The Validity of Real-Time Data Generated by a Wearable Microtechnology Device

Weaving D., Whitehead S., Till K., Jones B.

The purpose of this study was to investigate the validity of global positioning system (GPS) and micro-electrical-mechanical-system (MEMS) data generated in real-time via a dedicated receiver. Post-session data acted as criterion as it is used to plan the volume and intensity of future training and is downloaded directly from the device. Twenty-five professional rugby league players completed two training sessions wearing a MEMS device (Catapult S5, firmware version: 2.27). During sessions, real-time data was collected via the manufacturer receiver and dedicated software (Openfield v1.14) which was positioned outdoors at the same location for every session. GPS variables included total-, low- (0 to 3 m·s⁻¹), moderate- (3.1 to 5 m·s⁻¹), high- (5.1 to 7 m·s⁻¹), and very-high-speed (> 7.1 m·s⁻¹) distances. MEMS data included total session PlayerLoad™. When compared to post-session data, mean bias for total-, low-, moderate-, high-, and very-high-speed distances were all trivial, with the typical error of the estimate (TEE) small, small, trivial, trivial, and small respectively. Pearson correlation coefficients for total-, low-, moderate-, high-, and very-high-speed distances were nearly perfect, nearly perfect, perfect, perfect, and nearly perfect respectively. For PlayerLoad™, mean bias was trivial whilst TEE was moderate and correlation nearly perfect. Practitioners should be confident that when interpreting real-time speed-derived metrics, the data generated in real-time is comparable to that downloaded directly from the device post-session. However, practitioners should refrain from interpreting accelerometer-derived data (i.e., PlayerLoad™) or acknowledge the moderate error associated with this real-time measure.

View this research