Improving an Ergonomics Testing Procedure via Approximation-based Adaptive Experimental Design

[+] Author and Article Information
Michael J. Sasena

 Emmeskay, Inc., Plymouth, MI 48170msasena@umich.edu

Matthew Parkinson

 University of Michigan,Department of Biomedical Engineering, Ann Arbor, MI 48109-2125mparkins@umich.edu

Matthew P. Reed

 University of Michigan Transportation Research Institute, Associate Research Scientist, Ann Arbor, MI 48109-2150mreed@umich.edu

Panos Y. Papalambros

 University of Michigan, Department of Mechanical Engineering, Ann Arbor, MI 48109-2125pyp@umich.edu

Pierre Goovaerts

 Biomedware, Inc. and PGeostat, LLC, Ann Arbor, MI 48103goovaerts@biomedware.com

J. Mech. Des 127(5), 1006-1013 (Dec 12, 2004) (8 pages) doi:10.1115/1.1906247 History: Received August 12, 2003; Revised December 12, 2004

Adaptive design refers to experimental design where the next sample point is determined by information from previous experiments. This article presents a constrained optimization algorithm known as superEGO (a variant of the EGO algorithm of Schonlau, Welch, and Jones) that can create adaptive designs using kriging approximations. Our primary goal is to illustrate that superEGO is well-suited to generating adaptive designs which have many advantages over competing methods. The approach is demonstrated on a novel human-reach experiment where the selection of sampling points adapts to the individual test subject. Results indicate that superEGO is effective at satisfying the experimental objectives.

Copyright © 2005 by American Society of Mechanical Engineers
Your Session has timed out. Please sign back in to continue.



Grahic Jump Location
Figure 4

Example of a test subject reaching for the target. The chair can be rotated and the target can be moved up and down, left and right.

Grahic Jump Location
Figure 5

Example of the static, ray-based experimental design, shown for the lateral plane

Grahic Jump Location
Figure 1

Flowchart of the superEGO algorithm

Grahic Jump Location
Figure 2

Kriging predictions for interpolating (solid) and smoothed (dashed) models (top). A close up of the boxed region is shown at bottom. Data points are shown as circles.

Grahic Jump Location
Figure 3

Kriging variance for interpolating (solid) and smoothed (dashed) models (top). A close up of the boxed region is shown at bottom.

Grahic Jump Location
Figure 6

Plots of “iso-reach difficulty.” The wire box indicates the encroachment zone around the test subject. Distances are measured in mm.

Grahic Jump Location
Figure 7

Average separation distance of sample points for each test subject. Points and x’s are results from adaptive and static experiments, respectively.

Grahic Jump Location
Figure 8

Summary metrics of the distance (in mm) to the predicted maximum reach envelope for each test subject. Points and x’s are from adaptive and static experiments, respectively.



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