AN AUTOMATED METHOD FOR THE ESTIMATE OF VERTICAL JUMP POWER THROUGH INERTIAL MEASUREMENT UNITS

Vertical jump performance analysis allows for assessing the ability of the lower limb to generate mechanical power. The analysis performed with inertial measurement units (IMUs) is affected by inertial effects of wobbling masses. To compensate for them, an automated method was developed to estimate peak and mean concentric power based on anthropometric and time-frequency features. IMU data of 47 countermovement- (CMJ) and 50 squat- jumps (SJ) performed by 17 participants were used. Force platform data were used to obtain reference power values. Features were chosen according to the best subset regression method, devising a multiple linear regression for each estimated power parameter and jump. The regressions explained 88% and 96% variation, for CMJ peak and average power respectively, while explaining 75% and 74% of the variation for the SJ