Contact Classification using Tactile Arrays

This paper proposed a contact recognition algorithm capable of determining contact surfaces types by tactile sensor fusion. This algorithm can recognize grasped objects using a 2- fingered robot hand, on which tactile sensors are mounted. Experiments have demonstrated reliable surface classification and accurate transformation independent of object shape and orientation. The contribution of this paper regards the relations between a surface and the points cloud defining it. Particularly, we investigate on the possibility that there is some intrinsic feature of a points cloud under quadric surface, allowing us to discriminate or classify a dataset defining a surface from the acquisition noise.