REAL-TIME PATTERN RECOGNITION OF FINGER MOVEMENTS USING REGENERATIVE PERIPHERAL NERVE INTERFACES AND IMPLANTED ELECTRODES

Authors

  • Alex Vaskov
  • Philip Vu
  • Alicia Davis
  • Theodore Kung
  • Paul Cederna
  • Cynthia Chestek

DOI:

https://doi.org/10.57922/mec.38

Abstract

Commercial myoelectric control systems using surface electromyography are unable to obtain consistent control signals for finger-specific motions because the desired signals are either obscured by more superficial muscles or non-existent due to the level of amputation. Intramuscular recording techniques and Regenerative Peripheral Nerve Interfaces (RPNIs) can potentially resolve each of these issues. Two persons with transradial amputations had bipolar electrodes surgically implanted into residual musculature and RPNIs. Participants used a low latency pattern recognition system to intuitively distinguish 7 individual finger postures with 100% online success and complete a functional task requiring multiple grasps with a commercially available prosthetic hand. A classifier with the same architecture was also used to distinguish movements in a simultaneous and proportional 2 degree of freedom control scheme. Both participants used this controller in real-time to complete a virtual target matching task with success rates of 99%.

Downloads

Published

2020-07-23

How to Cite

[1]
A. Vaskov, P. Vu, A. Davis, T. Kung, P. Cederna, and C. Chestek, “REAL-TIME PATTERN RECOGNITION OF FINGER MOVEMENTS USING REGENERATIVE PERIPHERAL NERVE INTERFACES AND IMPLANTED ELECTRODES”, MEC Symposium, Jul. 2020.

Conference Proceedings Volume

Section

Myoelectric Control Algorithms

Categories