0
Research Papers

Accuracy of Wearable Sensors for Estimating Joint Reactions

[+] Author and Article Information
Ryan S. McGinnis

Mem. ASME
Department of Mechanical Engineering,
University of Michigan,
Ann Arbor, MI 48109
e-mail: ryanmcg@umich.edu

Jessandra Hough

Department of Mechanical Engineering,
University of Michigan,
Ann Arbor, MI 48109
e-mail: jesshough3@gmail.com

Noel C. Perkins

Fellow ASME
Department of Mechanical Engineering,
University of Michigan,
Ann Arbor, MI 48109
e-mail: ncp@umich.edu

1Corresponding author.

Contributed by the Design Engineering Division of ASME for publication in the JOURNAL OF COMPUTATIONAL AND NONLINEAR DYNAMICS. Manuscript received March 12, 2016; final manuscript received December 28, 2016; published online January 24, 2017. Assoc. Editor: Sotirios Natsiavas.

J. Comput. Nonlinear Dynam 12(4), 041010 (Jan 24, 2017) (10 pages) Paper No: CND-16-1129; doi: 10.1115/1.4035667 History: Received March 12, 2016; Revised December 28, 2016

Miniature wireless inertial measurement units (IMUs) hold great promise for measuring and analyzing multibody system dynamics. This relatively inexpensive technology enables noninvasive motion tracking in broad applications, including human motion analysis. This paper advances the use of an array of IMUs to estimate the joint reactions (forces and moments) in multibody systems via inverse dynamic modeling. In particular, this paper reports a benchmark experiment on a double-pendulum that reveals the accuracy of IMU-informed estimates of joint reactions. The estimated reactions are compared to those measured by high-precision miniature (6 degrees-of-freedom) load cells. Results from ten trials demonstrate that IMU-informed estimates of the three-dimensional reaction forces remain within 5.0% RMS of the load cell measurements and with correlation coefficients greater than 0.95 on average. Similarly, the IMU-informed estimates of the three-dimensional reaction moments remain within 5.9% RMS of the load cell measurements and with correlation coefficients greater than 0.88 on average. The sensitivity of these estimates to mass center location is discussed. Looking ahead, this benchmarking study supports the promising and broad use of this technology for estimating joint reactions in human motion applications.

FIGURES IN THIS ARTICLE
<>
Copyright © 2017 by ASME
Your Session has timed out. Please sign back in to continue.

References

Figures

Grahic Jump Location
Fig. 1

Schematic and photograph of the instrumented double-pendulum with definitions of reference frames, geometric, and inertial properties ((a) full pendulum schematic and photograph, (b) top link, and (c) bottom link). Front and back views of the inertial sensor secured to each link are shown in (d).

Grahic Jump Location
Fig. 2

Free body diagrams for the bottom (a) and top (b) links of the double-pendulum

Grahic Jump Location
Fig. 7

Force (a) and moment (b) components at j1 and force (c) and moment (d) components at j2 using updated mass center positions. Solid curves correspond to IMU-predicted reactions, while dashed curves correspond to load cell measurements. The three colors distinguish components resolved in frame G: blue =  Ê1T, Ê1G, green =  Ê2T, Ê2G, and red =  Ê3T, Ê3G.

Grahic Jump Location
Fig. 8

Normalized force (a) and moment (b) components at j1 and normalized force (c) and moment (d) at j2 using updated mass center positions. Estimated values are plotted versus measured values. The three colors distinguish components resolved in frame G: blue =  Ê1T, Ê1G, green =  Ê2T, Ê2G, and red =  Ê3T, Ê3G and where the unit line is shown in black.

Grahic Jump Location
Fig. 6

Three components of normalized force (a) and moment (b) acting at j2 as predicted by IMU-enabled inverse dynamic modeling plotted against those measured directly by the load cell. The three colors distinguish components resolved in frame G: blue =  Ê1G, green =  Ê2G, and red =  Ê3G and where the unit line is shown in black.

Grahic Jump Location
Fig. 5

Reaction force (a) and moment (b) at j2 as measured by the load cells (dashed) and estimated using IMU data (solid). The three colors distinguish components resolved in frame G: blue =  Ê1G, green =  Ê2G, and red =  Ê3G.

Grahic Jump Location
Fig. 4

Drift-polluted (gray) and corrected (colored) Euler parameters defining the orientation of the bottom link up until 10 s after t3 during the example trial

Grahic Jump Location
Fig. 3

IMU acceleration (a) and angular velocity (b) history for an example trial sampled from the bottom link IMU. The pendulum begins at rest in its stable equilibrium position (t < t1), is perturbed from this position by hand (t1 < t < t2), and then released (t = t3).

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