Human Factors Principles Enhance the Design of a Machine Learning-Based Lower-Limb Exercise System for Older Adults with Knee Pain

Knee pain in older adults significantly reduces their physical function and quality of life. While lower-limb exercises are effective for treating this health issue, individuals’ access to supervised programs is often limited. To address this, our research team developed a machine learning-based lower-limb exercise system that offers video demonstrations of exercises, provides real-time movement feedback, and tracks performance and progress. Initial evaluation of the system revealed usability barriers. Consequently, we applied human factors principles to improve the system design and subsequently evaluated the usability and acceptance of the enhanced system. Ten adults (aged 60–73) with knee pain used the system to perform lower-limb exercises. Their opinions on usability and acceptance were gathered. We found that the enhanced system was perceived as both usable and acceptable. These findings underscore the necessity of integrating human factors principles into the design of such digital health applications, as this approach allows developers to create intuitive user interfaces that align with users’ preferences and expectations.