Education
PhD, Electrical and Computer Engineering
MS, Electrical and Computer Engineering
BS, Computer Engineering
Research Interests
- Cybersecurity
- Information security
- Threat analysis
- In vivo communication
- Wireless communication
Selected Publications
- (2023) Uni-temporal Sentinel-2 imagery for wildfire detection using deep learning semantic segmentation models, Geomatics, Natural Hazards and Risk, 14:1,
- H. H. Al-Khshali and M. Ilyas, "Impact of portable executable header features on malware detection accuracy," Computers, Materials & Continua, vol. 74, no.1, pp. 153–178, 2023.
- Salam Alyassri, Muhammad Ilyas, Ali Marhoon, Oguz Bayat, Novel partial overlapped gaussian pulse multi-access system aided by data analysis, Computers and Electrical Engineering, Volume 101, 2022, 108007, ISSN 0045-7906,
- Alsanad, H.R., Sadik, A.Z., Ucan, O.N. Ucan., M. Ilyas. And O. Bayat., YOLO-V3 based real-time drone detection algorithm. Multimed Tools Appl (2022).
Teaching Courses
- Computer Networks
- Cybersecurity Law and Policy
- Mobile Applications and Security
- Discrete Structures
- Cryptography and Computer Network Security
- Intrusion Analysis and Incident Management
- Secure Systems Architectures and Mechanisms
Memberships
IEEE Senior Member
Expertise related to UN Sustainable Development Goals
In 2015, UN member states agreed to 17 global Sustainable Development Goals (SDGs) to end poverty, protect the planet and ensure prosperity for all.
This person’s work contributes towards the following SDG(s):
Modified YOLOv8x model for coronary stenosis detection and troponin risk stratification
Published in: Discover computing
Jan 20, 2026
Detection of coronary artery stenosis and risk stratification of troponin play a pivotal role in offering early diagnosis and treatment of cardiovascular diseases. In this paper, an improved deep learning framework that allows using both spatial and frequency-based attention mechanisms will be proposed using a modified YOLOv8x framework. Upon benchmarking YOLOv8, YOLOv9, and YOLOv10 models, YOLOv8x was chosen due to its excellent baseline, and the enhancement was done to make it more clinically relevant. The proposed model was found to have a precision of 0.991, a recall value of 0.960, an F1-score of 0.980, and a mAP of 0.976. These findings show significant possibilities of real-world applications. The effectiveness of the improvements is additionally validated by large-scale ablation studies, and the results overcome the problem of detecting fine lesions and disparate clinical information. The work has added value in the form of a reliable end-to-end diagnostic cardiovascular imaging and biomarker-based risk analysis.
PUF-Enabled Key-Exchange Protocol for Vehicular Ad-Hoc Networks
Published in: IEEE Transactions on Intelligent Transportation Systems
Dec 11, 2025
The Internet of Vehicles (IoV) enables data exchange among individuals, cloud resources, road infrastructures, and vehicles, interconnected through Vehicular Ad Hoc Networks (VANETs). VANETs comprise vehicles with Onboard Units (OBUs), Roadside Units (RSUs), and a Trusted Party Agent (TPA). The data transmission among these entities supports seamless interaction and collaborative traffic management. However, data transmission on public communication channels in VANETs presents significant challenges, including security, privacy, and authentication of participating entities. Although numerous key exchange and authentication protocols have been introduced to tackle these issues, many protocols remain vulnerable to various attacks, such as a vehicle, RSU, TPA impersonation, denial of service, physical cloning, and desynchronization attacks. Therefore, to address these vulnerabilities, we propose a key exchange protocol that leverages hash functions and Advanced Encryption Standard (AES) encryption. Our protocol also integrates the Physical Unclonable Function (PUF), enhancing its resistance to physical or cloning attacks. Additionally, it effectively counters threats like impersonation, session key leakage, ephemeral secret leakage, and desynchronization attacks. We validate the security and reliability of our protocol through both formal and informal analysis. Informal analysis highlights the protocol’s essential security features, while formal analysis provides robust substantiation. Performance evaluation reveals that our protocol achieves an average reduction of 35.53%, and 53.77%, in communication and computation overheads.
