As described in Chaps. 2–5, neuroscientific studies showed that the control of the human hand is mainly realized in a synergistic way. Recently, taking inspiration from this observation, with the aim of facing the complications consequent to the high number of degrees of freedom, similar approaches have been used for the control of robotic hands. As Chap. 12 describes SynGrasp, a useful technical tool for grasp analysis of synergy-inspired hands, in this chapter recently developed analysis tools for studying robotic hands equipped with soft synergy underactuation (see Chap. 8) are exhaustively described under a theoretical point of view. After a review of the quasi-static model of the system, the Fundamental Grasp Matrix (FGM) and its canonical form (cFGM) are presented, from which it is possible to extract relevant information as, for example, the subspaces of the controllable internal forces, of the controllable object displacements and the grasp compliance. The definitions of some relevant types of manipulation tasks (e.g. the pure squeeze, realized maintaining the object configuration fixed but changing contact forces, or the kinematic grasp displacements, in which the grasped object can be moved without modifying contact forces) are provided in terms of nullity or non-nullity of the variables describing the system. The feasibility of such predefined tasks can be verified thanks to a decomposition method, based on the search of the row reduced echelon form (RREF) of suitable portions of the solution space. Moreover, a geometric interpretation of the FGM and the possibility to extend the above mentioned methods to the study of robotic hands with different types of underactuation are discussed. Finally, numerical results are presented for a power grasp example, the analysis of which is initially performed for the case of fully-actuated hand, and later verifying, after the introduction of a synergistic underactuation, which capacities of the system are lost, and which other are still present.
This work was supported by the European Commission under the CP-IP grant
no. 248587 “THE Hand Embodied”, within the FP7-2007-2013 program, by the
grant no. 600918 “PaCMan” - Probabilistic and Compositional Representations of
Objects for Robotic Manipulation - within the FP7-ICT-2011-9 program, the grant
no. 611832 “Walk-Man” within the FP7-ICT-2013-10 program, and the grant no.
645599 “SOMA: Soft-bodied Intelligence for Manipulation”, funded under H2020-