Development of algorithm(s) to process data obtained from electrochemical glucose sensors

• Main Author: Xu, Jiayi
• Other Authors: Poenar Daniel Puiu
• Format: Thesis-Master by Coursework
• Language: English
• Published: Nanyang Technological University 2024
• Subjects: Engineering; Algorithm; Mobile application; Remove outliers; LOF; Smooth data; WMA
• Online Access: https://hdl.handle.net/10356/181796
• Institution: Nanyang Technological University

An electrochemical glucose sensor was used in a system which extracted biological signals from human interstitial fluid using microneedle technology and converted these signals into user-readable health information through efficient data processing algorithms, thus providing users with real-time and accurate glucose monitoring. In this dissertation, an algorithm was designed to process data obtained from electrochemical glucose sensors, including the preprocessing, feature processing, and linear conversion of signals. Additionally, a mobile application capable of independently performing data processing and terminal display was developed. By comparing several methods, the optimal approach for removing outliers using the Local Outlier Factor (LOF) method, smoothing data using the Weighted Moving Average (WMA) method, and calculating the maximum current value as the distance between the minimum value and the value in the first ten seconds of the minimum was identified. This algorithm, tested in a mobile application, achieved 90.129% accuracy, demonstrating the correctness and feasibility of the approach.