This paper presents an embedded system-based solution for sensor arrays to estimate blood glucose levels from volatile organic compounds (VOCs) in a patient’s breath. Support vector machine (SVM) was trained on a general-purpose computer using an existing SVM library. A training model, optimized to achieve the most accurate results, was implemented in a microcontroller with an ATMega microprocessor. Training and testing was conducted using artificial breath that mimics known VOC footprints of high and low blood glucose levels. The embedded solution was able to correctly categorize the corresponding glucose levels of the artificial breath samples with 97.1% accuracy. The presented results make a significant contribution toward the development of a portable device for detecting blood glucose levels from a patient’s breath.
Electromagnetic interference (EMI) is often re- garded as a disturbance in electrical systems, and most research focuses on developing techniques to mitigate its impact. EMI is particularly problematic in many sensing applications, but there are circumstances when EMI can be used as the sensing tool. This paper describes a method to determine the health of the DC link capacitance in a three-phase inverter using EMI diagnostics. The proposed approach relies on changes in a capacitor's resonant frequency and impedance, which occur naturally as it ages, to predict its health. These parameters influence the conducted EMI measurements in a manner that is detectable. Through simulation and experimentation, it is demonstrated that this approach can detect changes in a capacitor's health via a machine learning algorithm that is trained to inverters EMI spectrum.
This paper presents a microcontroller-based solution to classify blood glucose levels using acetone and ethanol breath volatile organic compounds. Two metal oxide semiconductor-based chemical sensors able to detect acetone and ethanol at parts per million concentrations were used. The sensors were tested in a controlled setup with humidified air spiked with acetone and ethanol, mimicking human breath corresponding to low and high blood glucose groups. A support vector machine algorithm was trained and implemented in a microcontroller. In a real time-time test, the trained algorithm classified low and high blood glucose groups with 97% accuracy. Subsequently, a smart wristband prototype that integrates the two sensors was developed. An Arduino-based wearable microcontroller platform was used for its small formfactor and a low-power operation. The wristband is enclosed in a 3D printed housing and powered by an onboard 3.7 V 500 mAh rechargeable Li-ion battery. A smartphone app communicates with the wristband through Bluetooth, allows data visualization, and saves data in the cloud. The presented work makes a significant contribution towards the development of a wearable device for detecting blood glucose levels from a patient’s breath.
As power electronics permeate critical infrastructure in modern society, more precise and effective diagnostic methods are required to improve system reliability as well as reduce maintenance costs and unexpected failures. Prognostic and Health Management (PHM) systems that analyze changes in the electromagnetic spectrum (E-PHM) of a circuit can be implemented to determine the health of the equipment under test. This research demonstrates the use of E-PHM techniques to measure the junction temperature of a silicon carbide (SiC) MOSFET. The results show the feasibility of training machine learning algorithms to recognize this relationship and determine the junction temperature within 10 °C. This is accomplished, in situ, without interruption of device operation and without altering the system’s performance