AN ARTIFICIAL HAND
BY TIME CLUSTERING AND TAKAGI-SUGENO MODELING
Rainer Palm1, Boyko Iliev2
1Adjunct professor at the AASS, Department
of Technology, Orebro University
SE-70182 Orebro, Sweden, also Siemens Co. Research, Frankfurt, Germany
2AASS, Department of Technology, Orebro
University SE-70182 Orebro, Sweden
e-mail: rub.palm@t-online.de,
boyko.iliev@tech.oru.se
Abstract. The focus of the paper is the learning of grasp primitives
for a five-fingered anthropomorphic robotic hand via programming-by-demonstration
and fuzzy modelling. In this approach, a number of basic grasps is demonstrated
by a human operator wearing a data glove which continuously captures the
hand pose. The resulting fingertip trajectories and joint angles are clustered
and modelled in time and space so that the motions of the fingers forming
a particular grasp are modelled in a most effective and compact way. Classification
and learning are based on fuzzy clustering and Takagi-Sugeno (TS) modelling.
The presented method allows to learn, imitate and recognize the motion
sequences forming specific grasps.
Key words: Grasp recognition, manipulation robots, programming-by-demonstration,
TS-modelling, time clustering
UČENJE ZAHVATA VEŠTAČKOM RUKOM VREMENSKIM
GRUPISANJEM I TAKAGI-SUGENO MODELIRANJEM
U ovom radu fokus je na učenju primitivnog zahvata petoprstnom antropomorfnom
robotskom šakom preko programiranja demonstracijom i fazi modeliranjem.
U ovom pristupu, demonstrira se veliki broj osnovnih zahvata ljudskim operaterom
koji nosi rukavicu podataka koja neprekidno zahvata položaj ruke. Rezultirajuća
prstne putanjei i spojeni zglobovi sakupljeni su i modelirani u vremenu
i prostoru tako da pokreti prstiju koji formiraju odgovarajući zahvat
mogu da se modeliraju na najefektniji i kompaktniji način. Klasifikacija
i učenje su zasnovani na fazi i Takagi-Sugeno (TS) modeliranju. Predstavljeni
metod omogućava učenje, imitaciju i prepoznavanje pokreta koji slede
formirajući specifične zahvate.
Ključne reči: Zahvat prepoznavanja, programiranje demonstracijom,
manipulacioni roboti, robotska šaka; TS-modeliranje, vremensko grupisanje
