Skip to main content

A Theory of Soft Synergies for a New Generation of Artificial Hands

Final Report Summary - SOFT HANDS (A Theory of Soft Synergies for a New Generation of Artificial Hands)

"Soft Hands: A Theory of Soft Synergies for a New Generation of Artificial Hands" started from the observation that – notwithstanding the many advances made in the mechatronics and computational hardware of artificial hands - the state of the art appeared to be only marginally closer to a satisfactory functional approximation of the human hand than it was twenty years before. In my analysis, the main reasons for this were not merely technical, but concerned some fundamental issues in the understanding of the organization and control of hands, and ultimately the lack of a theory to guide us in the search for a principled approach to taming the complexity of hands – meant in this project as the physical embodiments of the sense of active touch, and comprised of the sensorimotor apparatus that ultimately creates the link between perception and action.
Building on a coherent theoretical infrastructure, that of soft synergies as constraints that enable and determine cognition, SoftHands is contributing to the science and technology of hands in the following ways:
1) To investigate and clarify the conceptual organization and the geometric structure of synergies underpinning the relation between the hand embodiment and its functions and provide analytic, quantitative models to interpret and predict experiments in human and robot grasping, manipulation, and haptic exploration;
2) To understand the role of variable compliance of the hand motor system in adapting the rigid synergy-based grasping prototypes to direct physical interaction with unknown objects and environments, and to develop a new technology of muscle-like soft actuators for hands to replicate it;
3) To develop and build the sensory apparatus of Soft Hands including distributed tactile sensing, contact area sensing, proprioceptive touch sensing and novel methods for impedance estimation, and to organize and integrate these components in accordance with the principles of sensorial synergies as enabling constraints;
4) To design, build and control hands for robots of a new generation, exploiting soft synergies to achieve a minimalistic realization of the hand sensorimotor and control architecture capable of being programmed more simply and to adapt more robustly to different task/environment conditions;
5) To provide new and enhanced learning and adaptation capabilities of manipulation and haptic perception by using the new hands on a humanoid robot made available at IIT, thus demonstrating that the synergy approach to manipulation can be integrated seamlessly and purposefully into a two-arm, multisensory-equipped, developmentally-oriented architecture.

Our teamwork has produced results in the planned directions, and we are overall well aligned with the original plan. However, under some regards, we have been exceeding our most optimistic expectations. The main theoretical advances we have achieved are the proposition and formal description in mathematical terms of the ideas of soft motor synergies and touch sensory synergies. Our technological developments were firmly based on these ideas, and achieved notable results especially with the design of a new generation of robot hands (SoftHands) adopting a revolutionary adaptive approach that drastically reduces complexity without compromising effectiveness, and a series of new concepts and devices to measure and display touch interaction of the human hand.
Our work has met with wide and warm recognition from the scientific community, general public and the media. Most importantly, applications of our work are progressing well in the planned domains, with several industries and a number of research centers taking up the new technologies we developed. Unforeseen application perspectives have sprung off, such as in prosthetics, which will possibly lead to an even larger impact on societal and economic aspects than initially expected.