HANDBOT- Biomechatronic hand prostheses endowed with bio-inspired tactile perception, bi-directional neural interfaces and distributed sensori-motor control
phone: + 39 06 22 541 9607
phone: + 39 06 22 541 9632
The HandBot project tackles two grand challenges: a) to understand the precise role of human mechanoreceptors in sensorimotor control of grasping and manipulation tasks; b) to deliver a multi-fingered robotic prosthetic hand, embedding biomimetic mechanoreceptors and a bio-inspired tactile perception system, and featuring break-through fine manipulation capabilities.
The system will make use of advanced bi-directional invasive neural interfaces to the Peripheral Nervous System (PNS) in order to enable direct, truly neuromorphic sensorimotor control by the end-user (Fig. 1). This approach is expected to result in a significant enhancement of the overall performance of the hand prosthesis in terms of the variety and quality of enabled grasping and manipulation tasks, and thus of its final acceptability by the amputee.
FIG 1. HandBot scenario
In order to achieve such ambitious objectives the project, on one hand, relies on a solid scientific, strongly multidisciplinary partnership built around a core group of partners having a long-lasting cooperation on this specific research area. They are Università "Campus Bio-Medico" di ROMA, Politecnico di MILANO, Università degli Studi di CAGLIARI, Scuola Superiore di Studi Universitari e Perfezionamento S.Anna di PISA, Consiglio Nazionale delle Ricerche, Scuola Internazionale Superiore di Studi Avanzati di TRIESTE. On the other hand, the proposed developments are based on previous breakthrough achievements and technologies developed along the last 5 years by the HandBot core partners and by a network of EU partners who agreed to co-operate in the framework of this new initiative.
State-of-the-art components (Figs. 2 and 3), such as a biomechatronic 5-fingered prosthesis prototype, thin-film longitudinal (LIFE) and transversal (TIME) intrafascicular neural multi-electrode interfaces will be integrated with the main novel HandBot expected outputs, i.e. a neuromorphic tactile perception system and a neuroinspired hierarchic reinforcement learning, multi-layered control architecture embedded in a miniaturized, customized and fully implantable electronic control unit. Efferent signals will be recorded and classified to extract high-level user commands (e.g. type of required grasping and manipulation tasks), whereas tactile feedback will be used both to close the control loop on the hierarchical control architecture of the biomechatronic hand and to reproduce biomimetic afferent tactile artifacts to be transferred directly to the human brain by using neurostimulation of the PNS. In order to overcome some known limitations of PNS electric stimulation when used to elicit tactile sensations, a novel approach based on PNS electromagnetic stimulation will be for the first time investigated, and preliminarily compared to state-of-the-art technologies and approaches. All developments on artificial tactile perception will be based on the inputs directly deriving from the studies on humans and animal models (rats) that will be carried out by a multidisciplinary HandBot team of experts in neurophysiology, psychophysics, biomechanics, movement and behavioral analysis, human motor control, and neuroengineering. Such studies are expected to generate breakthrough neuroscientific results, by unveiling key mechanisms on the role of tactile perception, but also of tactile sensor morphologies and spatial distribution in our skin, in hand sensorimotor control.
An anthropomorphic robotic arm-hand system will be used to perform preliminary integration and verification of the system and for optimization of the proposed control architecture. Experiments on the robotic platform shall allow demonstrating the advantages deriving from the use of the novel HandBot neuromoprhic tactile perception system in controlling a variety of grasping and manipulation tasks, including handling of critical situations like object slippage and fine, fingertip-based manipulation activities in constrained and unconstrained motion (e.g. opening a bottle of water or using a touch-pad pen). After verification, the HandBot system will be integrated and validated on at least one human subject, by a multidisciplinary equipe at the Campus Bio-Medico Research Policlinic, where similar experiments have been already hosted in the recent past. Advanced neuroimaging techniques (including the use of a 7T fMRI apparatus) will be used for prepost assessment of brain plasticity, interconnectivity and other physiological parameters relevant to the use of the HandBot prosthesis. A rigorous surgical and training experimental protocol will be derived from clinical protocols used for previous experiments by the same team, and submitted for prior approval to the local Ethics Committee and to the Ministry of Health. Validation of the HandBot system shall provide measurable, objective evidence of the advantages deriving from the use of the proposed solution when compared to state-of-the-art results and commercial prostethic systems. If successful, HandBot could generate a significant impact not only in the field of hand prosthetics, but also in robotic manipulation systems, neuroscience, tactile sensor systems for several application domains, and more.
Fig. 2 HanBot robotic and biomechatronic technologies: KUKA LWR and DLR-HIT-Hand II (left); SSSA prosthetic hand (right).
Fig. 3 Neural electrodes: tf-LIFE (top) and TIME (bottom)
UNIVERSITA’ CAMPUS BIO-MEDICO DI ROMA, COORDINATOR
SCUOLA SUPERIORE SANT’ANNA DI PISA
SCUOLA INTERNAZIONALE SUPERIORE DI STUDI AVANZATI DI TRIESTE
POLITENICO DI MILANO
CONSIGLIO NAZIONALE DELLE RICERCHE
UNIVERSITA’ DEGLI STUDI CAGLIARI