A pilot study for detecting release instant using a single inertia measurement unit in shot put

Abstract

Evaluating and understanding the release parameters is crucial in the throwing events of athletics, however, calculating the release parameters is time-consuming for data collection and can be expensive. The present study aimed to investigate a method for detecting the release instant from data collected using a single inertia measurement unit (IMU) in shot put. Two male shot putters participated in the study. Each participant performed six competitive throws with a 9-axis IMU (100 Hz) attached to the back of the participant’s throwing hand. Three different methods were examined from IMU data: threshold values of resultant acceleration, waveform of resultant acceleration, and waveform of the angular velocity around the radio-ulnar axis of the hand. The release instant as a true value was obtained from the video recorded in sync with the IMU, and the error of the release instant detected from the IMU was calculated. The final number of trials analysed was four for participant A and six for participant B. As results, there were various cases where the threshold value of acceleration did not exceed the value depending on the participant and trial, and the release instant could not be detected. The release instant detected from the moment the acceleration decreased based on the acceleration waveform had a large error (4.00 ± 2.26 frames). The release instant, which was detected from the moment when the angular velocity became negative based on the angular velocity waveform, was detected in all trials for both participants, and the error was low (1.20 ± 0.92 frame).It was found that utilizing the detection of dorsiflexion timing of the wrist joint, with the angular velocity around the rotational axis parallel to the radio-ulnar axis of the hand as a cue, proved to be a highly accurate approach for determining the release instant.