LSOARP: A Link Stability and Obstacle-Aware Routing Protocol for UAV Networks
Published in: Journal of Soft Computing and Data Mining
Jun 30, 2025
As using Unmanned Aerial Vehicles (UAVs) continues to grow across military, environmental, and public safety sectors, we are seeing a fast development of Flying Ad Hoc Networks (FANETs). Despite this progress, creating reliable routing protocols for UAVs remains complex because of their high mobility, constantly changing network topology, frequent link drops, and physical obstacles in the environment. Current protocols often overlook the importance of link stability and obstacle-aware navigation, which can lead to decreased performance in real-world applications. We present LSOARP: a Link Stability and Obstacle-Aware Routing Protocol customized for UAV networks. This new protocol combines Bézier-curve-based trajectory adjustments for better obstacle avoidance with a multi-criteria link evaluation that considers residual link lifetime, energy efficiency, and route availability. We model UAV movement using a realistic prediction mechanism that captures various states such as high, low, idle, and paused. Routing decisions are then made using a weighted cost function to select the most stable and energy-efficient paths, ensuring strong network performance. Simulation experiments conducted under different conditions—including varying node density, speed, pause times, and traffic loads—show that LSOARP considerably outperforms traditional protocols like RLPR and AODV. It offers higher packet delivery ratios, lower end-to-end delays, reduced energy consumption, and less control overhead. These promising results demonstrate that LSOARP is both scalable and reliable in complex UAV environments, making it a strong candidate for real-time FANET applications.
Generalizing location-centric variations to enhance contactless human activity recognition
Published in: Frontiers in Computational Neuroscience
Jun 19, 2025
Contactless Human Activity Recognition (HAR) has played a critical role in smart healthcare and elderly care homes to monitor patient behavior and detect falls or abnormal activities in real time. The effectiveness of non-invasive HAR is often hindered by location-centric variations in Channel State Information (CSI). These variations limit the ability of HAR models to generalize across new unseen cross-domain environments; for instance, a model trained in one location might not perform well in another physical location. To address this challenge, in this study, we present a novel federated learning (FL) algorithm designed to train a robust global model from local datasets in different localizations. The proposed Federated Weighted Averaging for HAR (Fed-WAHAR) algorithm mitigates location-induced disparities, including heterogeneity and non-Independent and Identically Distributed (non-IID) data distributions. Fed-WAHAR employs a dynamic weighting approach based on local models' accuracy to improve global model classification accuracy and reduce convergence time effectively. We evaluated the performance of Fed-WAHAR using various metrics, including accuracy, precision, recall, F1 score, confusion matrix, and convergence analysis. Experimental results demonstrate that Fed-WAHAR achieves an accuracy of 85% in recognizing human activities across different locations, enhancing the ability of the model to infer across new unseen locations.
A Novel Flip-Filtered Orthagonal Frequency Division Multiplexing-Based Visible Light Communication System: Peak-to-Average-Power Ratio Assessment and System Performance Improvement
Published in: Photonics
Jan 15, 2025
Filtered orthogonal frequency division multiplexing (F-OFDM), employed in visible light communication (VLC) systems, has been considered a promising technique for overcoming OFDM’s large out-of-band emissions and thus reducing bandwidth efficiency. However, due to Hermitian symmetry (HS) imposition, a challenge in VLC involves increasing power consumption and doubling inverse fast Fourier transform (IFFT/FFT) length. This paper introduces the non-Hermitian symmetry (NHS) Flip-F-OFDM technique to enhance bandwidth efficiency, reduce the peak-average-power ratio (PAPR), and lower system complexity. Compared to the traditional HS-based Flip-F-OFDM method, the proposed method achieves around 50% reduced system complexity and prevents the PAPR from increasing. Therefore, the proposed method offers more resource saving and power efficiency than traditional Flip-F-OFDM. Then, the proposed scheme is assessed with HS-free Flip-OFDM, asymmetrically clipped optical (ACO)-OFDM, and direct-current bias optical (DCO)-OFDM. Concerning bandwidth efficiency, the proposed method shows better spectral efficiency than HS-free Flip-OFDM, ACO-OFDM, and DCO-OFDM.
