0
Research Papers

Multidirectional In Vivo Characterization of Skin Using Wiener Nonlinear Stochastic System Identification Techniques

[+] Author and Article Information
Matthew D. Parker

Auckland Bioengineering Institute,
The University of Auckland,
Private Bag 92019,
Auckland 1142, New Zealand
e-mail: mpar145@aucklanduni.ac.nz

Lynette A. Jones

BioInstrumentation Laboratory,
Massachusetts Institute of Technology,
77 Massachusetts Avenue,
Cambridge, MA 02139
e-mail: ljones@mit.edu

Ian W. Hunter

BioInstrumentation Laboratory,
Massachusetts Institute of Technology,
77 Massachusetts Avenue,
Cambridge, MA 02139
e-mail: ihunter@mit.edu

A. J. Taberner

Department of Engineering Science,
Auckland Bioengineering Institute,
The University of Auckland,
Private Bag 92019,
Auckland 1142, New Zealand
e-mail: a.taberner@auckland.ac.nz

M. P. Nash

Department of Engineering Science,
Auckland Bioengineering Institute,
The University of Auckland,
Private Bag 92019,
Auckland 1142, New Zealand
e-mail: martyn.nash@auckland.ac.nz

P. M. F. Nielsen

Department of Engineering Science,
Auckland Bioengineering Institute,
The University of Auckland,
Private Bag 92019,
Auckland 1142, New Zealand
e-mail: p.nielsen@auckland.ac.nz

1Corresponding author.

Manuscript received May 24, 2016; final manuscript received October 7, 2016; published online November 4, 2016. Assoc. Editor: Kristen Billiar.

J Biomech Eng 139(1), 011004 (Nov 04, 2016) (11 pages) Paper No: BIO-16-1221; doi: 10.1115/1.4034993 History: Received May 24, 2016; Revised October 07, 2016

A triaxial force-sensitive microrobot was developed to dynamically perturb skin in multiple deformation modes, in vivo. Wiener static nonlinear identification was used to extract the linear dynamics and static nonlinearity of the force–displacement behavior of skin. Stochastic input forces were applied to the volar forearm and thenar eminence of the hand, producing probe tip perturbations in indentation and tangential extension. Wiener static nonlinear approaches reproduced the resulting displacements with variances accounted for (VAF) ranging 94–97%, indicating a good fit to the data. These approaches provided VAF improvements of 0.1–3.4% over linear models. Thenar eminence stiffness measures were approximately twice those measured on the forearm. Damping was shown to be significantly higher on the palm, whereas the perturbed mass typically was lower. Coefficients of variation (CVs) for nonlinear parameters were assessed within and across individuals. Individual CVs ranged from 2% to 11% for indentation and from 2% to 19% for extension. Stochastic perturbations with incrementally increasing mean amplitudes were applied to the same test areas. Differences between full-scale and incremental reduced-scale perturbations were investigated. Different incremental preloading schemes were investigated. However, no significant difference in parameters was found between different incremental preloading schemes. Incremental schemes provided depth-dependent estimates of stiffness and damping, ranging from 300 N/m and 2 Ns/m, respectively, at the surface to 5 kN/m and 50 Ns/m at greater depths. The device and techniques used in this research have potential applications in areas, such as evaluating skincare products, assessing skin hydration, or analyzing wound healing.

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

Microrobot device used to perturb skin in vivo. (a) A close-up of the robot components. (b) A potential skin site positioned on the aluminum-acrylic support structure prior to testing. Note that the relative positions of the robot and subject's arm are chosen for demonstrative purposes and do not reflect exact test conditions presented in this study.

Grahic Jump Location
Fig. 2

Schematic system diagram. The system refers exclusively to the combination of the robot and tissue.

Grahic Jump Location
Fig. 3

Stochastic input with stepwise increase in mean, referred to as “protocol B.” Preconditioning at each step was performed by holding the mean value. The first 10 s show a full-scale preconditioning step where the maximum output force is held for demonstrative purposes and do not reflect exact test conditions presented in this study.

Grahic Jump Location
Fig. 4

Location, direction, and order of applied tangential stretches. Dashed line indicates approximate proximal–distal axis that intersects the test site.

Grahic Jump Location
Fig. 5

Representative experimental results from nonlinear stochastic system identification on the volar forearm. The measured input force (a) is used to generate the linear impulse response function (b), shown blue in measured form and red in parameterized form. The linear dynamics are then passed through the Wiener nonlinearity, as shown in blue in (c). The nonlinearity has been parameterized, shown by the red line. The output of the Wiener nonlinearity is shown in (d), where the Wiener-predicted output (red) is shown against the potentiometer-measured output (blue).

Grahic Jump Location
Fig. 6

Representative static nonlinearity plot for a volar forearm using incremental loading schemes, protocols A and B under (a) normal indentation and (b) extension

Grahic Jump Location
Fig. 7

Representative experimental results from linear stochastic system identification on a forearm using incremental loading, under normal indentation and across-surface extension. (a) The tissue stiffness estimated at various stretches is shown. (b) The perturbed mass estimated as various stretches is shown, after the actuator mass is subtracted. (c) The tissue damping at various stretches is shown after the actuator damping is subtracted. (d) The VAF for each site is plotted against actuator tip position.

Grahic Jump Location
Fig. 8

Representative experimental results from linear incremental stochastic system identification of a volar forearm and thenar eminence test are shown. Each plot shows a different linear output property as produced by an incremental loading scheme, in different perturbation directions and/or sites. (a) The tissue stiffness estimated at various stretches is shown. (b) The perturbed mass estimated at various stretches is shown, after the actuator mass is subtracted. (c) The tissue damping at various stretches is shown, after the actuator damping is subtracted. (d) The VAF for each site is plotted against actuator tip position.

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