Recent Research Interests

Dr. Hamdy S. Soliman's research is centered on machine learning and neural network modeling, with emphasis on classification, association, and generalization in complex systems. His work spans a wide range of ANN models (LVQ, BP, BAM, Hopfield, ART, KSOFM, Deep Learning) applied to big data analytics, cloud computing management, intelligent sensor networks, and image processing.

His research focuses on smart and secure wireless sensor networks (SS-WSNs) for early detection of critical asynchronous events — forest fires, border intrusion, volcanic eruptions, tsunamis, and disease progression — enabling real-time prediction and intelligent decision-making. Most recent work extends to cancer subtype identification and breast cancer detection using high-dimensional RNA expression data.

He has contributed to network security with novel encryption mechanisms exceeding AES performance (ten machine code instructions per byte, resilient to super-user insider attacks), resulting in five U.S. patents and a peer wireless security protocol comparable to the CCMP US standard.

Small-scale forest-fire emulation test bed — NMT field experiment

Active Research Areas

Smart & Secure Wireless Sensor Networks (SS-WSNs)

Development of smart and secure WSNs for very early prediction of asynchronous critical events — forest fire, border intrusion, tsunami, earthquake, volcanic activity, cancer development. Neural network models are integrated into sensor nodes for intelligent local decision-making, reducing communication overhead while improving detection accuracy. Field experiments with Mica-Z motes have validated the approach in realistic New Mexico outdoor conditions.
Funded by NSF CRI ($210,000) — 2007–2010

Cancer Subtype Detection & Healthcare ML

Early detection and subtype identification of cancers (breast cancer, TCGA pan-cancer RNA expression) using machine learning. Recent work includes VGG16 transfer learning for breast cancer detection (2025), and a capacity-matched evaluation of K-means vs. Self-Organizing Maps for minority cancer subtype recovery on 10,095 samples across 19 cancer types (AI for Engineering, 2026). The latter introduced the stability–recovery–interpretability trade-off framework now applicable to high-dimensional imbalanced engineering datasets.

Network Security & Dynamic Encryption (5 US Patents)

Novel synchronous dynamic encryption system that exceeds AES performance (ten machine code instructions per byte), resilient to super-user insider attacks. This work has yielded five U.S. patents and a peer wireless security protocol to the CCMP US standard. Recent extensions include cross-Atlantic security communication and secure cellphone applications, managed by NMT Research Foundation.

Cloud Computing & Big Data Analytics

Smart cloud computing manager using neural/fuzzy systems to monitor all activities and provide early anomaly warnings based on operational history. Experimental pilot model developed under NASA $25,000 grant (2012–2013). CloudFinder system for big data on volunteered federated clouds published in IEEE Transactions on Big Data (2017).

Neural Networks for Image Compression

Hybrid neural network model (ART/Kohonen) as an image compression engine that competes with the well-known Wavelet technology. Results published in the Applied Soft Computing journal and IEEE conference proceedings, and the model is patented with the USPTO. Satellite image compression using universal codebooks has also been explored.

Research Laboratories

Smart & Secure Sensor Lab

NSF-funded ($210,000 MRI grant) laboratory for developing intelligent and secure WSNs to detect asynchronous events — forest fire, volcanic eruption, border intrusion, seismic activity. Supports undergraduate, master's, and doctoral research with interdisciplinary collaborations.

Security Lab

Secures cyber and cellphone communication. Novel encryption algorithm exceeds AES and CCMP government standards. Most recent activity includes cross-Atlantic security communication and secure cellphone apps. Managed and supported by NMT Research Foundation.

Neural Networks Lab

Applies LVQ, BP, BAM, Hopfield, ART, KSOFM, and deep learning models to classification, pattern recognition, and analysis across healthcare, security, and environmental monitoring domains.

Field Experiments

The Smart & Secure Sensor Lab has conducted multiple outdoor field experiments in New Mexico to validate sensor-based event detection in realistic conditions, using Mica-Z motes in small-scale forest-fire emulation setups and border-intrusion detection trials.

Field experiment setup

Sensor mote deployment

Border intrusion detection trial

Data collection in the field

Forest fire emulation and border intrusion detection experiments

Mica-Z mote deployment, forest-fire emulation, and border-intrusion detection field trials

Igniting the forest fire emulation test bed

Igniting the small-scale forest-fire emulation test bed with sensor mote deployment

Collaborators & Funding Agencies

NSF
National Science Foundation

NASA / NMSGC
Space Grant Consortium

Sandia National Labs
Albuquerque, NM

LANL
Los Alamos National Lab

U.S. Dept. of Defense
Information Assurance

NMT Res. Foundation
Presidential & Parke Grants

KFUPM
King Fahd Univ., Saudi Arabia

International Universities
Algeria, Tunisia, others

Selected Grants Summary

Project	Agency	Amount	Period	Role
Experimental Network of Sensors Lab (SS-WSN)	NSF CRI	$210,000	2007–2010	PI
Information Assurance Scholarship Program	U.S. DoD	$156,906	2002–2003	Co-PI
Image Compression & Network Security	NMT / Presidential	$170,000+	Ongoing	PI
Equipment for Computer Networks Lab	NSF ILI	$48,674	1992–1995	PI
Volunteer Cloud Federation (Big Data)	NASA/NMSGC	$25,000	2012–2013	Co-PI
BIG DATA + Sensors for Healthcare	NASA/NMSGC	$24,999	2014–2016	PI
Task-Grain VHLL for Data-Driven Multiprocessing	Sandia National Labs	$60,000	1991–1993	PI
Advances in ML Methods: Data Science Approach to Genetics	Los Alamos National Lab (LANL)	$91,944	—	Co-PI
Research Student Support for Patent Development (Sensors, ML & Security)	NMT Research Park / President's Office	Undisclosed	Ongoing	PI