LimbMATE: A Versatile Platform for Closed-Loop Research in Prosthetics

The rapid advancements in machine learning methodologies carry a wide potential to match the increasing control demands of modern robotic hands. In particular, myoelectric pattern recognition is recognized nowadays as a viable means to facilitate the multi-functional control of prosthetic devices, f...

Full description

Saved in:
Bibliographic Details
Main Authors: Erik Gasparini, Rodolfo Cerqueira, Carlo Preziuso, Lucia Angelini, Robinson Guachi, Marco Controzzi, Christian Cipriani, Enzo Mastinu
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Transactions on Neural Systems and Rehabilitation Engineering
Subjects:
Online Access:https://ieeexplore.ieee.org/document/11015816/
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The rapid advancements in machine learning methodologies carry a wide potential to match the increasing control demands of modern robotic hands. In particular, myoelectric pattern recognition is recognized nowadays as a viable means to facilitate the multi-functional control of prosthetic devices, for both lower and upper limb amputations. However, such solutions require dedicated and specific hardware: versatile yet robust, low-power yet computationally powerful. Here, we present LimbMATE: a compact and modular embedded system for the intuitive real-time control of prostheses by decoding either differential or pre-amplified bioelectric signals. It is compatible with various prosthetic components and feedback actuators, disposing of major communication peripherals to adapt to other devices’ proprietary protocols. Moreover, it includes Bluetooth capabilities, and it implements an advanced real-time clock calendar-based logging system for tracking prosthesis usage, opening for long-term at-home longitudinal studies. The LimbMATE system was bench-verified to prove suitable power consumption (<200 mW), real-time operation, and long-term data storage (1 sec/day tolerance). Finally, it was validated with twelve non-disabled participants and two end-users with upper limb amputation via both virtual and real-life-inspired functional assessments. LimbMATE showed the capability for robust and responsive inference of the intended movements to enable intuitive bidirectional control of a multi-articulated prosthesis.
ISSN:1534-4320
1558-0210