KINESIS obtains biomechanical features using a 3D camera. Information from other sensors such as heart rate monitors is also fused to obtain a biomechanical profile of the skier. Then, using two types of machine learning algorithms, poling phases and recovery phases of the double poling technique is inferred. Moreover, real-time feedback is provided when the skier reaches a selected reference movement, for example when the current movement performance is close to a reference previous training.
The development of KINESIS is supported by the Umeå Sports Science School and is being currently tested with the elite cross-country skiing team of Umeå University.
Correlation between biomechanical analysis provided by KINESIS and physiological aspects (oxygen consumption, heart rate, lactate threshold, etc.) will be investigated to find new lines of research that can be further explored in a continuation of the collaboration between researchers in computing science and sports science beyond this project.