Design and Development of Emerging Platforms for Personal Healthcare

This thesis explores the design and development of advanced sensor platforms to address specific applications in personalised healthcare, which focuses on empowering individuals to monitor and manage their health through tailored and accessible technologies. The work focuses on innovative approaches in silent speech recognition, oral health monitoring, and biomarker detection. A wearable graphene-based strain gauge was developed to classify silent speech using machine learning algorithms. Two sensor designs were evaluated: an initial Lycra-based sensor, which demonstrated the feasibility of the approach, and an improved bamboo/elastane sensor, which incorporated improved fabrication techniques to enhance sensitivity, durability, and consistency. Machine learning methods, including feature-based classifiers and convolutional neural networks, were applied to the resistance signals generated by the sensors to predict intended speech from the neck-worn strain gauges. These methods achieved high classification accuracies, demonstrating the potential for reliable silent speech recognition. For oral health, two distinct approaches were developed. The first involved a point-of-care diagnostic platform incorporating a multiplex enzymatic biosensor for the detection of salivary biomarkers linked to periodontal disease. This system combined biomarker detection with a custom potentiostat and mobile application, offering a compact and accessible tool for early disease detection. The second approach utilised machine learning and computer vision to create a smartphone app capable of classifying and segmenting common oral health conditions from user-captured images. By focusing on smartphone-based diagnostics, the app provided a scalable, cost-effective solution for at-home monitoring of oral health. The final study focused on the development of an enzymatic microwire biosensor for detecting metabolic biomarkers associated with neurodegenerative diseases. This work introduced novel strategies for addressing non-redox measurable reactions, enabling the detection of malate, fumarate, and succinate at physiologically relevant concentrations. The biosensor design incorporated features that could support potential use in minimally invasive applications with a microwire configuration to maximise surface area and efficiency. Together, these studies demonstrate the versatility and potential of emerging sensor technologies to improve accessibility, precision, and personalisation in healthcare. Personalised healthcare aims to empower individuals with tailored tools for monitoring and managing their health, reducing reliance on traditional clinical settings and supporting proactive interventions. By addressing challenges in sensor design, data analysis, and practical implementation, this thesis contributes to advancing personalised healthcare and provides a foundation for future innovation in sensor platforms and their applications.

Name dc:type.qualificationname

Doctor of Philosophy (PhD)

Level dc:type.qualificationlevel

Doctoral

Grantor dc:publisher.institution

University of Cambridge

Year dc:date.issued

2024

Author dc:creator

• Ravenscroft, Dafydd

Advisor dc:contributor.advisor

• Occhipinti, Luigi

• Sensors
• Personal Healthcare
• Graphene Strain Gauge
• Silent Speech Recognition
• Oral Healthcare
• Computer Vision
• Point-of-Care Sensing
• Krebs Cycle
• Enzymatic Sensors
• Microwire Sensors
• Succinate
• Fumarate
• Malate

Licence

• https://apollo8-f-pro.lib.cam.ac.uk/bitstreams/8a25e41d-1dd8-4a12-b6ca-52d9b6cc75c8/download
• http://purl.org/NET/rdflicense/allrightsreserved

DOI dc:identifier.doi

https://doi.org/10.17863/CAM.119470

OAI identifier oai:identifier

oai:www.repository.cam.ac.uk:1810/386107