Non-orthogonal multiple access (NOMA) has been presented as an alternative technology due to its ability to accommodate immense connectivity and enhances spectral efficiency in 5G and future wireless networks. NOMA, in contrast to orthogonal systems, allows numerous users to share the same radio resource at the same time, breaking the orthogonality of traditional multiple access. However, power allocation is a critical challenge in designing an effective NOMA system. In this work, the authors investigated fair power allocation and considered it pertinent to improve the scheme further. To ensure user fairness, we proposed an improved fair power algorithm that can be modified dynamically based on target rate requirements and channel state information. The simulation results show that the proposed improved fair power allocation performs better than other traditional schemes.

Anomaly detection is crucial in modern data-driven applications such as healthcare, industrial monitoring, and cybersecurity, where traditional methods often face challenges with complex, high-dimensional data. Autoencoders, a form of unsupervised neural network that learns compact data representations, have become valuable tools for identifying anomalies through reconstruction errors. This paper reviews a range of autoencoder architectures used for anomaly detection, carefully evaluating their advantages, limitations, and applicability across different types of data and anomaly patterns. The discussion begins with basic models including vanilla, sparse, and denoising autoencoders, before moving on to more advanced forms such as variational, adversarial, recurrent, and attention-based variants. Key design considerations like the depth of encoder and decoder networks, the structure of latent spaces, and regularization techniques are examined to support effective model development. Real-world applications, including fraud detection and IoT fault diagnosis, are highlighted, along with challenges related to computational demands and robustness in the face of adversarial interference and streaming data. Standard performance metrics like precision, F1-score, and AUC-ROC are covered, together with reconstruction error measures such as mean squared error and mean absolute error, to evaluate detection accuracy. Finally, the review identifies ongoing challenges, including the selection of decision thresholds, interpretability of results, and risks of overfitting, and proposes future research directions focused on hybrid models, self-supervised learning, and improved explainability to enhance the deployment of autoencoder-based anomaly detection systems.

Machine-learning-based intrusion detection for Internet of Things systems has often been evaluated through model-centered pipelines that use weakly governed partitioning, limited leakage auditing, and closed-set assumptions. Consequently, reported performance could reflect data-handling artifacts rather than reliable security intelligence. This paper introduces a scientific integrity framework that treats preprocessing as a primary research object for open-set Internet of Things intrusion detection. The framework integrated device-disjoint split governance, feasibility-aware zero-day isolation, quantified leakage control, train-only preprocessing, shared-safe feature selection, diagnostic-harness verification, baseline split comparison, and auditable artifact generation. Applied to the CICIoT-DIAD 2024 corpus with Institute of Electrical and Electronics Engineers Organizationally Unique Identifier-based vendor enrichment, the protocol locked 28 canonical classes, eight semantic attack families, and five policy labels before constructing a device-disjoint, vendor-aware grouped split. When strict device-level zero-day holdout was infeasible, the framework activated an audited row-level fallback that preserved contamination-free holdout isolation without claiming strict device-novel zero-day evaluation. On 35,672,407 flows from 180 files, the accepted run achieved zero device overlap, zero flow-signature Jaccard leakage risk, 100 percent zero-day purity, a Feature Distribution Stability Score of 0.00518, a Device-Feature Dependency Index of 0.00000, an Attack Invariance Score of 0.92964, and an Attack Semantic Consistency Score of 0.90714. The diagnostic harness produced zero hard failures and zero warnings, while baseline comparison showed stronger preprocessing integrity than random stratified and simple device-disjoint splitting. This study did not claim downstream classifier superiority; rather, it established an auditable preprocessing substrate for later classifier-level experiments.

The exponential growth of Enhanced Machine-Type Communication (eMTC) in 5G and beyond networks is fundamental to the evolving Internet of Things (IoT) landscape, connecting billions of low-power devices. This rapid growth, while transformative, significantly broadens the attack surface, introducing critical security challenges that traditional measures struggle to address. This study proposes an advanced anomaly detection framework leveraging Convolutional Autoencoders (Conv AEs) to identify unusual patterns and potential security threats in eMTC-enabled IoT data traffic. The Conv AE model was developed and rigorously evaluated using the CICIoT2023 dataset, which comprises over 46 million records of both normal and anomalous network traffic. Through binary stratified sampling and z-score normalization, a balanced and representative dataset was prepared for training and evaluation. The model was trained exclusively on normal traffic to learn benign behavior. The Conv AE architecture, comprising an encoder with 1D convolutional layers and a decoder with 1D transposed convolutions, optimizes a Mean Squared Error (MSE) loss function to reconstruct input data, flagging anomalies based on high reconstruction errors exceeding an empirically selected threshold of 0.0262. The evaluation metrics demonstrate the framework\'s exceptional effectiveness and efficiency. The model achieved an impressive Accuracy of 0.9892, a Precision of 0.9987, a Recall (True Positive Rate) of 0.9903, and an F1-Score of 0.9945. These results indicate a strong ability to detect anomalous events with remarkably few false positives. Furthermore, the model exhibited efficient operational performance, with a per-sample inference time of 0.0809 seconds and the ability to process a full test set of 102,578 samples in 52.87 seconds, enabling real-time anomaly detection. The False Alarm Rate (FAR) was 0.0530, and the True Negative Rate (TNR) was 0.9470, confirming the model\'s proficiency in distinguishing normal from malicious traffic. The high Kappa Score of 0.8000, Matthews Correlation Coefficient (MCC) of 0.8096, Geometric Mean (G-Mean) of 0.9684, AUC-ROC of 0.9895, and AUC-PR of 0.9996 further validate the model\'s balanced and robust performance. This research contributes a robust and efficient solution for enhancing the security and safe deployment of IoT systems within the eMTC framework on current and future generation networks. Future research may adapt this binary framework to a multi-class classification approach, enabling classification of attack types for more detailed threat insight in eMTC-enabled IoT applications on 5G and beyond networks.