The rapid proliferation of Internet of Things (IoT) devices has introduced significant security challenges, particularly in ensuring the reliability and resilience of RFID systems, a critical component of IoT architectures. Among these challenges, denial of service (DoS) attacks pose a significant threat, potentially disrupting operations and compromising sensitive data. This study proposes a comprehensive framework that integrates lightweight cryptographic techniques and machine learning (ML)-based algorithms for enhanced RFID security. The research begins with an analysis of IoT security challenges, focusing on the vulnerabilities of RFID systems and the need for resource-efficient cryptographic solutions. Lightweight cryptography (LWC) is employed to ensure secure and efficient data encryption, while ML-based authentication and anomaly detection algorithms are utilized to identify unauthorized access and mitigate DoS attacks. The proposed system is modeled and evaluated using a layered RFID architecture and a detailed flowchart that incorporates data decryption, feature extraction, and real-time anomaly detection. By leveraging state-of-the-art ML algorithms such as Isolation Forest, the framework ensures robust authentication and anomaly detection capabilities. The results demonstrate the feasibility of combining lightweight cryptography with ML techniques to address IoT security challenges effectively, providing a scalable and resource-efficient solution for safeguarding RFID systems against evolving cyber threats.

Denial-of-Service (DoS) attacks continue to threaten the availability and dependability of Internet of Things (IoT) networks. Because many IoT devices have limited processing power, memory, and built-in protection, they are frequently exploited to generate abnormal traffic that blocks legitimate communication. Signature-driven security tools are often unable to cope with changing attack behaviour. This paper presents a hybrid intrusion detection model that combines Artificial Neural Networks (ANNs) and Random Forest (RF) classification for DoS detection in IoT networks. The ANN component learns non-linear traffic representations, while the RF component performs the final ensemble-based classification. The model was implemented in MATLAB and evaluated using detection accuracy, false positive rate (FPR), and latency. The results show that the hybrid ANN-RF model reached 93% detection accuracy and 5% FPR, outperforming standalone ANN and RF models. The findings indicate that the proposed approach can support reliable and scalable real-time intrusion detection for IoT applications such as smart homes, healthcare systems, and industrial automation.

The increasing prevalence of Man-in-the-Middle (MITM) attacks poses a significant threat to the security and privacy of data transmitted over public Wi-Fi networks. This study introduces a novel MITM attack detection mechanism that leverages the combined strengths of Support Vector Machines (SVMs) and Convolutional Neural Networks (CNNs). Existing machine learning approaches for detecting MITM attacks often struggle with limited accuracy and high false positive rates. To address these limitations, this research proposes a hybrid SVM-CNN model. CNNs excel at extracting intricate patterns from network traffic data, while SVMs are highly effective for classification tasks. By integrating these complementary strengths, the SVM-CNN model effectively identifies subtle anomalies indicative of MITM attacks. Simulation results demonstrate the superior performance of the SVM-CNN model compared to standalone SVM and CNN models. The SVM-CNN achieved a significantly higher detection accuracy (93

Accurate crop-yield prediction is increasingly important for strengthening food security, improving farm planning, and supporting efficient resource allocation under changing climatic conditions. This study presents a comprehensive review of recent machine learning algorithms for crop-yield prediction in smart agriculture. A structured literature search identified 24 studies published between July 2024 and March 2025, of which 11 met the final review criteria. The selected studies were analysed according to the algorithms applied, input features, evaluation metrics, predictive performance, practical limitations and future research opportunities. The findings show that Random Forest (RF) was the most frequently used algorithm, accounting for 16

The Internet of Things (IoT) is a crucial driver of the fourth industrial revolution, enabling devices to operate intelligently without human intervention. IoT devices, including sensors, smart devices, and RFID systems, collect and transmit data over the Internet. However, IoT security remains a major challenge due to the vast number of connected devices and their limited computing power, making them vulnerable to cyberattacks. Traditional cryptographic security measures often fail to fully address IoT security needs. Given these limitations, machine learning (ML) techniques have emerged as a promising solution to enhance IoT security by embedding intelligence into devices and networks. ML-based approaches can detect and mitigate security threats more effectively than conventional methods, offering new research opportunities. This paper reviews various IoT attacks, explores ML techniques to counter these threats, and provides recommendations for future research in securing IoT networks.

