Applying fuzzy logic and neural networks to total hand evaluation of knitted fabrics

Citation
Sw. Park et al., Applying fuzzy logic and neural networks to total hand evaluation of knitted fabrics, TEXT RES J, 70(8), 2000, pp. 675-681
Citations number
22
Categorie Soggetti
Material Science & Engineering
Journal title
TEXTILE RESEARCH JOURNAL
ISSN journal
00405175 → ACNP
Volume
70
Issue
8
Year of publication
2000
Pages
675 - 681
Database
ISI
SICI code
0040-5175(200008)70:8<675:AFLANN>2.0.ZU;2-U
Abstract
This study of two new total hand simulating methods for knits uses fuzzy th eory and neural networks. One method, a neural network system trained with a back-propagation algorithm, performs functional mapping between mechanica l properties and the resulting total hand values of the fuzzy predicting me thod. The second method, a fuzzy-neural network system, uses the fuzzy memb ership function, weighted factor vector, and error back-propagation algorit hm. The principal mechanical properties of stretchiness, bulkiness, flexibi lity, distortion, weight, and surface roughness of the knitted fabrics are correlated with experimentally determined Kawabata total hand values and fu zzy transformed overall hand values. Fuzzy and neural networks agree better with the subjective test results than the KES-FB system. The mechanical pr operties are fuzzified by fuzzy membership functions, then trained to predi ct the total hand value of outerwear knitted fabrics. In each case, the pre diction error is less than the standard deviation of experimentation, and t he optimum structure is investigated. These two systems, which use the Pasc al programming language, produce objective ratings of outerwear knit fabric s.