Research Papers

Intra-Articular Knee Contact Force Estimation During Walking Using Force-Reaction Elements and Subject-Specific Joint Model2

[+] Author and Article Information
Yihwan Jung, Cong-Bo Phan

School of Mechanical Engineering,
Chung-Ang University,
84 Heukseokro, Dongjakgu,
Seoul 06974, South Korea

Seungbum Koo

School of Mechanical Engineering,
Chung-Ang University,
84 Heukseokro, Dongjakgu,
Seoul 06974, South Korea
e-mail: skoo@cau.ac.kr

1Corresponding author.

2This paper was declared the winner of the 6th Knee Grand Challenge Competition.

Manuscript received October 14, 2015; final manuscript received December 28, 2015; published online January 27, 2016. Editor: Beth A. Winkelstein.

J Biomech Eng 138(2), 021016 (Jan 27, 2016) (9 pages) Paper No: BIO-15-1511; doi: 10.1115/1.4032414 History: Received October 14, 2015; Revised December 28, 2015

Joint contact forces measured with instrumented knee implants have not only revealed general patterns of joint loading but also showed individual variations that could be due to differences in anatomy and joint kinematics. Musculoskeletal human models for dynamic simulation have been utilized to understand body kinetics including joint moments, muscle tension, and knee contact forces. The objectives of this study were to develop a knee contact model which can predict knee contact forces using an inverse dynamics-based optimization solver and to investigate the effect of joint constraints on knee contact force prediction. A knee contact model was developed to include 32 reaction force elements on the surface of a tibial insert of a total knee replacement (TKR), which was embedded in a full-body musculoskeletal model. Various external measurements including motion data and external force data during walking trials of a subject with an instrumented knee implant were provided from the Sixth Grand Challenge Competition to Predict in vivo Knee Loads. Knee contact forces in the medial and lateral portions of the instrumented knee implant were also provided for the same walking trials. A knee contact model with a hinge joint and normal alignment could predict knee contact forces with root mean square errors (RMSEs) of 165 N and 288 N for the medial and lateral portions of the knee, respectively, and coefficients of determination (R2) of 0.70 and −0.63. When the degrees-of-freedom (DOF) of the knee and locations of leg markers were adjusted to account for the valgus lower-limb alignment of the subject, RMSE values improved to 144 N and 179 N, and R2 values improved to 0.77 and 0.37, respectively. The proposed knee contact model with subject-specific joint model could predict in vivo knee contact forces with reasonable accuracy. This model may contribute to the development and improvement of knee arthroplasty.

Copyright © 2016 by ASME
Your Session has timed out. Please sign back in to continue.



Grahic Jump Location
Fig. 1

Sixteen reaction elements were placed on each of the medial and lateral surfaces of the tibial insert (left). Each reaction element could pull and push in three orthogonal directions, with a maximum strength of 3000 N.

Grahic Jump Location
Fig. 2

The developed knee contact model was embedded in a full-body model from AnyBody Managed Model Repository. The full-body model had 19 segments, 38DOF, and more than 1000 muscle actuators.

Grahic Jump Location
Fig. 6

A hinge joint and normal limb alignment were assumed in model A (left). A ball–socket joint with 3DOF was employed in model B, and the skin markers were adjusted to implement the valgus deformity observed in the frontal long-leg radiograph (right).

Grahic Jump Location
Fig. 5

The locations of the seven markers on the lower-limb were perturbed in blinded and unblinded predictions. The mean and standard deviation of the medial, lateral, and total forces of the knee predicted with perturbation were presented along with experimentally measured data.

Grahic Jump Location
Fig. 4

Medial, lateral, and total contact force of the knee from the experimentally measured data, blinded predictions, and unblinded predictions using nominal marker locations

Grahic Jump Location
Fig. 3

A subject-specific leg model was created using (a) the subject's bone geometry, (b) a template with muscle attachment sites, and (c) an anatomical landmark prediction algorithm [28] to estimate (d) the locations of muscle attachment sites using local bone geometries

Grahic Jump Location
Fig. 7

Marker location and joint type variation in model B were tested separately. Medial and lateral contact forces in the knee were predicted during smooth gait when only marker locations were changed in model A (solid line) and when the knee joint type was changed to a ball–socket joint without marker location change (dashed line).




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