On automatic identification of upper-limb movements using small-sized training sets of EMG signals

We evaluate the performance of a variety of neural and fuzzy networks for d iscrimination among three planar arm-pointing movements by means of electro myographic (EMG) signals, when learning is based on small-sized training se es. The aim of this work is to underline the importance that the sparse dat a problem has in designing pattern classifiers with good generalisation pro perties. The results indicate that one of the proposed fuzzy networks is mo re robust than the other classifiers when working with small training sets. (C) 2001 IPEM. Published by Elsevier Science Ltd. All rights reserved.