Myoelectric Prosthetic Hands: A Review of Muscle Synergy, Machine Learning and Edge Computing

Abstract

Over the past decade, the integration of electromyography (EMG) techniques with machine learning has significantly advanced prosthetic device control. Researchers have developed sophisticated deep learning classifiers for gesture recognition and created EMG controllers capable of simultaneous proportional control across multiple degrees of freedom. However, the increasing complexity of these machine learning models demands greater computational power, creating challenges for real-time deployment on embedded prosthetic controllers. Various optimization techniques - including hyperdimensional computing, pruning, and quantization - have demonstrated effectiveness in reducing computational requirements while preserving system performance. Concurrently, biomedical research has explored muscle and task synergies as methods to simplify inputs for machine learning models. This review examines synergy extraction in upper limb prosthetics research and identifies the need for standardized hardware specifications to facilitate proper validation and comparison of research outcomes. Furthermore, it explores how optimization techniques from Internet of Things (IoT) applications could enhance EMG controllers in biomedical settings. The analysis identifies sensor fusion and high-density EMG as particularly promising approaches for achieving robust, generalized control of upper limb prosthetics.