The increasing complexity of 6G-IoT networks presents challenges in ensuring real-time trust assessment, computational efficiency, and security against adversarial threats. Existing frameworks struggle to dynamically adapt to evolving threats and high-volume data streams, leading to compromised decision reliability. This study proposes a trustworthy adaptive neural framework (TANF), an advanced deep learning-driven trust evaluation system incorporating hierarchical processing, multidomain trust layers, and holo-recursive memory (HRM) for adaptive optimization. TANF prioritizes high-trust data streams using sensory stream balancing, dynamically allocates resources through task-specific synergy layers, and enhances memory recall by integrating past, present, and predictive state representations. The simulation, conducted in Edge-IIoTset, IoT-23, and CICIDS2017, evaluated trust assessment, computational latency, scalability, and adversarial detection. TANF achieves a precision of 92. 8%, a latency reduction of 34. 5% and an adversarial detection rate of 95. 6%, outperforming ERAI, ROBUST-6G, and IMCS.

widespread adoption of Internet of Things (IoT) devices increases the need for trust management systems that adapt to dynamic conditions and maintain reliability under diverse threats. This article introduces StackTrust, a trust management framework designed for scalable and precise IoT security. The framework integrates decision trees, support vector machines (SVMs), and random forests within a logistic regression metalearner to enhance classification robustness. A central feature is the adaptive weighting mechanism, which periodically adjusts the influence of each base model according to current performance metrics. To further stabilize predictions, a logarithmic historical-trust function incorporates long-term behavioral evidence while reducing sensitivity to short-term fluctuations. The combined trust score converges to a stable equilibrium under bounded model outputs. StackTrust supports both centralized and decentralized architectures and is validated through NS-3 simulations across multiple datasets and attack scenarios. Results on 45 000 instances confirm precision, recall, and F1-scores of 0.99, with computational complexity of O(N & times; T) and O(M & times; T) to ensure efficiency for resource-constrained IoT environments.

There are serious security issues with the quick growth of IoT devices, which are increasingly essential to Industry 4.0. These gadgets frequently function in challenging environments with little energy and processing power, leaving them open to cyberattacks and making it more difficult to implement intrusion detection systems (IDS) that work. In order to address this issue, this study presents a unique feature selection algorithm based on basic statistical methods and a lightweight intrusion detection system. This methodology improves performance and cuts training time by 27-63% for a variety of classifiers. By utilizing the most discriminative features, the suggested methods lower the computational overhead and improve the detection accuracy. The IDS achieved over 99.9% accuracy, precision, recall, and F1-Score on the dataset IoTID20, with consistent performance on the NSLKDD dataset.

Brain tumors pose a severe health risk, often leading to fatal outcomes if not detected early. While most studies focus on improving classification accuracy, this research emphasizes prediction certainty, quantified through loss values. Traditional metrics like accuracy and precision do not capture confidence in predictions, which is critical for medical applications. This study establishes a correlation between lower loss values and higher prediction certainty, ensuring more reliable tumor classification. We evaluate CNN, ResNet50, XceptionNet, and a Proposed Model (VGG19 with customized classification layers) using accuracy, precision, recall, and loss. Results show that while accuracy remains comparable across models, the Proposed Model achieves the best performance (96.95 % accuracy, 0.087 loss), outperforming others in both precision and recall. These findings demonstrate that certainty-aware AI models are essential for reliable clinical decision-making. This study highlights the potential of AI to bridge the shortage of medical professionals by integrating reliable diagnostic tools in healthcare. AI-powered systems can enhance early detection and improve patient outcomes, reinforcing the need for certainty-driven AI adoption in medical imaging.

