The aim of this study was to analyze the stability of the aiming process of elite biathlon athletes. Nine elite athletes performed four series of five shots onto the same target and onto targets next to each other in a shooting hall. A video-based system reconstructed the horizontal and vertical motion of the muzzle. The time period starting after repeating the rifle and ending with the shot was divided in 10 intervals of equal duration. Eight kinematic parameters describing the motion in these intervals were calculated. Based on the parameter values obtained a special variant of an artificial network of type SOM (self-organizing map) was trained. Similar neurons were combined to clusters. For each shot the 10 data sets describing the aiming process were then mapped to the corresponding neurons. The sequence of the related clusters in the respective succession was used as representation of the complex aiming motion. In a second processing step types of shots were identified applying a second net. A more stable pattern could be inferred for the members of the national squad compared to the biathletes classified in the next best performance level. Only small differences between the two shooting conditions could be observed.
Copyright © 2010 Elsevier B.V. All rights reserved.