Water Distribution Networks (WDNs) frequently experience significant water losses due to pipeline leakages. These losses not only create economic challenges for water utilities but also intensify global concerns regarding water scarcity. This study aims to enhance the accuracy and reliability of leak detection and localisation within WDN infrastructures. Traditional leak detection techniques often exhibit limitations such as high operational costs, inefficient detectionprocesses, and susceptibility to false alarms, particularly when sensors are deployed randomly across the network. Furthermore, detecting concealed or low-intensity leaks remains a difficult task. To address these challenges, this study introduces a hybrid supervised machine learning framework that combines Support Vector Machines (SVM), Artificial Neural Networks (ANN), and Graph Theory (GT). The integration of these techniques enables the proposed model to analyse multiple parameters influencing leak behaviour and improve the reliability of detection outcomes. The hybrid model, referred to as the SVM-ANN-GT algorithm, is evaluated using the EPANET hydraulic simulation environment and compared with conventional machine learning approaches. Experimental results indicate that the proposed hybrid model significantly improves leak detection performance. The model achieves an average detection accuracy of approximately 96%, outperforming standalone SVM and ANN models, which achieved accuracies of 85% and 80%, respectively. The improved performance is primarily attributed to the integration of graph-theoretic optimisation for sensor placement, which enhances monitoring coverage and reduces redundancy within the network.

Water leakage in Water Distribution Networks (WDNs) remains a critical challenge, particularly in developing regions where Non-Revenue Water levels are high. This paper presents an optimisation approach for real-time water leak detection using a hybrid Machine Learning (ML) model, KM-LightGBM, deployed in an edge computing environment and supported by LoRa-based communication. The proposed framework integrates K-Means clustering for data preprocessing and dimensionality reduction with Light Gradient Boosting Machine (LightGBM) for accurate leak classification. The system processes sensor data including flow rate, pressure, and temperature collected from distributed IoT nodes, enabling efficient local decision-making at the edge gateway. The performance of the proposed model was evaluated using simulations conducted in EPANET 2.2, configured to model a realistic water distribution network consisting of 25 nodes, a pump station, and a storage tank under varying operational and environmental conditions. Experimental results demonstrate that the KM-LightGBM model achieves superior performance compared to standalone methods. It attained an average accuracy of approximately 97% with a peak of 98%, while recall consistently exceeded 95%, reaching up to 98%, indicating strong leak detection capability. Furthermore, the proposed model achieved a significantly reduced latency of 10-12 ms, outperforming K-Means (up to 26 ms) and LightGBM (around 18 ms). These results confirm that the hybrid approach enhances both detection accuracy and computational efficiency. The integration of edge computing and LoRa communication further reduces transmission delay and supports scalable real-time monitoring. The proposed system demonstrates strong potential for deployment in smart water infrastructure to improve sustainability and reduce water losses.

Water leakage in distribution networks poses significant challenges, including resource wastage and increased operational costs. This paper presents the ANN-XGBoost algorithm, an innovative approach that combines Artificial Neural Networks (ANN) with Extreme Gradient Boosting (XGBoost) to enhance water leak detection in water distribution networks (WDNs). By leveraging multi-dimensional sensor data, specifically pressure, flow rate, and temperature, the algorithm achieves high detection accuracy and reduces false alarms. The simulation was conducted using EPANET, a widely recognized software for modeling water distribution systems. Simulation results demonstrate that the ANN-XGBoost model outperforms traditional models, achieving a detection accuracy of up to 98%, while the individual ANN and XGBoost models reached maximum accuracies of 95% and 89%, respectively. Precision metrics further indicate significant improvements, with the ANN-XGBoost model achieving a precision of 85% compared to 74% and 80% for ANN and XGBoost, respectively. Additionally, the ANN-XGBoost model achieved a False Positive Rate (FPR) of only 5%, significantly lower than the 12% and 10% rates observed in the ANN and XGBoost models, respectively. The ANN-XGBoost algorithm effectively integrates Machine Learning (ML) techniques with IoT technologies, facilitating real-time monitoring and automated responses to potential leaks.

The rapid development of smart sensors has enabled the integration of wireless sensor networks (WSNs) with the Internet. The emergence of the Internet of Things (IoT) represents a significant paradigm shift, with WSNs remaining essential to its framework. Most smart sensor nodes will serve as foundational components, which can cause network congestion. Congestion is a critical factor affecting the performance of IoT networks. The Constrained Application Protocol (CoAP) has been developed by the Internet Engineering Task Force (IETF) for IoT communications. CoAP is a lightweight, request/response-based RESTful protocol designed for resource-constrained nodes. Resource-constrained networks often face limitations in hardware and link capabilities, leading to challenges such as packet loss and increased delay. This review examines CoAP, highlighting its distinct fundamental features that facilitate deployment in constrained IoT environments. This paper provides an overview of WSNs as foundational elements for IoT applications, followed by a discussion of low-power wide area networks (LPWANs) and their standardisation as promising technologies for these applications. This paper discusses prominent state-of-the-art mechanisms for congestion control (CC) in CoAP, highlighting several research challenges and future directions. For controlling congestion in IoT resource-constrained networks, it suggests a three-state Markov model and machine learning-based strategies.

