Comparison of algorithms to determine jump height and flight time from body mounted accelerometers

The purpose of this study was to evaluate a new procedure of data processing from a body-mounted accelerometer to improve the assessment of vertical jump height. As the main difficulty when using an accelerometer is to detect the times of take-off and of landing, a new criterion was proposed to detect these times more accurately. Thirty physical education students participated in this study. They performed three squat jumps, three countermovement jumps with hands placed on the pelvis and three countermovement jumps with free arms ($$n = 270$$n=270 jumps). Flight time, vertical jump height and vertical velocity at take-off were collected from the accelerometer (the Myotest device and its specific software), the modified accelerometer (raw data of the Myotest and a specific treatment for the detection take-off and landing times) and a force platform, considered as the reference device. Concerning the flight time, systematic bias decreased from $$0.034 \pm 0.079$$0.034±0.079 s with the original accelerometer to $$-0.008 \pm 0.078$$-0.008±0.078 s with the modified accelerometer. Regarding the vertical jump height, systematic bias decreased from $$4.8 \pm 9.4$$4.8±9.4 cm with the original accelerometer to $$-1.3 \pm 9.2$$-1.3±9.2 cm with the modified accelerometer. Finally, the vertical velocity systematic bias was $$-0.12 \pm 0.28$$-0.12±0.28 and $$-0.19 \pm 0.29$$-0.19±0.29 m s$$^{-1}$$-1 with the original accelerometer and the modified accelerometer, respectively. This study showed that the improvement proposed for the body-mounted accelerometer decreased the systematic bias, especially for the vertical jump height assessment, but not the random error.