Detection of Sets and Repetitions in Strength Exercises Using IMU-Based Wristband Wearables

This study presents a model for classifying four strength exercises (chest, back, biceps, and triceps). The system is built on the data collected from an IMU sensor embedded in a wearable wristband device. We gathered data from multiple sessions involving three users while they performed these exercises on fitness machines in real-world scenarios. A Deep Learning model incorporating CNN and LSTM networks was developed to classify these exercises and estimate repetition progressions accurately. The performance results are highly promising, suggesting the feasibility of integrating such exercise recognition capabilities into both mobile and wearable devices.