This study explores the application of JAYA optimization algorithms to significantly enhance the performance of indoor optical wireless communication (OWC) systems. By strategically optimizing photo-signal parameters, the system was able to improve signal distribution and reception within a confined space using circular and randomly positioned diffuse spots. The primary objective was to maximize signal-to-noise ratio (SNR) and minimize delay spread (DS), two critical factors that affect transmission quality in OWC systems. Given the challenges posed by background noise and multipath dispersion, an effective optimization strategy was essential to ensure robust signal integrity at the receiver end. Key achievements of JAYA optimization include significant performance gains, such as a 29% improvement in SNR, enhancing signal clarity and reception, and a 23.3% reduction in delay spread, ensuring stable and efficient transmission. System stability also improved, with the standard deviation of SNR improving by up to 5%, leading to a more consistent performance, while the standard deviation of delay spread improved by up to 9.9%, minimizing variations across receivers. Resilience against environmental challenges: Optimization proved effective even in the presence of ambient light noise and complex multipath dispersion effects, reinforcing its adaptability in real-world applications. The findings of this study confirm that JAYA optimization algorithms offer a powerful solution for overcoming noise and dispersion issues in indoor OWC systems, leading to more reliable and high-quality optical wireless communications. These results underscore the importance of algorithmic precision in enhancing system performance, paving the way for further advancements in indoor optical networking technologies.

The shift to smart grids requires strong cybersecurity measures to protect critical infrastructure from evolving threats. The integration of automation systems increases vulnerability to cyberattacks, particularly targeting Supervisory Control and Data Acquisition (SCADA) systems. Ensuring secure communication and data integrity is essential for maintaining the reliability and resilience of smart grids. This study introduces a novel cybersecurity algorithm for Smart Grid SCADA systems, addressing modern cyber threats. The algorithm enhances authentication, availability, data integrity, and confidentiality while minimizing computational overhead. It integrates N-th degree Truncated polynomial Ring Units (NTRU) and Toom-Cook algorithms for optimized security. The proposed Enhanced Cybersecurity (ECS) algorithm was evaluated using Network Simulator Version-2 (NS-2) against Elliptic Curve Cryptography (ECC) and Rivest-Shamir-Adleman (RSA). Results show high-speed encryption (1.4 ms) and decryption (2.9 ms) with minimal processing resource use. Under DoS attacks, ECS experienced only 0.5

IPv6 stacks for wireless devices with low hardware resources have enabled many new applications and protocols. IETF's Constrained Application Protocol (CoAP) allows constrained devices with Internet access to manipulate resources. Due to restricted radio channel capacities and hardware resources, constrained device networks often experience congestion. This study looks at three congestion control (CC) protocols: the Markov model, CoAP Simple Congestion Control / Advanced (CoCoA) and Default CoAP and compares how well they handle data flow, lost packets, and data flow over different time periods. The proposed Markov model incorporates state transition probabilities to dynamically adjust to network conditions and effectively control varying congestion levels. The Markov model outperforms CoCoA and Default CoAP using Contiki OS and the Cooja simulator. It maintains a high throughput of 220.729 kbps and 13.637 % packet loss at a transmission interval of 15, with a higher CDF of throughput at 0.8, especially under increasing network congestion. The study highlights the use of proactive CC, which employs Markov state transitions to forecast transmission behaviours. In contrast, CoCoA and Default CoAP use reactive approaches, which perform poorly during network congestion. Our findings suggest that predictive models, such as the Markov model, offer substantial benefits for resource-constrained networks like the IoT, where efficient use of available resources is crucial.

The concept of Wireless Sensor Networks (WSNs) has generated significant attention on a global scale. According to the IEEE 802.11 standard, WSNs are wireless networks composed of sensor nodes (SNs) that interconnect with one another via wireless links. SNs can sense, process, and communicate wirelessly in a harsh environment. WSNs are primarily used for traffic communication in a wide range of environments such as Smart Cities, healthcare, smart homes, and military areas. The implementation of WSNs in various environments such as Smart City comes with different challenges like traffic congestion. The traffic congestion that occurs in the roads within a Smart City during peak hours is normally caused by numerous vehicles traveling at the same time. As a result, traffic jams, time delay, and accidents occur during traffic congestion. This paper proposes the Prioritybased Congestion Control (PBCC) algorithm which is designed by integrating the Fusion-based Intelligent Traffic Congestion Control (FITCC) algorithm and Traffic Congestion Detection (TCD) in order to reduce the on-road traffic congestion within the Smart City. In the future, simulation results will be presented to show the superiority of the proposed algorithm.

