Mobile App to Streamline the Development of Wearable Sensor-Based Exercise Biofeedback Systems: System Development...

Area of Science:

• Sports Science
• Biomedical Engineering
• Rehabilitation Technology

Background:

• Inertial measurement units (IMUs) offer objective assessment of exercise technique but face challenges in data collection and labeling.
• Personalized exercise technique classifiers are a promising solution to overcome these limitations.

Purpose of the Study:

• To develop a tablet application for automating the creation of personalized biofeedback systems using IMUs for exercise technique analysis.
• To evaluate the real-world accuracy of these individualized IMU-based systems.

Main Methods:

• A tablet app was designed to automate IMU data synchronization, processing, segmentation, labeling, feature computation, and classifier creation.
• A personalized single IMU-based system was tested on 15 volunteers performing 4 lower limb exercises.

Main Results:

• The developed app successfully automated the creation of personalized exercise biofeedback systems.
• The single IMU system achieved 89.50% accuracy, 90.00% sensitivity, and 89.00% specificity in assessing exercise technique deviations.

Conclusions:

• A novel tablet app automates the development of personalized exercise technique classification systems.
• The personalized models demonstrate excellent accuracy for cyclical, repetitive exercises, even with a single IMU.