Wireless sensor networks (WSNs) are subject to distributed denial-of-service (DDoS) attacks that impact data dependability, mobility of nodes, and energy drain. The remedy to these challenges in this work is a solution based on deep learning integrated with a blockchain-aided distance-vector hop (DV-HOP) localization algorithm for reliable and secure node localization. Incorporating a blockchain ledger makes the network more trustworthy as it verifies usual and unusual system activities, whereas the DV-HOP algorithm mitigates localization inaccuracies and enhances node placement. The system is evaluated according to different performance measures like localization error, accuracy ratio, average localization error (ALE), probability of location, false positive rate (FPR), false negative rate (FNR), energy utilization, network stability, node failure rate, node recovery rate, and malicious node detection rate. Experimental results reveal improved security, accuracy, and efficiency with 17% FPR and 15% FNR, outperforming the conventional methods. This model enhances WSN performance in different environments via precise data transmission from the source to the destination. The results confirm that integrating deep learning with blockchain and DV-HOP increases network robustness, thus making WSNs more secure against security attacks while reducing energy consumption and localization accuracy. The proposed model presents a strong solution for real-world applications in wireless network environments.

Detecting cloned and impersonated profiles on online social networks (OSNs) has become an increasingly critical challenge, particularly with the proliferation of AI-generated content that closely emulates human communication patterns. Traditional identity deception detection methods are proving inadequate against adversaries who exploit large language models (LLMs) to craft syntactically accurate and semantically plausible fake profiles. This article focuses on the detection of profile cloning on LinkedIn by introducing a multi-stage, content-based detection framework that classifies profiles into four distinct categories: legitimate profiles, human-cloned profiles, LLM-generated legitimate profiles, and LLM-generated cloned profiles. The proposed framework integrates multiple analytical layers, including semantic representation learning through attention-based section embedding aggregation, linguistic style modeling using stylometric-perplexity features, anomaly scoring via cluster-based outlier detection, and ensemble classification through out-of-fold stacking. Experiments conducted on a publicly available dataset comprising 3,600 profiles demonstrate that the proposed meta-ensemble model consistently outperforms competitive baselines, achieving macro-averaged accuracy, precision, recall, and F1-scores above 96%. These results highlight the effectiveness of leveraging a combination of semantic, stylistic, and probabilistic signals to detect both human-crafted and artificial intelligence (AI)-generated impersonation attempts. Overall, this work presents a robust and scalable content-driven methodology for identity deception detection in contemporary OSNs.

The exponential growth of Industrial Internet of Things (IIoT) is a major driving force behind Industry 4.0. Besides complete automation and transformation, industrial IoT has so far created plenty of opportunities in several sectors 1.3such as smart manufacturing, energy, healthcare, smart agriculture, retail, supply chain, and transportation. However, the increased pervasiveness, reduced human involvement, resource-constrained nature of underlying IoT devices, dynamic and shared spectrum of 4G/5G communication, and reliance on the cloud for outsourced massive storage and computation bring novel security challenges and concerns. A significant challenge currently confronting the Industrial Internet of Things (IIoT) is the increasing prevalence of sophisticated IoT malware threats and attacks. To address this, the authors propose a hybrid threat intelligence framework that is not only highly scalable but also incorporates self-optimizing capabilities, enabling it to counteract a wide range of persistent cyber threats and attacks targeting IIoT systems. For a comprehensive evaluation, the authors utilized the state-of-the-art TON_IIoT dataset, which includes over 3 million instances representing various adversarial patterns and threat vectors. In addition, both standard and extended performance evaluation metrics were employed to ensure a thorough assessment. The proposed approach was also compared against several contemporary deep learning-based architectures and existing benchmark algorithms. The results indicate that the proposed method achieves superior detection accuracy, with only a minimal compromise in speed efficiency. Finally, a 10-fold cross-validation was conducted to provide an unbiased evaluation of the framework’s performance.

