Asthma Monitoring using IoT and Edge Technology

Project Title

Project Area of Specialization

Artificial Intelligence

Project Summary

To control the escalating cost of treating chronic diseases such as the asthma, it is required that they need to self managed by the patients. The main goal of this project is to detect the severity levels of Asthmatic Patients, and people suffering from seasonal allergies. This can be done by designing filters that detect the respiratory sounds of a patient and then use that data to get further insights into the disease's condition and eventually, predict its severity. Data is acquired from various sources, such as the heart rate sensor, microphone, and a peak flow graph etc. Using some already availabe datasets of the Asthmatic patients, suitable machine learning models can be built to predict the severity and using edge computing, results can be displayed in real-time.

Project Objectives

To provide a low cost, effective, portable and immediate health monitoring system
To help doctors monitor asthma patients.
To help doctors understand if their treatment is helping reduce the severity of the disease.
And lastly to have a positive impact on the lives of asthmatic patients.

Project Implementation Method

It takes in the respiratory sound of the patient as input through a mobile device. Then our algorithm processes and analyzes the signal.
We then use signal processing techniques to filter out the required sound (wheeze, cough, crackle etc)
Lung Sound Analysis; Exploit the different features of respiratory sound signal to extract useful information.
Feed the extracted features to a machine learning model to classify the severity level as ‘High’, ‘Low’ or ‘Medium’.

Benefits of the Project

Low cost
Portable
Quick

Technical Details of Final Deliverable

An Android/iOS Mobile app capable of running on all smartphones and iphones, having the capability of interacting with the patient, recording their respiratory sounds, and giving back the results in real time
An edge device based on a raspberry pi performing the same functionality
The results include a severity level prediction, a few precautions to take, and a few suggestions to overcome the current condition