Cloud computing eliminates the need for expensive hardware and software expenditures by revolutionising access to computing resources through internet-based utility services. However, Data Integrity (DI) in this paradigm faces a variety of challenging issues, including complexity, security, privacy, control limits, fallibility of human beings, and financial limitations. The shortcomings of current DI solutions in terms of guaranteeing data verification, preventing replay attacks, and controlling computational overhead have led to an increasing need for access to cloud infrastructures by third-party verifiers. The suggested Cryptographic Accumulator Provable Data Possession with Merkle Hash Tree (CAPDP-MHT) scheme demonstrates significantly improved performance over Provable Data Possession (PDP) and Rivest Shamir Adleman (RSA) algorithms in various domains, as demonstrated by thorough simulation and MATLAB-based evaluation. In particular, CAPDP-MHT outperforms PDP and RSA with an average data verification success rate of 25%, compared to their respective rates of 10% and 5%. Moreover, it identifies replay attacks in about 30 seconds, compared to 45 and 70 seconds for PDP and RSA, respectively. Furthermore, the computational overhead of CAPDP-MHT is about 27 seconds, while that of PDP and RSA is 45 and 60 seconds, respectively. Therefore, as compared to PDP and RSA-based systems, CAPDP-MHT not only exhibits exceptional computing efficiency but also outperforms in reliability.

As organizations transition their ICT infrastructure to the cloud, this provides a new way of accessing data on ICT platforms from anywhere and at any time. The cloud offers users infrastructure, applications, and storage services that require protection through policies or procedures. Cloud security is important in securing user data and infrastructure from malicious users by ensuring confidentiality, integrity, and availability. An intrusion detection system's (IDS) primary goal is to detect fraudulent activities to secure user data and cloud services. This study conducted a systematic literature review based on peer-reviewed empirical studies to identify and evaluate technology solutions for addressing cloud computing cyberattacks. This study provides an in-depth understanding of cloud computing and intrusion detection methods, defining precise classification and standards for algorithm construction in this dynamic research field. This study examines the limitations of current intrusion detection techniques in cloud computing, categorizing them into distinct groups through a literature review. The study highlights the benefits of feature selection-based intrusion detection in cloud computing, highlighting vulnerabilities and providing insights for stakeholders, offering valuable resources for further exploration. Finally, the paper concludes with a summary and outlines potential avenues for future research.

Overcoming the digital divide in rural and remote areas of the Republic of South Africa (RSA) has been a challenging and daunting. This is because of the country's vast geographically landscape. As of 2023, only 70% of South Africans had reliable internet access. The COVID-19 pandemic has further worsened this gap, as education, business, government services were conducted online. The need for internet has risen significantly as the country is embracing the potential of Information and Communication Technology (ICT) as a stepping stone to economic and social development. However, the traditional way of deploying broadband is limited by the prohibitively expensive nature of extending high-capacity fibre and microwave backhaul to remote districts, making many business cases unviable for terrestrial operators. As a result, this paper examines the role of Non-Terrestrial Networks (NTNs) specifically Low Earth Orbit (LEO) satellites in bridging this digital divide. Furthermore, this paper examines two promising LEO satellite-based solutions. The first solution leverages LEO constellations as a backhaul for current 5G terrestrial networks. The other solution leverages direct-to-direct (D2D) LEO services to provide low-latency Internet access in remote and underserved areas. This paper further presents the challenges that are slowing down the adoption of LEO, including the regulatory barriers and high deployment costs. The recommendations to expedite LEO adoption and integration into 5G networks are also highlighted. Integrating 5G infrastructure sharing with LEO satellite networks reduces deployment costs, improves rural broadband coverage, and guides policy reforms that promote equitable access and efficient spectrum use in South Africa. This study enhances technical understanding of LEO deployment and provides a strategic reference for policymakers, researchers, and industry leaders working to bridge the rural digital divide.

