BENCH-HAND project

Data

  • Acronym: BENCH-HAND

  • Title: Development of benchmarks for the experimental evaluation of artificial hands. Application to the design of better robotic and prosthetic hands.

  • Dates: Jan. 2018- Dec. 2020 (extended to Sep.2021)

  • Funder: Spanish Ministry of Economy, Industry and Competitiveness and the ERDF under the grant DPI2017-89910-R.

Summary

The development of efficient artificial hands, both for use in robots and for prosthetic replacements, remains a complex challenge for design and control engineering. In the past, simulation tools have been used to evaluate the functionality of robotic grippers and anthropomorphic hands and to define grasping metrics for motion planning. However, these simulation tools are limited to consider the actual functionality of hand designs in daily life activities, since aspects as the friction, the deformation, the joints stiffness or the dynamic response of actuators or the control system, are difficult to incorporate into analytical metrics. Therefore, the experimental evaluation of prototypes of artificial hands is essential to move towards more efficient solutions. From the analysis of the literature and from the previous developments of the research team, is deduced the need for standardized equipment, protocols and metrics that permit the experimental evaluation of different prototypes of artificial hands, whether they are robotic or prosthetic hands, and the definition of quality metrics or indices compared to the human hand. In this project it is assumed that this experimental evaluation must be structured to analyse separately the three main factors that affect the functionality of an artificial hand: 1) the mechanical design (topology, materials) that affect to the attainable stable grip postures 2) the dynamic performance of their actuators and mechanical transmission, which condition the dynamic response as grasp forces and opening/closing times 3) the control system, affecting to the proprioception, sensitivity, responsiveness or versatility for different tasks. It is also assumed that the experimental metrics that allow evaluating hand prototypes according to these three factors must be based on the comparison with the performance of the human hand. Therefore, the project has three main objectives:

    1. To develop test equipment and protocols (benchmarks) that allow for the experimental evaluation of any artificial hand compared to the human hand, using metrics defined at the three levels above-mentioned (mechanical design, actuators performance, control system).

    2. To obtain an improved design of a robotic hand for service applications, using the methodologies developed in objective 1 at its three levels.

    3. To obtain an improved design of a low-cost motorized transradial hand prosthesis, using the methodologies developed in objective 1 at its three levels.

Background and motivation

The research team that proposes the project is composed by researchers from two research groups of the Universitat Jaume I (UJI) with extensive experience in grasping research: the Intelligent Robotics group and the Biomechanics and Ergonomics group. Both groups accumulate an experience of more than 15 years in the study of grasping with robots and with the human hand, with results widely cited for the synthesis of the robotic grip or for the human grip simulation. For more than 5 years they have started the development of joint research in order to take advantage of the synergies and multidisciplinarity applied to the simulation of grasping and to the design of more efficient manipulators. The DEVALHAND project "Design and evaluation of artificial anthropomorphic hands through grasping simulation. Application to the design and control of hand prostheses" has been a first approach to the design of anthropomorphic hands with application to the field of service robotics and the design of hand prostheses. In this project, a human and robotic grip simulation tool developed in a thesis co-supervised between both groups (León, 2014) was employed to obtain new hand designs with good grip quality indicators. However, the experience accumulated throughout the project has shown that the use of the simulation must be complemented by a stage of experimentation on the developed prototypes, to solve part of the limitations of the simulation models. Although the DEVALHAND project has carried out preliminary experimental tests on prototypes of the developed anthropomorphic hands, it has been observed the need for standardized benchmarks (devices and protocols) for the experimental evaluation of different prototypes of artificial hands, whether robotic or prosthetic, and the importance of having metrics or indices of quality in comparison to the human hand. Although there are some antecedents in the literature, both in the robotic field and in the evaluation of hand prostheses, these are not exhaustive, and do not define metrics that allow comparison with the human hand. Besides, they do not allow to separate the evaluation in the different scales of the design problem: mechanical or kinematic design of the manipulator, dynamic capacities of the actuators and efficiency of the control system, what is considered essential to obtain clear conclusions for the redesign of the prototypes.