Assessment method of indoor localization system for high-performance sports analysis

Since a decade indoor localization based on radio-frequency (RF) systems are developed because of the cost, the time consuming deployment and postprocessing that make devices easier to use than computer vision. The latter was commonly used in the early days of athlete training analysis to measure workload, performance and for strategy purposes. Indoor localization is a well known topic and it is largely study for static and pedestrian scenarios. While sports are a more challenging context requiring high accuracy systems (few tens of centimeters) evaluated in dynamic scenarios. This paper proposes a method to assess the performances of a positioning system in relevant scenarios for high-level sports, i.e., in dynamic situations, following a trajectory with different speeds and accelerations. This method evaluates the deviation between the athlete's position relative to the trajectory, bringing out the impact of speed and acceleration on dynamic accuracy. We apply this method to an Ultra-Wide Band (UWB) system which is a technology used for indoor sports, results confirm that dynamic scenarios cannot be neglected (dynamic accuracy is 20cm versus 10cm for static accuracy).