This paper presents a novel model-free approach for modeling friction for servo-motion systems. The proposed approach uses the support vector machine networks to parameterize the static friction mapping. The procedure of constructing such networks from a finite amount of training (sampling) data is developed based on support vector machine regression (SVMR). The validity of the proposed approach has been experimentally verified.
Published in:
Mechatronics, IEEE/ASME Transactions on
(Volume:9
,
Issue:
3
)
Date of Publication: Sept. 2004