Linear regression using intramuscular EMG for simultaneous myoelectric control of a wrist and hand system

Lauren H. Smith*, Todd A. Kuiken, Levi J. Hargrove

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contribution

5 Scopus citations

Abstract

Clinically available myoelectric prostheses are limited by the inability to control multiple degrees of freedom simultaneously. Linear regression-based control and parallel dual-site control (an extension of conventional amplitude-based methods using intramuscular EMG) are two frequently proposed approaches for simultaneous control. Both approaches assume linearity in the EMG features, but differ in whether users are required to independently modulate the EMG amplitudes from residual limb muscles. The objective of this preliminary study was to compare these two methods for the real-time control of a 3 degree-of-freedom (DOF) wrist/hand system. Both systems used intramuscular EMG amplitudes from six forearm muscles, and differed only in how the signals were used to predict intended prosthesis activity. Five able-bodied subjects were recruited to evaluate each control system (ten subjects total). Performance in a virtual Fitts' law task demonstrated that parallel dual-site control provided improved controllability when acquiring targets that required use of only one DOF, but linear regression control provided improved performance when acquiring targets requiring use of all three DOFs. Subjects using linear regression control were more easily able to activate multiple DOFs simultaneously, but at the expense of unintended movement when trying to isolate individual DOFs.

Original languageEnglish (US)
Title of host publication2015 7th International IEEE/EMBS Conference on Neural Engineering, NER 2015
PublisherIEEE Computer Society
Pages619-622
Number of pages4
Volume2015-July
ISBN (Electronic)9781467363891
DOIs
StatePublished - Jan 1 2015
Event7th International IEEE/EMBS Conference on Neural Engineering, NER 2015 - Montpellier, France
Duration: Apr 22 2015Apr 24 2015

Other

Other7th International IEEE/EMBS Conference on Neural Engineering, NER 2015
CountryFrance
CityMontpellier
Period4/22/154/24/15

ASJC Scopus subject areas

  • Artificial Intelligence
  • Mechanical Engineering

Fingerprint Dive into the research topics of 'Linear regression using intramuscular EMG for simultaneous myoelectric control of a wrist and hand system'. Together they form a unique fingerprint.

Cite this