In front crawl swimming, Seifert et al. (2014) explained that the variability of behaviours might be functional in enhancing performance at different levels (e.g. intra-cyclic, inter-cyclic and inter-individual level). If the energy cost should be minimized in swimming by maintaining a movement speed as stable as possible, it was observed that different coordination modes can be used. Therefore, computing coordination indexes can help to evaluate the intra-cyclic variability of behaviours. As a result, understanding and minimising the intra-cycle speed variations becomes relevant because it may indicate a high continuity between propulsive actions (showed in front crawl, Seifert et al., (2014)). Practically, coordination indexes have been proposed to attest coordination (crawl: Chollet et al. 2000; breaststroke: Chollet & Seifert, 2004; butterfly: Chollet et al. 2006; backstroke: Chollet et al. 2008). Such indexes are based on the simple collection of some relevant spatio-temporal parameters of each stroke cycle to obtain insights on the swimmer`s behaviour. Those spatio-temporal and kinematical parameters are generally collected with a video system or an optoelectronical camera-based system, considered as a gold standard to analyse movement in swimming (Seifert et al. 2015). However, in an aquatic environment, these systems suffer from a number of technological limitations (restricted field of view) and limitations in terms of adaptability to the environment (light refraction, effects of water turbulence, water clarity, parallax effect and problems of contrast and distortion), making data collection limited to 3 to 4 swimming cycles per sequence, which is not enough to register an entire 50m pool length or even a 25m swimming race. To overcome these limitations, IMUs (inertial measurement units) have become popular for analysing swimmers (James et al. 2011). An IMU is an electronic system consisting of a 3D accelerometer, a 3D gyroscope and a 3D magnetometer. These sensors respectively measure linear accelerations, angular velocities and the magnetic field to estimate the speed, orientation and position of the swimmer in the water, offering the scientific community new measurement possibilities. The present contribution seeks to propose an algorithm that would automatically detect the key points of the butterfly technique, in the manner of Dadashi et al. (2013) and Guignard et al. (2017) who analyzed front crawl with IMUs. We hypothesize that the kinematic similarities in the sagittal plane of upper limb movements between crawl and butterfly could allow for a fine-grained study of butterfly stroke phases and hence butterfly coordination based on the existing literature.
© Copyright 2023 XIVth International Symposium on Biomechanics and Medicine in Swimming Proceedings. Veröffentlicht von evoletics Media. Alle Rechte vorbehalten.