Abstract - The Internet of Things (IoT) represents a vast network of interconnected devices equipped with software, sensors, and other technologies that enable data exchange and autonomous operation with other devices and systems without human intervention over the internet. IoT applications span across various sectors, including agriculture, education, healthcare, and communication. However, Distributed Denial of Service (DDoS) attacks continue to pose significant risks to the IoT network due to current challenges of classification efficiency and response times by the existing algorithms, such as Decision Tree (DT), Linear Regression (LR), and K-means. This paper provides a comprehensive review of DDoS attack types within the IoT networks. Secondly, the paper critically examines and analyses the challenges and opportunities inherent in leveraging Machine Learning (ML) algorithms for detecting, preventing, and mitigating these attacks. Finally, it presents the categories of IoT performance metrics, and their statistics found in the Literature over the Past decade.

The building sector accounts for a significant portion of the global energy consumption and carbon dioxide (CO2) emissions, making it a critical area for improving energy efficiency. In Africa, the rapid energy demand and costs have further emphasized the urgency of developing effective solutions for reducing building energy use. This paper presents a comprehensive review of data-driven and physics-based modeling approaches for forecasting and optimizing energy performance in non-domestic buildings. The review highlights the evolution of statistical models, classical machine learning methods, deep learning, and hybrid approaches across various application scenarios. Emphasis is placed on the role of data pre-processing techniques, including data fusion and transfer learning, as strategies to address data limitations and improve model generalization. Furthermore, the study evaluates the strengths and limitations of different modeling methods in terms of accuracy, scalability, and applicability in real-world contexts. By integrating insights from recent literature, this paper identifies key research gaps such as the need for standard datasets, physics-informed hybrid modeling, and policy-oriented frameworks. The findings aim to guide building managers, policymakers, and researchers toward adopting robust data-driven solutions that enhance energy resilience, reduce operational costs, and support environmental sustainability in the built environment. The review also justifies the importance of these models for practical applications like energy benchmarking, retrofit planning, and CO2 reduction, providing a clear link between research and industry implementation.

The concept of Wireless Sensor Networks (WSNs) has garnered significant global attention due to their wide range of applications. According to the IEEE 802.11 standard, WSNs are wireless networks consisting of sensor nodes (SNs) that interconnect via wireless communication links. These SNs are capable of sensing, processing, and wirelessly transmitting data, even in challenging environments. WSNs are primarily utilized for communication in various domains, including smart cities, healthcare, residential areas, and military applications. However, the deployment of WSNs in environments such as smart cities come with some challenges, particularly traffic congestion. Traffic congestion in smart cities, particularly during peak hours, is often caused by a high volume of vehicles traveling at the same time, resulting in delays, accidents, and inefficiencies under various weather conditions, including sunny and rainy days. Therefore, this paper proposes a novel algorithm called the Fusion-based Intelligent Congestion Management (FICM) algorithm, developed through the integration of the Navigation Reference Spatial Data (NRSD) algorithm and Fusion-based Multimodal Abnormal Detection (FMAD) algorithm. The objective of FICM is to mitigate on-road traffic congestion within smart cities effectively. The algorithm’s performance was evaluated using Network Simulator 3 (NS-3) by comparing its effectiveness with the NRSD and FMAD algorithms. Under sunny weather conditions, the NS-3 simulation results revealed that the FICM algorithm achieved an average False Alarm Rate (FAR) of 0.83%, a Mean Time to Detection (MTTD) of 76.0%, and a Detection Rate (DR) of 84.3%, outperforming both the NRSD and FMAD algorithms. Similarly, under rainy weather conditions, the FICM algorithm demonstrated an average FAR of 14.09%, an MTTD of 57.03%, and a DR of 78.04%, surpassing the performance of the NRSD and FMAD algorithms within the smart city environment.

Livestock theft is a significant threat to the agricultural industry, necessitating innovative preventive strategies. This study proposes a Wireless Livestock Tracking System (WLTS) that uses real-time Internet of Things (IoT) technologies to prevent livestock theft. The WLTS integrates GPS sensors with Long Range Radio (LoRa) wireless communication modules, overcoming the limitations of Wi-Fi and Bluetooth-based systems. It uses a single LoRa network receiver to facilitate real-time communication between farmers and their livestock. Simulation results show the WLTS effectively mitigates livestock theft, enabling farmers to quickly identify and recover stolen animals. Geofencing alerts enhance the system's sensitivity to potential theft scenarios. The WLTS has a user-friendly interface, allowing farmers to remotely monitor their livestock. Data analytics capabilities enable predictive analysis of probable theft trends based on historical data. The findings pave the way for practical implementation, revolutionizing livestock protection and safeguarding farmers' livelihoods worldwide.