Automated Ski Velocity and Jump Length Determination in Ski Jumping Based on Unobtrusive and Wearable Sensors

Abstract

Although ski jumping is a widely investigated sport, competitions and training sessions are rarely supported by state-of-the-art technology. Supporting technologies could focus on a continuous velocity determination and visualization for competitions as well as on an analysis of the velocity development and the jump length for training sessions. In the literature, there are several approaches for jump analysis. However, the majority of these approaches aim for a biomechanical analysis instead of a support system for frequent use. They do not fulfill the requirements of unobtrusiveness and usability that are necessary for a long-term application in competitions and training. In this paper, we propose an algorithm for ski velocity calculation and jump length determination based on the processing of unobtrusively obtained ski jumping data. Our algorithm is evaluated with data from eleven athletes in two different acquisitions. The results show an error of the velocity measurement at take-off of (which equals -3.0 % ± 4.7 % in reference to the estimated average take-off velocity) compared to a light barrier system. The error of the jump length compared to a video-based system is 0.8 m ± 2.9 m (which equals 0.9 % ± 3.4 % of the average jump length of the training jumps in this work). Although our proposed system does not outperform existing camera-based methods of jump length measurements at competitions, it provides an affordable and unobtrusive support for competitions and has the potential to simplify analyses in standard training.