The study proposes an EMG and strain fusion patch-based muscle function assessment solution application.
To achieve this, the study used a patch-type medical device with an integrated electromyography and strain sensor manufactured by SMDsolutions Co., Ltd.
The developed fusion sensor was attached to 20 healthy adults and 20 stroke patients at 4 different locations, and strain and EMG signal data was acquired by performing 6 different movements.
Through the correlation analysis between EMG and strain signals, the timing of muscle contractions and relaxations was detected. The strain sensor demonstrated its utility by providing reliable information through spectral analysis of the induced EMG signals.
An eXtreme Gradient Boosting algorithm-based classification model was trained using the classified EMG and strain signals. The model was effectively performed with a classification accuracy of 80% in the new test set
In addition, the quality of exercise was scored by predicting muscle function through EMG and tension signals learned through personal data-based linear regression, and the degree of recovery was scored using the number of exercises, the number of goals achieved, and execution time parameters.
The proposed muscle function evaluation solution, based on EMG and strain fusion patches, holds immense potential in exercise monitoring, personalized exercise guidance, and rehabilitation services. It not only facilitates accurate muscle function assessment but also contributes to the advancement of strain sensor research in the field of rehabilitation.