In the contemporary landscape of vehicular communications, the role of vehicular ad-hoc networks (VANETs) has become increasingly pivotal, transcending the capabilities of traditional mobile ad-hoc networks (MANETs). These advancements in VANETs play a critical role in enhancing traffic management systems, promoting collision prevention, bolstering road safety, and efficiently handling emergency scenarios. Modern vehicles, equipped with advanced data collection tools, accumulate extensive information encompassing vehicle health, fuel requirements, and comprehensive location histories. This rich data repository is instrumental in forecasting future destinations and facilitating timely arrangements, embodying the essence of ambient intelligence within the Internet of Things (IoT) framework. In emergency contexts, the rapid analysis of vehicle data is crucial for identifying the nearest emergency facilities. This paper proposes an innovative approach that leverages machine learning and edge computing techniques to predict vehicles' subsequent locations using large-scale data, concurrently prioritizing user privacy. We employ federated learning for processing at the network's edge and integrate a blockchain-based distributed database to ensure robust data privacy and security. The application of blockchain and federated learning in training models on expansive datasets is particularly effective in estimating the proximity to medical facilities and emergency services. Furthermore, this study introduces an optimization method to monitor vehicle speed and outlines a comprehensive attack model, along with effective protection measures.

In the era of smart sensor networks for Internet of Things (SSN-IoT), interconnected sensors and the Internet of Things (IoT) enable what was previously unattainable. These networks are comprised of strategically placed sensor nodes that are carefully planned to gather, process, and transmit data seamlessly in a variety of settings. In this research, the Comprehensive Adaptive Sensing and clustering system (CASC-Sys) is carefully and quantitatively probed in the context of SSN-IoT, with a focus on how it fits in and what effects it might have on smart cities. When we look at key performance measures, CASC-Sys is much jester than other clustering algorithms including proficient bee colony-clustering protocol (PBC-CP), Enhanced PSO-based clustering (EPSO-C), backup cluster head (BCH) clustering. Moreover, its most adept quality is that it clusters efficiently, with a time of 17.5, which is faster than PBC-CP (18.5), EPSO-C (21.25), and BCH (20.5). This expresses that CASC-Sys can quickly organize groups, which is a very crucial feature in dynamic sensor networks. Concerning network stability, CASC-Sys has a higher reaffiliation rate (RR) of 1.25 compared to PBC-CP (2.53), EPSO-C (1.58), and BCH (0.8), indicating that it ameliorates at keeping consistent connections that are necessary for data flow to continue. With only 8.88

The Internet of Underwater Things (IoUTs) connects underwater devices to communicate, sense surroundings, and transmit data. Acoustic communication faces bandwidth limitations, making underwater wireless optical communication-free space optics (UWOC-FSO) hybrid systems a promising alternative. However, maintaining sufficient power budget and signal-to-noise ratio (SNR) is a challenging task, making wavelength translation (WT) from visible to infrared (IR) at the water-fiber-air interface crucial for reliable signal transmission. In this paper, we propose an underwater wireless optical communication-single mode fiber-free space optics (UWOCSMF-FSO) hybrid link based on a photo-detection, remodulate, and forwarding (PRF) relay and intensity modulation-direct detection (IM/DD) scheme for 8 x 1-Gb/s underwater optical wireless sensor network (UWOSN). The PRF relay is installed at a remotely operated underwater vehicle (ROV) to perform WT from visible range to IR. The performance of the sensors is analyzed for different water bodies and weather conditions of underwater and free space optics channels, respectively using metrics of Bit-error rate (BER) and Quality factor (Q-factor) employing Gamma-Gamma channel model. The simulation results show that forward-error correction (FEC) target BER of 10-4 for sensors is achieved under different water bodies and weather conditions. The results obtained from this study show that the proposed UWOC-SMF-FSO hybrid link is flexible, resilient to adverse channel effects, and can be a potential candidate for implementation of high-speed long-distance future IoUTs.

In the rapidly evolving domain of the Internet of Vehicles (IoV), ensuring robust trust management, privacy, and security presents significant challenges. This article proposes a novel approach integrating generative AI (GAI) and federated learning (FL) to address these challenges. FL allows distributed learning across vehicles without the need to share data, enhancing privacy compared to centralized methods. Our approach enhances trust management by raising the level of accuracy in detecting anomalies and preserving data privacy. As a result, the effectiveness of the proposed approach in practical real-world urban settings is illustrated by comprehensive evaluations using the CityPulse dataset. The results show a 20% improvement in trust scores under normal conditions, a 92% anomaly detection accuracy, and acceptable latency despite the added security measures. Additionally, 3-D visualizations illustrate the system's robustness and scalability. This solution aligns with the objectives of 6G wireless communications, laying the groundwork for future intelligent, ultrareliable, and secure vehicular networks. Future research will focus on expanding the application of GAI and FL for real-time decision-making in large-scale IoV networks and optimizing cryptographic protocols.

