Automatic event identification of para powerlifting bench press with a single inertial measurement unit

Maximal bench press exercise represents the main competition in Paralympic powerlifting. Several events can be defined during the barbell lifting movement that are of interest for sport performance. To date, inertial measurement units (IMUs) are considered a valid and reliable tool for obtaining parametric values of the barbell linear velocity. Although the identification of the different phases of the bench press movement is of interest for sports performance monitoring and optimization, algorithms for time event identification are still lacking. The present study proposes a single sensor-based method for the identification of key performance time events during the Paralympic bench press. The timing of events was identified from linear acceleration and velocity signals and validated against video recordings. The proposed algorithm showed an average event detection error rate of 2.8%. Average differences between measurement systems ranged between 0.02 and 0.12 s. Inter-subject variability proved to be the major constraint towards an accurate IMU-based event detection ( maximum absolute error = 0.50 s). Intraclass correlation coefficients indicated excellent to moderate test-retest reliability for most of the events. Overall, the proposed method represents a promising approach for monitoring specific aspects of the Paralympic bench press technique from which coaches and athletes could benefit to improve training protocols and sports performance.