Integrating Different Tracking Systems in Football: Multiple Camera Semi-Automatic System, Local Position Measurement, and GPS Technologies

Buchheit M., Allen A., Poon T.K., Modonutti M., Gregson W., Di Salvo V.

During the past decade, substantial development of computer-aided tracking technology has occurred. Therefore, we aimed to provide calibration equations to allow the interchangeability of different tracking technologies used in soccer. Eighty-two highly trained soccer players (U14–U17) were monitored during training and one match.

Player activity was collected simultaneously with a semi-automatic multiple-camera (Prozone), local position measurement (LPM) technology (Inmotio), and two global positioning systems (GPSports and VX). Data were analyzed with respect to three different field dimensions (small, <30 m^2 to full-pitch, match). Variables provided by the systems were compared, and calibration equations (linear regression models) between each system were calculated for each field dimension. Most metrics differed between the 4 systems with the magnitude of the differences dependent on both pitch size and the variable of interest. Trivial-to-small between-system differences in total distance were noted. However, high-intensity running distance (>14.4 km/h) was slightly-to-moderately greater when tracked with Prozone, and accelerations were small-to-very largely greater with LPM. For most of the equations, the typical error of the estimate was of a moderate magnitude.

Interchangeability of the different tracking systems is possible with the provided equations, but care is required given their moderate typical error of the estimate.