Wearable multimodal optical and acoustic sensing of physiological parameters – Team 2

Overview

Dr. Kothapalli and his team are developing a low-cost wearable device that monitors physiological parameters. The device detects blood oxygen saturation, tissue metabolism, glucose in the blood, measures body temperature, and recognizes cancer cells. Our team that consists of 6 undergraduate students is tasked to help advance the device, hardware, software, machine learning algorithms, and their integration.

Objectives

• Our main objective was to develop a low-cost wearable medical device for continuous monitoring of physiological parameters.
• On the experimental side of the project, the objective was to go to the lab to develop our own delay circuit that creates a delay between firing laser and data acquisition, and to design a circuit that can fire different wavelengths alternatively.
• On the software side of the project, the objective was to augment and improve the machine learning algorithm that resides on the Raspberry Pi, which will reduce the noise in the ultrasonic data that will be read through the USB ports.

Approach

● Conduct a weekly meeting with our sponsor to get feedback
● Complete the required purchases of Lase Driver (LDP-V 80-100 V3.3), and Raspberry Pi to complete tasks.
● Frequent lab visits to work on developing circuits with teammates and sponsors.
● Different algorithms were tested such as neural networks and TensorFlow models.
● Algorithms were written on computers to train models which were then loaded onto the Raspberry Pi.

Outcomes

● A delay circuit was developed to generate a delay between firing laser and data acquisition.
● A melanoma depth of penetration graph v. estimated penetration was produced to visualize the depth of melanoma penetration of the skin.
● A melanoma data signal was generated to display a sample signal of someone that has melanoma growth.