A call for improved navigation algorithms in GNSS-IMU-based tracking data in sport science

GNSS-IMU-based player tracking data is widely used in sports and sport science to design training programs and monitor training load. Based on GNSS and IMU technologies, manufactures provide data such as distances covered in several velocity bands, number and intensity of accelerations and decelerations or summated training loads. Accuracy of that data is crucial, to ensure reasonable decisions and recommendations. However, there is limited information regarding the concurrent accuracy, reliability and validity of different devices in applied sports settings. On the one hand, this paper illustrates results of a validation study including three commercially available devices for determining speed and distance measures during a soccer-specific setting as well as other applications of GNSS-IMU-based tracking data in comparison to gold-standard reference system. On the other hand, we discuss possible reasons for inaccuracies in GNSS-IMU-based data estimations with a focus on the special needs in high dynamic movements of sport science applications, research gaps and what is necessary to overcome current limitations. We conclude that due to unknown algorithms and pattern recognition strategies, validation of MEMS-devices is necessary to ensure ecological validity for the application under field-based conditions. Additionally, algorithms and signal processing procedures should be accessible, to detect causes for inaccuracies and to enable improvement in data quality.