A neural approach to robotic haptic recognition of 3-D objects based on a Kohonen self-organizing feature map

Abstract

This paper describes a novel approach to robotic haptic recognition, which exploits an unsupervised Kohonen self-organizing feature map for performing a match-to-sample classification of three-dimensional (3-D) objects. The results obtained, even though currently referring to a simulated environment and to some working assumptions, have emphasized the validity of the approach and its applicability in a variety of dextrous robotic systems.