Malware Detection with Subspace Learning-based One-Class Classification
Published in: IEEE Access
Jun 05, 2024
Detecting malware is crucial for ensuring the security of computer systems. Traditional machine learning models face challenges in effectively detecting malware, mainly due to the class imbalance problem, where the number of malware samples is significantly smaller than that of non-malware samples. Additionally, malware’s dynamic and evolving nature, continuously altering its structure and tactics, presents a substantial challenge for conventional artificial intelligence algorithms, further complicating the detection task. In pursuing an optimized malware detection technique, researchers initially explored traditional machine learning algorithms, focusing on the features of Portable Executable (PE) file headers. However, the inherent issues, such as imbalanced datasets and the deceptive nature of malware, have raised concerns about the credibility of the attained results. This can result in misclassifying malware as non-malware, leading to security vulnerabilities. One-Class Classification (OCC) methods have emerged as a promising approach to improve the detection of unknown malware. However, traditional OCC approaches face the challenge of the curse of dimensionality. This research proposes adapting subspace learning-based OCC methods to overcome the curse of dimensionality and effectively handle the class imbalance problem. We propose a pipeline for detecting malware using methods that jointly optimize a subspace and data description for OCC. We evaluate the performance of various one-class classifiers on three different datasets. We observed that the subspace-learning-based OCC is a promising approach. Evaluating various classifiers on three datasets reveals promising results, with a True Positive Rate (TPR) of 100% for subspace-learning-based OCC. The proposed pipeline can serve as a valuable tool for improving the security of computer systems by accurately detecting malware and protecting against potential attacks.
Enhancing spectral efficiency with low complexity filtered-orthogonal frequency division multiplexing in visible light communication system
Published in: ETRI
Mar 22, 2024
Hayder S. R. Hujijo
Muhammad Ilyas
The filtered-orthogonal frequency division multiplexing (F-OFDM) scheme has gained attention as a promising solution in the field of visible light communication (VLC) systems. One crucial aspect in VLC is the conversion of the complex F-OFDM signal into a real signal that corresponds with direct detection and intensity modulation. Traditionally, achieving a real F-OFDM signal has involved imposing Hermitian symmetry (HS) on the samples of the Inverse Fast Fourier transform (IFFT), which requires a 2N-point IFFT and obtains an N-point FFT, thus adding complexity. In this study, a novel approach is presented and implemented, aiming to enhance spectral efficiency and reduce system complexity by generating a real F-OFDM signal without relying on HS. This approach is then compared with HS-free (HSF)-OFDM, direct current biased optical OFDM, and asymmetrically clipped optical OFDM. The suggested method offers a remarkable improvement of ~50% in the required IFFT/FFT volume. Consequently, this method reduces hardware complexity and power usage compared with the traditional F-OFDM method. Moreover, regarding error rates, the proposed method demonstrates better spectral efficiency than HSF-OFDM.
Uni-temporal Sentinel-2 imagery for wildfire detection using deep learning semantic segmentation models
Published in: Geomatics, Natural Hazards and Risk
Apr 05, 2023
Wildfires are common disasters that have long-lasting climate effects and serious ecological, social, and economic effects due to climate change. Since Earth observation (EO) satellites were launched into space, remote sensing (RS) has become a more efficient technique that can be used in agriculture, environmental protection, geological exploration, and wildfires. The increasing number of EO satellites orbiting the earth provides huge amounts of data, such as Sentinel-2 with its Multi Spectral Instrument (MSI) sensor. Using uni-temporal Sentinel-2 imagery, we proposed a workflow based on deep learning (DL) semantic segmentation models to detect wildfires. In particular, we created a new big wildfire dataset suitable for semantic segmentation models. We tested our dataset using DL models such as U-Net, LinkNet, DeepLabV3+, U-Net++, and Attention ResU-Net. The results are analyzed and compared in terms of the F1 score, the intersection over union (IoU) score, the precision and recall metrics, and the amount of training time each model possesses. The best results were achieved using U-Net with the ResNet50 encoder, with an F1-score of 98.78% and an IoU of 97.38%, and we developed it into a pre-trained DL Package (DLPK) model that is able to detect and monitor the wildfire from Sentinel-2 images automatically.
