Wood, G. A. (June 1992)
"Of Movement and Measures." VIII Meeting of the European Society of
Biomechanics - Rome - Italy, June 21-24, p.274
Introduction - Procedures for the analysis of animal and human movement are now
well documented (see for example Winter, 1990), and typically involve linked rigid
body models together with inverse dynamic analysis using as a basis the time histories
of spatial coordinates which identify the locations of the instantaneous 'joint'
centres of rotation. Such models can yield a large number of variables which describe
the behaviour of the biokinematic system in mechanical terms (albeit in a non-deterministic
fashion), and these can provide a valuable basis for the study of normal, injurious
and pathological patterns of movement.
These procedures are, however, very error-prone, and appropriate attention is
often not given to the fidelity of linked rigid body model predictions, nor to
the underlying biological variability itself. Few comprehensive studies of the
sources and extent of errors have been reported, and the sensitivity of results
to these errors and the consequent uncertainty of predictions are rarely estimated.
This paper presents a synthesis of reported measurement errors and, in the context
of current technology, outlines some approaches which can be adopted to estimate,
as well as enhance, the precision (validity) and accuracy (reliability) of results.
<< BACK