Investigating Non-Invasive Glucose Level Sensing Technologies for Diabetic Patients

Other team members: Sourav Sarker, Mohammad Tahsin Alam.

Supervisor: Dr. Muhammad Abdullah Arafat.

Conventional glucose monitoring techniques, which involve pricking the skin to draw blood, can be uncomfortable and inconvenient for diabetes patients. These methods cause pain and disrupt daily activities, leading to poor compliance. Moreover, the frequent need for testing adds to the overall burden of managing the disease. This discomfort and inconvenience highlight the need for less invasive alternatives. Consequently, there is a growing demand for innovative, patient-friendly, and most importantly non-invasive glucose level monitoring solutions that minimize discomfort and streamline diabetes management.

During this research, we explored non-invasive methods including Near-Infrared (NIR) Spectroscopy, Photoacoustic Spectroscopy (PAS), and Pulse Oximetry Method, highlighting their principles, advantages, and limitations. The primary focus of this thesis was initially to determine variations in transducer output due to molecular vibrations induced by laser pulses. However, these methods did not yield promising results. Subsequently, NIR spectroscopy combined with ultrasonic waves was applied to observe changes in the captured micro-vibrations of glucose molecules, but this approach also lacked significant consistency. In the third approach, detailed experimentation with different configurations and analysis techniques was conducted to assess the reliability and effectiveness of the Perfusion Index (PI) method for continuous glucose monitoring. This involved combining photoplethysmogram (PPG) sensors with ultrasonic sensors. The results demonstrated potential for non-invasive glucose monitoring, though further refinement and testing are required. The photoacoustic method was revisited using single and double-lasing techniques with different experimental setups. However, the results were heavily distorted, indicating the need for higher-power lasers and more accurate transducers to achieve reliable measurements.

Some of the Experimental Setups

I designed this compact 3D-printed ring to hold two lasers and an ultrasonic transducer to generate and detect photoacoustic signal

This is one of the many PCBs we designed. Particularly, this one is a lock-in amplifier made with AD630 synchronous modulator-demodulator IC

This is the schematics of the lock-in amplifier designed in Altium

We are taking data from a prototype and analyzing the result in MATLAB.

We modified a pulse oximeter as shown in this diagram to examine the effect of ultrasound on the correlation of blood glucose level with the perfusion index of the PPG signal. We were trying to reproduce the findings of E. J. Argüello-Prada and S. M. Bolaños (2023)

One of the several in-vitro setups we made to test the photoacoustic effect. We shot laser pulses on different glucose solutions and tried to find any differences in the captured acoustic signal.

Testing the acoustic effect of laser pulses shot through the side of the finger. Grounded the arm with straps to reduce noise.

Examining the impacts of ultrasound on transmitted IR signal. In the center, we placed glucose solutions of different concentrations. We also tested by putting finger in the center.

I designed and 3D printed this compact setup for photoacoustic signal generation and detection. The use of two lasers increase the signal generation capability. The placement of the ultrasonic transducer is optimized to capture maximum amount of the generated signal. The signal acquisition board utilizes an instrumentation amplifier based on AD620.

Some of the Obtained Outputs

Doing FFT in oscilloscope to observe the changes in the frequency components of captured signals.

Observing the captured signal at the ultrasonic transducer right after the application of laser pulse on the finger. Yellow waveform is the applied pulse to the laser. Pink waveform is the captured signal.

Observing the generated ultrasound standing waves. The pink waveform are the triggers carefully programmed by microcontrollers to generate the standing waves.

Recorded signals from the ultrasonic transducer for various conditions

A glimpse of the data analysis we did to find out if there were any impact of ultrasound on the perfusion index of PPG signal

Google Sites

Report abuse