0
Research Papers

Robustness of Using Dynamic Motions and Template Matching to the Limb Position Effect in Myoelectric Classification

[+] Author and Article Information
Sungtae Shin

Department of Mechanical Engineering,
Texas A&M University,
College Station, TX 77843-3123
e-mails: sstmir@tamu.edu; sstmir@gmail.com

Reza Tafreshi

Associate Professor
Mem. ASME
Department of Mechanical Engineering,
Texas A&M University at Qatar,
P. O. Box 23874,
Doha, Qatar
e-mail: reza.tafreshi@qatar.tamu.edu

Reza Langari

Professor
Mem. ASME
Department of Mechanical Engineering,
Texas A&M University,
College Station, TX 77843-3123
e-mail: rlangari@tamu.edu

1Corresponding author.

Contributed by the Dynamic Systems Division of ASME for publication in the JOURNAL OF DYNAMIC SYSTEMS, MEASUREMENT, AND CONTROL. Manuscript received November 5, 2015; final manuscript received May 23, 2016; published online July 15, 2016. Assoc. Editor: Xiaopeng Zhao.

J. Dyn. Sys., Meas., Control 138(11), 111009 (Jul 15, 2016) (11 pages) Paper No: DS-15-1554; doi: 10.1115/1.4033835 History: Received November 05, 2015; Revised May 23, 2016

Myoelectric classification has been widely studied for controlling prosthetic devices and human computer interface (HCI). However, it is still not robust due to external conditions: limb position changes, electrode shifts, and skin condition changes. These issues compromise the reliability of pattern recognition techniques in myoelectric systems. In order to increase the reliability in the limb position effect when a limb position is changed from the position in which the system is trained, this paper proposes a myoelectric system using dynamic motions. Dynamic time warping (DTW) technique was used for the alignment of two different time-length motions, and correlation coefficients were then calculated as a similarity metric to classify dynamic motions. On the other hand, Fisher's linear discriminant analysis was applied on static motions for the purpose of dimensionality reduction and Naïve Bayesian classifier for classifying the motions. To estimate the robustness to the limb position effect, static and dynamic motions were collected at four different limb positions from eight human subjects. The statistical analysis, t-test (p < 0.05), showed that, for all subjects, dynamic motions were more robust to the limb position effect than static motions when training and validation sets were extracted from different limb positions with the best classification accuracy of 97.59% and 3.54% standard deviation (SD) for dynamic motions compared with 71.85% with 12.62% SD for static motions.

FIGURES IN THIS ARTICLE
<>
Copyright © 2016 by ASME
Your Session has timed out. Please sign back in to continue.

References

Figures

Grahic Jump Location
Fig. 1

(a) Static motion versus (b) dynamic motion; RMS in a static motion (hand close) is relatively steady, whereas RMS in a dynamic motion (finger snap) changes over time

Grahic Jump Location
Fig. 2

(a) Process diagrams of PMC and (b) the proposed SMC

Grahic Jump Location
Fig. 3

Static motions (left), the shape of a motion is equal over time: (a) rest, (b) hand close, (c) hand open, (d) wrist flexion, (e) wrist extension, (f) forearm pronation, and (g) forearm supination. Dynamic motions (right), the shape of a motion varies over time: (a) finger snapping, (b) finger gun, (c) finger beckon, (d) palm beckon, (e) wave, and (f) go away.

Grahic Jump Location
Fig. 4

Location of surface EMG electrodes to collect EMG data: (a) anterior view and (b) posterior view

Grahic Jump Location
Fig. 5

Four different limb positions: (1) straight arm hanging at side (P1), (2) straight arm reaching forward (P2), (3) straight arm reaching up (P3), and (4) humerus hanging at side and forearm horizontal (P4)

Grahic Jump Location
Fig. 6

Before and after alignment by DTW

Grahic Jump Location
Fig. 7

Six templates of dynamic motions

Grahic Jump Location
Fig. 8

Classification accuracy of subject 5 (the best case): (a) SMC and (b) PMC. SDs are in the parentheses. A 50 ms window size with 25 ms overlap for dynamic motion and a 200 ms window size with 100 ms overlap for static motions were chosen.

Grahic Jump Location
Fig. 9

Classification accuracy of subject 7 (the worst case): (a) SMC and (b) PMC. SDs are in the parentheses. A 50 ms window size with 25 ms overlap for dynamic motion and a 200 ms window size with 100 ms overlap for static motions were chosen.

Grahic Jump Location
Fig. 10

(a) Static motion (hand closed) versus (b) dynamic motion (finger snap) at the different limb positions; after changing a limb position, the same static motions are totally different. However, the same dynamic motions keep their similar RMS profiles even their RMS values are slightly shifted.

Tables

Errata

Discussions

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related eBook Content
Topic Collections

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In