An Algorithm for Estimating the Performance during Barbell Exercises with Acceleration Sensors

The aim of this thesis was to evaluate the possibility to estimate the performance during barbell exercises with acceleration data. Initially, it was tried to estimate the barbell velocity by using the sensor data. The barbell velocity has been found to be an indicator of exertion during the bench press or squats. In this thesis, only the bench press was examined. The first analysis showed that acceleration sensors, even with the addition of gyro sensors, are not able to estimate the barbell velocity. Based on these findings, it was tried to determine the duration of the concentric of the repetition during a set of bench press. To validate the findings, several sets were performed with changing weights and were also recorded by a motion capturing system. The experiments showed that it is possible to determine the duration of the concentric so that it is possible to draw conclusions on the performance of the athlete. Future studies could work on improving the detection of the repetitions and on the practical application of the findings. For example, a smartphone application could be developed which could give an estimate of the repetitions in reserve during a set of an exercise.