Introduction Nowadays, GPS devices are vital in controling training load by calculating total distance covered, distance covered in different speed zones or velocity values, thus allowing decision-making mechanism in professional soccer for sports scientists and coaches. Therefore, the accuracy of the data obtained from GPS devices is crucial, especially when data from different manufacturers are used and compared. Hence, this study investigates the accuracy of three commercial GPS devices (Catapult Sports, STATSports and Kinexon) in determining distances covered with high velocities and maximum velocity during a soccer specific setting in comparison to a 3D-motion capture system. Methods 25 youth U-19 male professional soccer players completed a soccer-specific circuit withprescribed movements and high movement intensities according to Linke et al. (2018). Each participant performed three trials in total, wearing each GPS device once. The devices were located in a pocket at the upper back as recommended by the manufactures. Speed (maximum velocity) and distance measures (distance covered in the speed Zones 14.4-19.8 km/h and 19.8-25.2 km/h) as provided by the three manufacturers algorithms were compared against a simultaneously recorded 3D-motion capture system (Qualisys AB, Göteburg, Sweden) including 14 cameras sampled at 100 Hz. Statistical data analysis included intraclass correlation coefficients (ICC) as well as Bland-Altman Limits of agreement (LoA). Results There were negative biases of all GPS devices for the high speed running distances, whereas positive biases were found for distances covered within 14.4-19.8 km/h. ICCs show poor (0.051) to excellent (0.983) accuracies for all systems during high speed running. Distances covered with lower velocities are estimated with good (0.725) to excellent (0.992) accuracy for the Catapult and STATSport systems, Kinexon shows poor (0.486) to excellent (0.930) results. Maximum velocity is underestimated by the Catapult and Kinexon system, whereas the STATSport device overestimates maximum velocity. ICC (0.977) and the respective CI (0.940-0.991) indicate excellent accuracy for the maximum velocity derived by the STATSport device only. Discussion Our results indicate meaningful differences of distance and speed parameters between all GPS systems against the camera-based reference systems. The differences range from 1.8% for maximum velocity with the STATSports system up to 36.6% for distance covered with velocities above 19.8 km/h with the Catapult system. Additionally, we find increasing errors with increasing velocities for all systems, which is in line with results from other studies (Rampinini et al. 2015; Scott et al. 2016). These findings have an important practical meaning as they underline that GPS devices commonly underestimate distances covered with velocities > 19.8 km/h. These underestimations typically lead to overestimations of distances covered with lower speed. Sport scientists and coaches should be aware of these inaccuracies when analysing GPS-derived data.
© Copyright 2024 15. Symposium der dvs-Sektion "Sportinformatik und Sporttechnologie": Zwischen Geistesakrobatik und praktischer Anwendung: Innovationen in der Sportinformatik und Sporttechnologie - Abstractband. Veröffentlicht von Technische Universität Dortmund. Alle Rechte vorbehalten.