An improved parameter estimation and comparison for soft tissue constitutive models containing an exponential function

Aggarwal, A. (2017) An improved parameter estimation and comparison for soft tissue constitutive models containing an exponential function. Biomechanics and Modeling in Mechanobiology, 16(4), pp. 1309-1327. (doi:10.1007/s10237-017-0889-3) (PMID:28251368) (PMCID:PMC5511618)

[img]
Preview
Text
169761.pdf - Published Version
Available under License Creative Commons Attribution.

5MB

Abstract

Motivated by the well-known result that stiffness of soft tissue is proportional to the stress, many of the constitutive laws for soft tissues contain an exponential function. In this work, we analyze properties of the exponential function and how it affects the estimation and comparison of elastic parameters for soft tissues. In particular, we find that as a consequence of the exponential function there are lines of high covariance in the elastic parameter space. As a result, one can have widely varying mechanical parameters defining the tissue stiffness but similar effective stress–strain responses. Drawing from elementary algebra, we propose simple changes in the norm and the parameter space, which significantly improve the convergence of parameter estimation and robustness in the presence of noise. More importantly, we demonstrate that these changes improve the conditioning of the problem and provide a more robust solution in the case of heterogeneous material by reducing the chances of getting trapped in a local minima. Based upon the new insight, we also propose a transformed parameter space which will allow for rational parameter comparison and avoid misleading conclusions regarding soft tissue mechanics.

Item Type:Articles
Status:Published
Refereed:Yes
Glasgow Author(s) Enlighten ID:Aggarwal, Dr Ankush
Authors: Aggarwal, A.
College/School:College of Science and Engineering > School of Engineering > Infrastructure and Environment
Journal Name:Biomechanics and Modeling in Mechanobiology
Publisher:Springer
ISSN:1617-7959
ISSN (Online):1617-7940
Published Online:01 March 2017

University Staff: Request a correction | Enlighten Editors: Update this record