Towards a unified
design of an affordable and versatile artificial hand valid for
prosthetics and collaborative robotics
Abstract:
The design of
artificial hands has had a considerable development in the last
decades, pushed forward by the need to improve existing solutions
for people with a limb loss and because of the increasing importance
of robotic grasping and collaborative robotics. Anthropomorphic
designs, similar to the human hand, are convenient for prosthetic
applications due to evident aesthetic reasons. They could also deal
with a greater variety of objects and grasping modes and allow to
reduce tool changes in industrial environments and to provide
versatility to assistive robotic arms, like those attached to
wheelchairs of disabled people. However, it is impossible to
replicate the complexity of the human hand with current technology.
Low-cost and simplicity are important goals in artificial hand
design, enabling accessibility to developing countries and people
with low incomes, as initiatives such as the Open Hand Project or
Enabling the Future have shown. They can also contribute to
widespread exponentially collaborative robotic applications, running
in parallel to the implementation of Industry 4.0. The two editions
of the Cybathlon Arm Race in 2016 and 2020 have shown that simple
hands with a low number of degrees of freedom can surpass more
sophisticated designs in terms of final performance. Previous
studies developed in our research group have also shown that a
fraction close to 80% of the variability observed in human grasping
in activities of daily living, can be adequately represented by five
synergies and that only two synergies are necessary to represent 80%
of the the variability observed in the tendon actuation of an
artificial hand with six degrees of freedom moved by healthy
subjects. Hence, in this project we rely on the hypothesis that a
unified design for an affordable hand, with a limited number of
degrees of freedom and anthropomorphic appearance can be a good
solution for both prosthetic and collaborative robotic applications.
Hence, the objective of the project is to design such a hand, which
must be robust but soft, simple and affordable, and with a
straightforward attachment to commercial robotic arms and to
prosthetic arm sockets and wrist units. This unified design would
concentrate the efforts for low level control on a unique
architecture and contribute to the standardisation of the interfaces
with high-level control for either robotic or prosthetic
applications. The related work previously developed by the team in
the framework of projects
DEVALHAND
(2014-2016) and
BENCH-HAND
(2017-2020), pave the way for achieving this goal and for applying
suitable design and testing methods developed in those projects to
the development of this new hand. The grasping ability of the new
hand will be demonstrated through realistic grasping environments
such as those of the Cybathlon competition for prosthetic arms or
the robotic competitions of IROS conference. The project will
contribute to solve important challenges and scientific problems,
such as the tradeoffs in finger transmission design between high
grasping force and fast free motion, or the transference of human
synergies to artificial hands to increase workspace without
increasing control complexity. Additionally, the project proposes to
develop new artificial intelligence and augmented reality
applications that could contribute to an easier control development
and user training.
Date:
Sep.
2021-Sep. 2025
PI UJI:
Antonio Pérez
González
Marta Mora Aguilar
Partners:
UJI, University of
Strathclyde, University of Salford, Université Paris-Saclay
Funding:
Spanish Ministry of
Economy, Industry and Competitiveness
(PID2020-118021RB-I00)
