Abstract
For ALS patients, communication is more than a convenience, it is survival. A well‐designed assistive device can mean the difference between isolation and expression for patients facing serious disabilities. My capstone project develops a BiPAP integrated blink detection system so that ALS patients that use noninvasive ventilation can communicate, even when their eyes are occluded by a BiPAP mask. I chose to do this project to learn how to integrate machine learning into mechanical engineering as well as do an interdisciplinary project that benefits the medical world practically. Concurrently, my STS research explores socioeconomic factors influencing equitable access to assistive technology for ALS patients and for healthcare in general. Understanding these sociotechnical barriers is essential because technology is meaningless if inaccessible to the populations that most require them. These projects are interconnected through their design elements, where the need for the product is apparent, but it does not just look at technical factors. There are economic factors involved such as material, size, or how easily integrable it can be, so that most patients can access it.
My capstone project targets the challenge posed by BiPAP masks, which currently limit ALS patients' access to eye-tracking communication devices, due to the occlusion of the mask over the eyes. By integrating a small, lightweight camera directly onto the mask, I developed a real-time blink-detection system using a machine learning algorithm called a convolutional neural network (CNN). The prototype features an adjustable, 3D-printed mount compatible with various mask models, ensuring adaptability and patient comfort.
The testing of the blink-detection system had an accuracy of 90% with minimal latency under varying environmental conditions like lighting and eye shape. Future steps involve trials involving healthy participants and feedback from caregivers to improve the system's potential effectiveness in emergency situations. It is also to relieve caregivers of a constant burden to watch their patients. The technical report concludes that integrating blink detection into BiPAP masks is feasible, effective, and potentially transformative for ALS patients' communication and quality of life.
My STS paper poses the research question: What sociotechnical factors influence the equitable distribution and accessibility of assistive communication technologies for ALS patients? This question is significant and relevant since medical technologies often fail to reach those who need them most due to economic and systemic barriers. The methodology involved applying the Social Construction of Technology (SCOT) and Actor-Network Theory (ANT) frameworks to analyze healthcare policies, patient demographics, and medical device design.
Evidence collected through interviews, policy analyses, and case studies reveals systemic challenges such as high cost, insurance not covering everything whether it is private or public, limited provider training, and profit over healthcare. My findings show that technical prowess is not the only important factor, but systemic alignment involving policy reform, clinician education, and stakeholders. My STS paper concludes that using innovations like the BiPAP-integrated blink-detection system must concurrently address these broad sociotechnical barriers to truly enhance patient care.
