Waveform-level agreement between global positioning system and sacral inertial measurement unit acceleration during overground running in male elite rugby athletes.

Abstract

Thoracic-mounted global positioning system (GPS) units embed accelerometers to characterise locomotor mechanics in applied sport. However, it's unclear whether acceleration waveforms captured by these devices preserve the signal structure of sacral-mounted inertial measurement units (IMUs), which are used as biomechanical references for gait characteristics. This study evaluated waveform-level similarity and between-session reliability of resultant acceleration from thoracic GPS and sacral IMUs during running. Ten elite rugby union athletes completed two 4-min steady-state running sessions at individualised sub-maximal speeds. Triaxial acceleration was recorded simultaneously using thoracic-mounted GPS (100 Hz) and sacral-mounted IMU (downsampled to 100 Hz) devices. Resultant acceleration waveforms were temporally aligned, standardised, and compared using cosine similarity. Between-session reliability was assessed using Bland-Altman analysis. Waveform similarity was high across sessions (Session 1: 0.90 ± 0.02; Session 2: 0.89 ± 0.02). Mean absolute error (0.019) and root mean squared error (0.022) were trivial. Between-session bias was negligible (-0.004), with narrow limits of agreement (-0.048 to 0.040). Thoracic-mounted GPS accelerometers preserve the waveform structure of sacral-derived resultant acceleration during steady-state overground running. These findings support using GPS-embedded accelerometry for whole-waveform analyses of running mechanics in applied environments, while highlighting the need for further validation under more demanding conditions.