Sammendrag
Due to the complexity of human locomotion, a quantitative analysis of technique in sports is often difficult. For that reason a qualitative approach is most widespread, in particular amongst practitioners. The qualitative approach offers fast and individually adjusted feedback from a technical coach. However, it is also prone to bias due to the coaches (or researchers) prior assumptions. A quantitative approach is less prone to such bias. This thesis suggests two different methods for a quantitative approach to technique analysis. Both methods are applied to a group of 6 elite cross country skiers using the V2 (or double dance) technique on a roller skiing treadmill. The methods are used to assess links between technique and performance, quantified by FIS-points. The first method was an extensive analysis of the skiers 3D movement patterns, quantified by the coordinates of 41 markers positioned on the athletes skin and equipment. These markers determined the skiers posture. A dimensional reduction technique (PCA) was used to decompose the complex, but highly redundant set of postures into a comprehensible amount of uncorrelated variables. Each of these uncorrelated variables represented multi-segment movements, which could be visualized as movements by a stick figure. Also, the center of mass (COM) of the athletes were determined by a segment model based on the markers, which enabled an assessment of the effect of postural movements to whole body movements. Normalization and weighting procedures novel to the field of sports science enabled a direct comparison of the postural movements between athletes. The second method used a much simpler approach, and consisted solely of measurements from an accelerometer and a gyroscope (both 3 axis) positioned at the athletes sacrum. The aim was to assess whether such a system could record interesting differences between athletes. If it could, the simplicity of the experimental setup, and the light weight of the sensor suggest that quantitative measurement of technique would be feasible both in regular training, and even in competition situations. Both methods proved able to identify differences in skiing technique, even in a group consisting solely of elite skiers. Some of the differences appeared to relate to the FIS-point ranking of the athletes, which suggested that these features could be important for performance. In particular, the coordination between major hip flexor musculature and vertical COM motion appeared relevant, and suggested a more beneficial utilization of potential in the best ranked skiers. A second aspect appeared to be a preference in the best ranked skiers to use a smaller lateral COM excursion, which was closely linked to the axial rotation of the pelvis during the leg push. Also included in this thesis are two appendices. Appendix A outlines the method used to obtain drift free measurements of displacements from the accelerometer and gyroscope output, and assess the accuracy of these measurements. Appendix B is included to show that sensors of similar specifications as those used in appendix A are incorporated in current marked smart phones, and investigates the possibility use smart phones as a tool for technique analysis.