Underwater environmental exploration using sensor nodes has emerged as a critical endeavor fraught with challenges such as localization errors, energy, and costs attributed to the dynamic nature of underwater environments. This paper proposes a KNN-based cost-efficient machine-learning algorithm aimed at optimizing underwater context acquisition with sensor nodes. By addressing existing localization challenges, the algorithm minimizes localization errors, energy consumption and Time costs while significantly enhancing localization accuracy to 99.98 3.88 × 10^-8 m, Reducing localization energy consumption rate 0.0045J in addition for the first time we have also computed the localization Time cost rate which is 0.06762s. we assumed that in real-time and in NS-3 simulations on the Aqua-sim model indicate communication speed at 1500m/s. This research presents an innovative and practical approach to resolving challenges associated with underwater context acquisition through sensor nodes, it offers a comprehensive understanding and emphasizes the real-time implementation of the KNN-based cost-efficient approach.

The convergence of the Internet of Things (IoT) and Software-Defined Networking (SDN) has paved the way for a new technological paradigm in Healthcare Industry 5.0. This integration addresses the complexity, heterogeneity, and dynamic nature of smart IoT devices within healthcare systems. However, it also increases the risk of cyberattacks, particularly Distributed Denial of Service (DDoS) attacks, which pose significant threats to such critical infrastructure. While Deep Learning (DL)-based intrusion detection methods have demonstrated high accuracy in detecting these attacks, their opaque decision-making process often leads to their characterization as black-box models, limiting their practical use for security analysts. To overcome these challenges, this study proposes an explainable hybrid model for DDoS attack detection in SDN-IoT-based Healthcare Industry 5.0 environments. The model combines the strengths of Convolutional Neural Networks (CNNs) for spatial feature extraction and Long Short-Term Memory (LSTM) networks for capturing temporal dependencies in network traffic. Implemented using an SDN controller, the model accurately classifies DDoS and IoT attacks while providing transparency through the SHapley Additive exPlanation (SHAP) method, which identifies the most influential features in the model’s decision-making process. Simulation results on the CICDDoS2019 and IoT Healthcare Security datasets demonstrate the model’s effectiveness, achieving detection accuracy of 99.59% and 98.12%, respectively. These findings confirm the robustness of the proposed hybrid model compared to state-of-the-art methods for detecting potential attacks in Healthcare Industry 5.0 systems.

Breast cancer poses a real and immense threat to humankind, thus a need to develop a way of diagnosing this devastating disease early, accurately, and in a simpler manner. Thus, while substantial progress has been made in developing machine learning algorithms, deep learning, and transfer learning models, issues with diagnostic accuracy and minimizing diagnostic errors persist. This paper introduces MobNAS, a model that uses MobileNetV2 and NASNetLarge to sort breast cancer images into benign, malignant, or normal classes. The study employs a multi-class classification design and uses a publicly available dataset comprising 1,578 ultrasound images, including 891 benign, 421 malignant, and 266 normal cases. By deploying MobileNetV2, it is easy to work well on devices with less computational capability than is used by NASNetLarge, which enhances its applicability and effectiveness in other tasks. The performance of the proposed MobNAS model was tested on the breast cancer image dataset, and the accuracy level achieved was 97%, the Mean Absolute Error (MAE) was 0.05, and the Matthews Correlation Coefficient (MCC) was 95%. From the findings of this research, it is evident that MobNAS can enhance diagnostic accuracy and reduce existing shortcomings in breast cancer detection.

