Physical activity plays a major role in daily life of a human being. This is especially true when considering bipedal locomotion but can also encompass swimming or riding a bicycle for example. The analysis of human motion becomes particularly interesting when it has to be learned (during growth), when it is limited (in cases of disease and rehabilitation), or for the purpose of improving physical performance (as for sport). In this cumulative dissertation a modern approach, the so-called attractor method, is applied to kinematically differentiate human cyclic motion, especially running. Two theoretical papers have been published in international journals. Two additional works provide insight into the fields of application of the method and its significance for sports practice. The first theoretical work describes the kinematic components of cyclic human running motion using the attractor method. The model presented can be described by six superimposed components: first, the individual attractor, i.e., the running movement itself, and morphing, which represents continuous deviations from it during the course of the run. In addition, the transient effect, which explains fluctuations in running behavior that occur especially in the first minutes after the onset of a session. In addition, there are less strong factors such as short-term fluctuations, a control mechanism that directs the movement back to the attractor should it deviate too much from it, and finally technical noise, which is derived from the measurement technology. The corresponding analyses and validation were performed with both, measured and simulated data. In a second, continued work concerning the above-mentioned factors, the transient effect was investigated in more details. This aspect is particularly interesting from a practical point of view, since many athletes report an arrhythmic feeling at the beginning of a running session. However, this sensation should diminish after a few minutes, once the athletes have found their rhythm. With this study it was possible to objectively quantify exactly this initial feeling. In addition, the methodology also made it possible to estimate the duration of this transient phase and to check whether it is related to athletic performance and training experience. These findings provide novel information for a proper estimation of human running and thus a possible tool to steer athletic running performance. Two further application studies support the theoretical works described above in terms of its practical relevance. It is known that familiar people can often be recognized by their posture or gait pattern even from a distance. With the help of the attractor method, we created a gaitprint for each recorded runner in the first of the two application-oriented papers. The gaitprint refers to an individual attractor that is so typical and inherent that the respective person can be identified with a very high probability solely on the basis of it. Despite the strong individuality of the gaitprints, it is still possible to create a global attractor for the running movement itself. Another practical project was conducted at Northern Michigan University in Marquette (MI, USA) with the local cross-country skiing college team. The skating technique of 16 athletes was examined and the attractor method was used to identify the respective techniques, the so-called V1 or V2 style. In addition, individual technique assessments were carried out for case studies, which made it possible to obtain a differentiated impression of leg, arm and trunk movements. Overall, the findings of both studies open up new possibilities to analyze movements of cyclic sports in a cost- and effort-efficient way and to upgrade identification algorithms to the detection of movements. published