Selected examples in biomechanical modelling and simulation
In the past decades, multibody and numerical models have become very popular to analyse and optimize human motion. A variety of measuring techniques have been developed to determine subject-specific parameter values, such as for describing muscle properties (e.g. the force–velocity relationship of a specific muscle). Special attention is given to the development of individual models of human body segments. Tools and methods from image acquisition, image processing and medical imaging are applied to obtain shapes, density distributions and mechanical parameter values of these segments. Sophisticated models of human joints help to better understand the mechanical properties within these joints or to predict the loads on the joints. Models for simulating muscle behaviour show large differences in complexity; some are even based on the molecular structure of the muscle. Finite elements or computational fluid dynamics forms the basis for numerical models to be applied in a variety of biomechanical research questions, such as in swimming or sports equipment design.
Behavioural processes like those in motor activities are often characterized by a complex structure; measurements considering them may produce a huge amount of data. It is an interesting challenge to transform them into useful information. Artificial neural networks turn out to be an appropriate tool to transform abstract numbers into informative patterns that help us to understand complex behavioural phenomena. Principal component analysis (PCA) techniques provide a promising alternative. They also help to reduce multidimensional biomechanical data to fewer dimensions for further analysis.
This special issue is concerned with manifold aspects of biomechanical modelling and simulation. Theoretical and practical results from different research areas are reported. The articles are based on selected contributions that were presented in a special session of the 6th MATHMOD conference in February 2008 in Vienna, Austria.
In the first article, Karen Römer, Uwe Jungnickel, Frank Lindner and Thomas L. Milani adapt and implement an individualized knee model into a multibody system. They show how knee and femoropatellar model types influence the magnitude of calculated knee forces.
Irene Reichl, Winfried Auzinger, Heinz-Bodo Schmiedmayer and Ewa Weinmüller compare two different knee models with regard to the accuracy of the results when reconstructing knee joint motion from noisy kinematic data. The model resulting in more accurate results is not unambiguous but depends on the magnitude of noise.
Model equations describing the relationship between individual neuromuscular properties and the associated sports performance are presented by Sigrid Thaller, Markus Tilp and Martin Sust. The authors describe methods to identify values of these subject-specific properties and simulate effects of individual differences in these parameters on performance.
In the article by Peter Dabnichki and Angel Zhivkov two different methods are proposed for reconstructing biological objects from discrete noisy data. The methods described allow to process data directly from optical or general image devices, such as cameras or scans.
The aim of the article presented by Falk Hildebrand and Axel Schüler is to explain swimming propulsion from a phenomenological point of view. The authors' conclusions are deduced from three-dimensional video analyses of swimming motions.
Mario Heller simulates a certain behaviour (‘doublet firings’) of motor units (a motor unit consists of a single motor neuron and all of the muscle fibres it innervates) to investigate some contractile properties in force production. He shows that motor unit doublet firings substantially increase isometric muscle force in motor unit pools, which mainly contain small, slow twitch motor units.
Some concepts and current approaches in the field of net-based analysis of biomechanical data are presented in the article by Jürgen Perl. Artificial neural networks of type Kohonen feature map (KFM) are applied to extract useful information from multidimensional motion data.
Kerstin Witte, Nico Ganter, Christian Baumgart and Christian Peham discuss how such multidimensional data sets may be reduced to lower dimensions by applying PCA techniques. They show how these techniques could successfully be applied to gait analysis, in rehabilitation and in triathlon as well as in horse riding.
In this special issue some insight into the interdisciplinary area of biomechanical modelling and simulation shall be given. I would be very pleased if it could stimulate some future research in this challenging and prosperous field.
Guest Editor
Arnold Baca
Biomechanics/Kinesiology and Applied Computer Science
University of Vienna, Aufder Schmelz 6A
A-1150, Vienna, Austria