3D Grasp Synthesis Based on Object Exploration

Many approaches to robotic grasping have focused on a specific aspect of the problem only, without considering its integrability with other related procedures in order to build a more complex task. The model for grasp synthesis presented in this paper, inspired on human neurophysiology, is built upon an architecture that allows its scalability and its integration within more complex tasks. The grasp synthesis is designed as integrated with the extraction of a 3D object description, so that the object visual analysis is driven by the needs of the grasp synthesis: visual reconstruction is performed incrementally and selectively on the regions of the object that are considered more interesting for grasping. Our approach, inspired by the efficiency of our visual cortex, allows for an easy integration of additional modules and different grasp synthesis criteria.