Loading...

Designing and Fabrication of a Wearable Colorimetric Biosensor Based on Polyvinyl Alcohol Hydrogel with the Help of Machine Learning to Detect Glucose Concentration and PH Level of Body Sweat

Chenani, Hossein | 2024

9 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 56934 (07)
  4. University: Sharif University of Technology
  5. Department: Materials Science and Engineering
  6. Advisor(s): Simchi, Abdolreza; Ekrami, Aliakbar
  7. Abstract:
  8. Simultaneous detection of pH and glucose levels in sweat is a promising tool for early skin disease detection and diabetes diagnosis. Hydrogels have attracted a lot of attention in the field of wearable biosensors due to sweat absorption capability, biocompatibility, and the ability to store enzymes and reagents in their cavities. In this research, we have presented an innovative wearable colorimetric biosensor based on polyvinyl alcohol (PVA) hydrogel, which can simultaneously measure pH level (in the range of 3-9) and glucose concentration (in the range of 0.025-0.5 mM) of sweat with the help of machine learning (ML). This wearable sensor consists of two main parts: sensors and a patch, in the preparation of the sensors, pure PVA hydrogel is used due to its excellent absorption and swelling, and in the production of the labels, PVA/sucrose hydrogel is used, which has high flexibility and adhesion strength. Next, by adding colorimetric reagents including red cabbage anthocyanin (RCA), phenol red (PR) and potassium iodide (KI), respectively, pH level detection sensors in the range of 3-6 and 9-6 and glucose sensor in the range of 0.025 -0.5 mM were produced and the limit of detection (LOD) of glucose was 22 μM. With the help of random forest (RF) machine learning model, these sensors will be able to achieve correlation coefficient (R2) of 0.9973 and 0.9980 for quantitative detection of pH in the mentioned ranges by RCA and PR. The detection of glucose level was done using the deep learning regression model of convolutional neural network (CNN), which achieved a correlation coefficient of 0.9849. Finally, an Android software has been developed using these selected models, showing that the developed sensor and software work well on an athlete. This research paves the way for the development of reliable and efficient biosensing devices with applications in healthcare and beyond
  9. Keywords:
  10. Hydrogel ; Poly Vinyl Alcohol ; Biosensor ; Colorimetric ; PH Level ; Glucose Concentration ; Artificial Intelligence ; Machine Learning ; Deep Learning

 Digital Object List

 Bookmark

No TOC