Development and Characterization of Ultrasensitive Carbon Nanotube Paper-Based Capacitive Sensors for Human-Machine Interfaces and Machine Learning-Enhanced Eye Tracking

Abstract

Capacitive sensors are playing an important role in human-machine interfaces (HMI) for their simplicity, speed, and robustness. However, in wearable applications, traditional metal capacitive sensors’ proximity sensitivity is limited by their small size. In this dissertation, I present my research on studying the proximity sensitivity of a capacitive sensor made of carbon-nanotube-paper-composite (CPC) and its application in HMI and wearable eye-trackers. CPC capacitive sensors had conductive cellulose fibers whose high aspect ratio structures increased the electric field strength and total surface area, leading to a higher capacitive sensitivity. The effect of cellulose fiber geometries on sensor sensitivity was studied with finite element analysis. The proximity and pressure sensitivity of CPC capacitive sensors were characterized and benchmarked against identically sized silver sensors. The CPC sensor provides a highly-sensitive, multi-modal sensing solution for wearable devices and other HMI applications. A non-human primate capacitive eye-tracker used four CPC capacitive sensors that sensed the displacement changes at the scleral-corneal junction during eye movements. The capacitive eye-tracker was benchmarked against the gold standard scleral search coil on two rhesus monkeys, showing a high signal correlation. Various data processing methods and machine learning techniques were studied to optimize signal quality, achieving gaze detection accuracy comparable to top-end commercial solutions while having superior portability. This lightweight, non-invasive capacitive eye tracker offered potential as an alternative to the traditional coil and camera-based systems in oculomotor research and vision science. CPC capacitive sensors were further developed into a cylindrical format in human eye trackers. This sensor format allowed easier integration, minimum vision obstruction, and better safety. This cylindrical format also exhibited a superior capacitive response compared to identically sized copper sensors. Sensor placements were optimized through eye area 3D scanning, allowing the capacitive eye-tracker to achieve a high accuracy over a wide gaze angle range. The capacitive eye-tracker was also explored for fatigue monitoring using blink rate and eye closure duration biomarkers, showing a low deviation from manual counting and computer vision algorithms. The compact wearable capacitive eye-tracker allowed continuous gaze and fatigue tracking throughout the day, enabling unique applications for fatigue monitoring, cognitive research, and medical diagnostics. In summary, the dissertation covered the design, fabrication, characterization, and eye-tracking applications of CPC capacitive sensors. The high proximity sensitivity at a small size could be ideal for wearable sensing applications.