A hybrid approach to product prototype usability testing based on surface EMG images and convolutional neural network...

Area of Science:

• Human Factors Engineering
• Biomedical Engineering
• Ergonomics

Background:

• Office workers frequently report muscle fatigue when resting in reclined office chairs.
• Assessing physical comfort in supine positions is crucial for product design and usability.

Purpose of the Study:

• To investigate physical factors influencing comfort in a supine office chair position.
• To test the efficacy of a newly designed product prototype using a hybrid approach.
• To develop a novel method for product usability testing.

Main Methods:

• A hybrid approach combining subjective questionnaires (body mapping, impact comfort scale) and surface electromyography (sEMG) measurements.
• Utilized deep learning algorithms, specifically Convolutional Neural Networks (CNNs), for data analysis.
• sEMG signals were pre-processed, and feature maps were generated using Mean Power Frequency (MPF) for analysis.

Main Results:

• Subjective assessments identified 10 body parts significantly impacting comfort, with the neck showing the highest effect.
• sEMG measurements indicated reduced sternocleidomastoid (SCM) fatigue in the experimental group compared to the control, suggesting improved comfort.
• The developed CNN model achieved high accuracy (0.99) in classifying datasets.

Conclusions:

• The hybrid methodology is effective for studying physical comfort in supine sitting positions.
• This approach can validate the comfort of similar products and inform future design iterations.
• The study demonstrates the potential of integrating sEMG and deep learning in human factors engineering.