Header Information

NPRP13S-0205-200270
NPRP13S
Qatar University
Award Active
11 Apr 2021
Dr. Khalid Abualsaud
2 Year(s)
11 Oct 2023
New
Privacy-Preserving Health Monitoring System Using AI and Non-Intrusive Smart Sensors

Project Summary
Smart health technology, powered by Internet of Things (IoT) sensors and mobile devices, is rapidly transforming the care-giving industry from the perspectives of both physicians and users (patients, family of patients, etc). This proposal stresses on the need for incorporating intelligence locally into the IoT sensors so that they can truly become powerful edge nodes to complement the powerful user-smartphone devices in addition to measuring and monitoring various health data. Among the numerous electronic health (eHealth) and mobile health (mHealth) use-cases, cardiac conditions leading to hypertension, strokes, and other heart failures, is a significant problem in Qatar and the rest of the world including North America. Cardiac monitoring, in a remote, home-based setting, can offer the chance to care-givers to detect heart conditions over a prolonged period of time. This proposal will change the paradigm of monitoring of cardiac conditions by proposing a smart IoT sensor architecture. The conventional sensors monitor the electric current produced by the heart. Traditionally, ECG (electrocardiogram) machine is used to measure this. ECG is the alphabet of the heart, and physicians and other care-givers are well-trained to interpret the ECG signals. ECG can tell a doctor whether the person has heart disease such as arrhythmia. Clinical ECG machines use multiple electrodes or leads, which are placed on a patient’s body. This is intrusive because it interferes with the patient’s activities and movement. There are wearable devices, which use some algorithms to monitor ECG; but the wearable devices do not provide accurate ECG readings. In this research, for the aforementioned use-case of cardiac monitoring, we aim to introduce a type of magnetic IoT sensor, which is non-intrusive and can be used to measure magnetic field produced by the electrical current of the heart. Such magnetic sensors have intrinsic noise, which existing signal processing algorithms cannot process easily. Therefore, we propose an AI model to process the noise and convert the magnetic reading to corresponding ECG, which doctors can easily understand. Thus, we propose non-intrusive smart sensors for ECG monitoring. The ECG signal is used to train a local AI model in the user-smartphone to make an early warning system of heart diseases such as arrhythmia. The early warning output can be further improved using federated learning without sharing the sensitive health data of the patients. By doing so, the privacy of the patients is preserved also. Privacy incorporation will be made into not only the user-smartphones, referred to as edge devices, but also across the entire IoT system.
Health Monitoring; Artifical intelligence; Smart Sensors; Federated Learning; Privacy-preservation
Mhealth; smart sensors ; wban; body area network; edge computing
Applied research
2. Engineering and Technology
2.02 Electrical, Electronic, and Information Engineering
telecommunications
Yes
No
2. Engineering and Technology
2.06 Medical Engineering
Medical Engineering
No
Yes

Institution
Qatar University
Qatar
Submitting Institution
Al Ahli Hospital
Qatar
Collaborative Institution
Tennessee Technological University
United States
Collaborative Institution
Lakehead University
Canada
Collaborative Institution
Trio Investment
Qatar
Collaborative Institution

Personnel
Lead PI
Dr. Khalid Abualsaud
Qatar University
PI
Dr. Mohamed Mahmoud
Tennessee Technological University
PI
Dr. Hessa Al-Jaber
Trio Investment
PI
Dr. Zubair Fadlullah
Lakehead University
PI
Dr. Ahmed Badawy
Qatar University
PI
Dr. Elias Yaacoub
Qatar University
PI
Prof. Mohsen Guizani
Qatar University
PI
Dr. Abdurrazzak Gehani
Al Ahli Hospital
Post Doctoral Fellow
Dr. Mostafa Fouda
Idaho State University

Outputs/Outcomes
Journal Paper
A Proof-of-Concept of Ultra-Edge Smart IoT Sensor: A Continuous and Lightweight Arrhythmia Monitoring Approach
Sadman Sakib, Mostafa M. Fouda, Zubair Md. Fadlullah, Nidal Nasser, and Waleed Alasmary
ISSN:21693536
Journal Paper
Deep Learning Models for Magnetic Cardiography Edge Sensors Implementing Noise Processing and Diagnostics
Sadman Sakib, Mostafa M. Fouda, Muftah Al-Mahdawi, Attayeb Mohsen, Mikihiko Oogane, Yasuo Ando, and Zubair Md. Fadlullah
ISSN:21693536
Conference Paper
A Circuit-embedded Reservoir Computer for Smart Noise Reduction of MCG Signals
Biraj Shakya, Mostafa M. Fouda, Steve C. Chiu, and Zubair Md Fadlullah
DOI:10.1109/IoTaIS53735.2021.9628824
Conference Paper
Asynchronous Federated Learning-based ECG Analysis for Arrhythmia Detection
Sadman Sakib, Mostafa M. Fouda, Zubair Md Fadlullah, Khalid Abualsaud, Elias Yaacoub, and Mohsen Guizani
DOI:10.1109/MeditCom49071.2021.9647636
Conference Paper
Noise-Removal from Spectrally-Similar Signals Using Reservoir Computing for MCG Monitoring
Sadman Sakib, Mostafa M. Fouda, Muftah Al-Mahdawi, Attayeb Mohsen, Mikihiko Oogane, Yasuo Ando, and Zubair Md. Fadlullah
DOI:10.1109/ICC42927.2021.9500993
Journal Paper
Multidata-Owner Searchable Encryption Scheme Over Medical Cloud Data With Efficient Access Control
Sherif Abdelfattah, Mohamed Baza, Mohamed M. E. A. Mahmoud, Mostafa M. Fouda, Khalid A. Abualsaud, and Mohsen Guizani
ISSN:19379234
Conference Paper
A Lightweight Central Learning Approach for Arrhythmia Detection from ECG Signals
Abdulla Abumadi, Elias Yaacoub, and Khalid Abualsaud
DOI:10.1109/iThings-GreenCom-CPSCom-SmartData-Cybermatics53846.2021.00021