Author(s): DEVES, M., SAURET, C.2, HAYS, A.3, FAUPIN, A.1, Institution: JAP2S, Country: FRANCE, Abstract-ID: 1291

INTRODUCTION:
Wheelchair tennis (WT) involves intermittent aerobic activity with multidirectional movements, including high-speed linear and rotational actions, and rapid accelerations [1]. Quantifying physical loading during matches is crucial for trainers to develop effective training programs. The use of onboard sensors, particularly inertial sensors, has become common for studying wheelchair movements, enabling detailed descriptions of specific movements in WT [2]. However, theres a lack of studies focusing on precise quantification of physical load during WT matches. Generic training programs may lead to poor performance and injuries due to interindividual variability within the wheelchair sports population [3]. Our study aims to propose a method for characterizing and describing locomotor activities during WT matches.
METHODS:
Nine French WT international players, comprising six men and three women, participated in our study. We recorded thirteen matches, totaling 26 sets and 213 games. Athletes utilized their own wheelchairs equipped with Xsens Inertial Measurement Units (IMUs) on the wheels and frame, enabling kinematic data collection during all games. Time series data of linear velocity, rotation velocity and radius of curvature were processed using a PAA-SAX method for data reduction and symbolic representation. A final logical pattern search on these signals identified different locomotion tasks.
RESULTS:
Across all recorded matches, players spend an average of 25 ± 14% of the time in a static state, 41 ± 10% of the time moving in a straight line, and 24 ± 9% of the time executing wide curves. Movement characteristics and intensity are measured through segmentation. During games, players engage in various types of straight lines: 433 ± 152 within 0 to 3 meters, 154 ± 54 within 3 to 8 meters, and 55 ± 22 longer than 8 meters, with respective intensities of 4 ± 3%, 7 ± 4%, and 3 ± 4% at very high intensity (> 75% of maximal acceleration). They also execute wide turns (715 ± 222) and tight turns (363 ± 82) during the match. Among wide turns, 41 ± 5% are at high rotational velocity (>= 120°/s), with an average rotation of 84.47 ± 67.65°, covering 2.14 ± 2.02 meters. Among tight turns, 66 ± 5% are executed at high rotational velocity (>= 120°/s), with an average rotation of 78.93 ± 48.38°, covering 0.55 ± 0.36 meters.
CONCLUSION:
The employed method confirms our initial hypothesis by enabling quick and simplified visualization of entire games, sets, or matches. Subsequently, this method allows us to characterize different types of activities to describe the intensity of play. Physical trainers can use this processing method to determine the external load of the activity and adapt their training accordingly. Future studies could be designed to compare the intensity of matches at different level.