An innovative study called "iMove" demonstrates that the integration of bioimpedance and inertial (IMU) sensors significantly improves the accuracy of fitness exercise recognition. The experiment tracked six types of upper body exercises with 10 volunteers over five days. The results show that the system using only IMU sensors achieves an average Macro F1 of 81.49%, but after including bioimpedance in the training, the accuracy increases to 84.71%.
The researchers also prove that with sensor fusion (both sensors active even during execution), an impressive 89.57% accuracy is achieved. These findings offer new perspectives for fitness devices and applications, where personalization is improved, and the accuracy and reliability of automated exercise tracking are enhanced. This is crucial for effective home workouts and remote monitoring.