Integración sensorial utilizando clasificación difusa en tareas de ensamblado en robótica

Abstract

Some approaches to the automation of assembly tasks with robots have been proposed. However, nearly all of them face the same problem: the need of identifying the current contact situation from sensory data. The confidence of assembly tasks with robotic manipulators is affected by uncertainty. The knowledge of the generalized reaction force is the natural complement to configuration information in the presence of geometric uncertainty, since a reaction force indicates a constraint in Configuration Space. Sensorial integration techniques have as objective to use information provided for several sensors to build and update a world model in order to achieve a goal. In this paper we present an approach to multisensory integration based on fuzzy c-means clustering algorithm for detection and identification of states during a dexterous manipulation task in robotics, considering modeling and sensing uncertainties.