Research Projects Executed

1. Artificial Intelligence for Energy Efficiency in 5G Wireless Networks

5G wireless technology requires significant power in order to search for the best available beam pairs prior to establishing communications, leading to rapid power drain in the equipment. Dr. Memon researched the application of artificial intelligence (AI) techniques to optimize beam searching. He developed an AI-based discontinuous reception (AI-DRX) heuristic that predicts the next packet’s arrival time. This enables the radio circuitry to be turned off until the arrival of the packet, thereby saving energy. Dr. Memon’s algorithm was shown to enable a 70% improvement in energy efficiency for certain types of packets.

1) “Artificial Intelligence-Based Discontinuous Reception for Energy Saving in 5G Networks,” Electronics, 2019 

2) “Deep‐DRX: A framework for deep learning–based discontinuous reception in 5G wireless networks,” Transactions on Emerging Telecommunications Technologies, 2019.

2. Green and Emerging Wireless Networks

In this research, Dr. Memon investigated the use of radio-frequency (RF) signals as an energy source for battery-free communications. He specifically studied the use of backscatter communication (BackCom) to harvest energy from incident RF. Dr. Memon detailed various types of BackCom, including ambient BackCom (Amb-BackCom), its evolution and architecture, and its modes, including half-duplex and full-duplex. He then examined range extension and security considerations. Dr. Memon’s research provides a comprehensive guide to BackCom, enabling researchers to evaluate its suitability within the context of their particular investigations and designs.

Dr. Memon has described his findings related to his analysis of backscatter communication and emerging wireless networks in the following three peer-reviewed papers

1) “Backscatter communications: Inception of the battery-free era—A comprehensive survey,” Electronics, 2019.

2) “Ambient Backscatter Communications to Energize IoT Devices,” IETE Technical Review, 2020.

3) “Femtocell: What, Why, and How?” IJCSNS International Journal of Computer Science and Network Security, 2019.

3. Intelligent Solutions for Preventive Healthcare Systems

In this research, Dr. Memon investigated several unique, artificial intelligence-based solutions to be used for preventive healthcare, including a machine learning (ML) based architecture for a personalized glucose monitoring system (PGMS). Using both samples acquired through invasive and non-invasive methods, the ML model was trained and repeatedly refined using a unique adaptive boosting algorithm. Once trained, the model was personalized for individual patient's characteristics. Dr. Memon then analyzed the performance of the system, finding that his PGMS significantly reduced the error rate as measured against previously measured data using non-invasive glucose values. Continuing his research into ML-based solutions in preventive healthcare, Dr. Memon developed a wearable system for health monitoring using a unique event similarity search algorithm. Dr. Memon’s research has provided several intelligent solutions for use in preventive healthcare.

Dr. Memon’s research has resulted in a total of five peer-reviewed papers. These include:

1) “Adaptive Boosting Based Personalized Glucose Monitoring System (PGMS) for Non-Invasive Blood Glucose Prediction with Improved Accuracy,” Diagnostics, 2020.

2) “Personalized Non-Invasive Blood Glucose Monitor Using Machine Learning Models,” Test Engineering and Management, 2020.

3) “A CNN-based Automated Activity and Food Recognition using Wearable Sensor For Preventive Health Care,” Electronics, 2019.

4) “Accelerated Reliability Growth Test for Magnetic Resonance Imaging System Using Time-of-Flight Three-Dimensional Pulse Sequence,” Diagnostics, 2019.

5) “MRI Gradient Subsystem Accelerated Reliability Test Using Nominal Day Usages,” Test Engineering and Management, 2020.

4. Image Recognition and AI for Real-Life Applications

In this research, Dr. Memon explored the use of convolutional neural networks (CNNs) for two image recognition applications: satellite detection of ships and automated recognition of human actions. The first, he trained a CNN to recognize ships from satellite data that improved the accuracy of existing systems, particularly when operating on noisy data due to weather conditions or the presence of high waves. Dr. Memon validated this performance against open-source datasets that provided exhaustive scenarios. The second topic to which Dr. Memon applied CNNs was the recognition of human actions. After training the CNN, he measured its performance using the stanford40 dataset, obtaining an 87.3% accuracy

Dr. Memon has published three peer-reviewed papers based on his work on machine-learning techniques to improve automated image recognition. These are:

1) “Ship Detection in Satellite Imagery by Multiple Classifier Network,” IJCSNS International Journal of Computer Science and Network Security, 2019.

2) “Feature Fusion Based Human Action Recognition in Still Images,” IJCSNS International Journal of Computer Science and Network Security, 2019.

3) “Finger-vein Image Dual Contrast Enhancement and Edge Detection,” IJCSNS International Journal of Computer Science and Network Security, 2019.

4) "Multi-Path Deep CNN with Residual Inception Network for Single Image Super-Resolution" Electronics, 2021. 

Technical Projects Executed