Impact of portable executable header features on malware detection accuracy
Published in: Computers, Materials & Continua
Sep 22, 2022
One aspect of cybersecurity incorporates the study of Portable Executable (PE) file maleficence. Artificial Intelligence (AI) can be employed in such studies, since AI has the ability to discriminate benign from malicious files. In this study, an exclusive set of 29 features was collected from trusted implementations; this set was used as a baseline to analyze the presented work in this research. A Decision Tree (DT) and Neural Network Multi-Layer Perceptron (NN-MLPC) algorithms were utilized during this work. Both algorithms were chosen after testing a few diverse procedures. This work implements a method of subgrouping features to answer questions such as, which feature has a positive impact on accuracy when added? Is it possible to determine a reliable feature set to distinguish a malicious PE file from a benign one? When combining features, would it have any effect on malware detection accuracy in a PE file? Results obtained using the proposed method were improved and carried few observations. Generally, the obtained results had practical and numerical parts. For the practical part, the number of features and which features are included are the main factors impacting the calculated accuracy. Also, the combination of features is as crucial in these calculations. Numerical results included finding accuracies with enhanced values; for example, NN_MLPC attained 0.979 and 0.98, and for DT, an accuracy of 0.9825 and 0.986 was attained.
Novel partial overlapped gaussian pulse multi-access system aided by data analysis
Published in: Computers and Electrical Engineering
Jul 01, 2022
Orthogonal frequency-division multi-access (OFDMA) systems have limited flexibility to improve efficiency due to their dependency on subcarrier orthogonality. As a result of this restriction, attention has shifted to a new multi-access communication method. The popularity of non-orthogonal multi-access (NOMA) systems is growing. Because the NOMA systems may broadcast and receive signals at various power levels, more complicated reception devices are required. Partially overlapped subcarriers or a non-orthogonal multi-access system are presented in this study. Instead of relying on the power level of sending signals, as is the case in present NOMA systems, the proposal relies on the benefit of modifying the shape of subcarriers to build a more efficient system. The authors propose Gaussian-pulse signals as an alternative to sites of concern for nature conservation (SINC) (SINC is the shape of subcarriers in OFDM). In this paper, there are several algorithms designed for this work. These algorithms control the distribution of frames on the transmitting subcarriers. They also calculate the width of the subcarriers as well as the spacing between the subcarriers to produce the lowest possible data error rate. The similarity values between the frames that will be sent will influence the values generated by these algorithms. So, these algorithms are to reduce the computational complexity of the system and obtain efficient channel capacity. The proposed model presents encouraging results for the bit error rate (BER) compared with OFDMA and ordinary NOMA systems. Also, Gaussian pulses with data analysis, as in the proposed schema, can achieve a reduction value in spectrum requirements by up to 13.8%. Besides, there is a decrease in out-of-band compared to OFDMA, which increases the spectrum efficiency. Finally, as compared to OFDMA, an improvement in BER with multipath fading and a Doppler frequency shifting environment was discovered in this research.
YOLO-V3 based real-time drone detection algorithm
Published in: Multimedia Tools and Applications
Mar 26, 2022
Drones are currently being used in a wide range of useful tasks that are too dangerous and/or expensive to be performed by humans. However, this is increasingly developing security breaching issues due to the possibility of misuse of unmanned aircraft in illegal activities such as drug smuggling, terrorism, etc. Thus, the detection and tracking of drones are becoming crucial topics. Unfortunately, due to the drone’s small size, its detection methods are generally unreliable: high false alarm rate, low accuracy rate, and low detection speed are well-known aspects of this detection. The new emerging real-time algorithm based on the improved “You Only Look Once” (YOLO-V3) algorithm is proposed here for drone detection. This newly designed algorithm comprises multiple phases and has shown the potential to outperform the traditional detection approaches. The proposed algorithm enhances the performance of YOLO-V3 by designing and building a CNN to solve the problem of a large number of YOLO-V3 parameters, using densely connected modules to enhance the interlayer connection of CNNs and further strengthen the connection between dense neural network blocks, and finally improving the YOLO-V3 multiple-scale detection by expanding the three-scale to four-scale detection to increase the accuracy of detecting small objects like drones. The evaluation results of our algorithm obtain 96% average precision and 95.60% accuracy.