A study of back-error propagation networks in the domain of noisy tactile impressions

The authors have designed an artificial neural system (ANS) consisting of coupled back-error propagation (BEP) networks that perform feature extraction, clustering. and categorization of tactile surface impressions. The network and its characteristics are reviewed, with particular focus on its performance in the presence of noisy input patterns. Simulation results indicate that, regarding geometry-size- and activation-constrained grey-scale patterns, the BEP classifier is sensitive to both additive low-amplitude spike noise and additive white Gaussian noise. Most of the misclassifications occur among patterns that differ only by small variations in force gradients. The network´s performance gradually improves when noisy patterns are included in the training set, but indications are that large training sets or alternative error functions in BEP will be required to achieve robust performance in the tactile domain