Diabetes is a growing health concern in developing countries, causing considerable mortality rates. While machine learning (ML) approaches have been widely used to improve early detection and treatment, several studies have shown low classification accuracies due to overfitting, underfitting, and data noise. This research employs parallel and sequential ensemble ML approaches paired with feature selection techniques to boost classification accuracy. The Pima India Diabetes Data from the UCI ML Repository served as the dataset. Data preprocessing included cleaning the dataset by replacing missing values with column means and selecting highly correlated features using forward and backward selection methods. The dataset was split into two parts: training (70%), and testing (30%). Python was used for classification in Jupyter Notebook, and there were two design phases. The first phase utilized J48, Classification and Regression Tree (CART), and Decision Stump (DS) to create a random forest model. The second phase employed the same algorithms alongside sequential ensemble methods—XG Boost, AdaBoostM1, and Gradient Boosting—using an average voting algorithm for binary classification. Evaluation revealed that XG Boost, AdaBoostM1, and Gradient Boosting achieved classification accuracies of 100%, with performance metrics including F1 score, MCC, Precision, Recall, AUC-ROC, and AUC-PR all equal to 1.00, indicating reliable predictions of diabetes presence. Researchers and practitioners can leverage the predictive model developed in this work to make quick predictions of diabetes mellitus, which could save many lives.

This article aims at analyzing and comparing an adaptive algorithm-based method for improving the performance of Internet of Things (IoT) systems through simulation studies. Concentrating on active and complex scenarios, the study presents new proposals for secure and smart learning of routes, activity forecasting for nodes, link stability estimation, and flexible resource management. These methods are benchmarked against conventional algorithms to evaluate the effectiveness of the proposed solution based on routing efficiency, traffic prediction, link, resource consumption, network response time, and energy requirements. The findings are encouraging, the adaptive algorithms do improve dramatically on the standard ones making the system slower and consuming much less power. From the findings of the study it can be concluded that using adaptive algorithms in IoT can have a high impact in terms of improvement in efficiency as well as sustainability. We conclude this work by providing some directions for further research and development in the IoT field.

The growth of the internet and big data has spurred the demand for more extensive information hoarding to store and distribute information. In today's digital era, ensuring the security of data transmission is paramount. Advancements in digital technology have facilitated the proliferation of high-resolution graphics over the Internet, raising security concerns and enabling unauthorized access to sensitive data. Researchers have increasingly explored steganography as a reliable method for secure communication because it plays a crucial role in concealing and safeguarding sensitive information. This study introduces a novel and comprehensive steganography framework using the discrete cosine transform (DCT) and the deep learning algorithm, generative adversarial network. By leveraging deep learning techniques in both spatial and frequency domains, the proposed hybrid architecture offers a robust solution for applications requiring high levels of data integrity and security. While conventional steganography methods are typically classified into spatial and transform domains, extensive research and analysis demonstrate that the hybrid approach surpasses individual techniques in performance. The experimental results validate the effectiveness of the proposed steganography approach, showcasing superior visual image quality with a mean square error (MSE) of 93.30%, peak signal-to-noise ratio (PSNR) of 58.27%, root mean squared error (RMSE) of 96.10%, and structural similarity index measure (SSIM) of 94.20%, in comparison to existing leading methodologies. The proposed model achieved reconstruction accuracies of 96.2% using Xu Net and 95.7% with SR Net. By combining DCT with deep learning algorithms, the proposed approach overcomes the limitations of spatial domain methods, offering a more flexible and effective steganography solution. Furthermore, simulation results confirm that the proposed technique outperforms state-of-the-art methods across key performance metrics, including MSE, PSNR, SSIM, and RMSE.

Ensuring data security and privacy in Internet of Things (IoT) is increasingly critical due to the growing interconnectedness of devices and the sensitivity of the data they handle. This article presents a novel approach to enhancing data security in IoT through the integration of homomorphic encryption and blockchain technology. We conduct simulations using the Kaggle Smart Home Dataset to evaluate the effectiveness of our proposed methodology on smart home devices and wearable technology. Our approach not only secures data transmission but also guarantees data integrity and privacy through decentralized verification and secure aggregation techniques. Specifically, our evaluation demonstrates an encrypted data transmission rate exceeding 99.5%, a complete absence of unauthorized access instances in the simulated environment, and a verified data integrity rate of over 99.8%. Additionally, our method supports real-time processing and scalability, making it suitable for various IoT applications, including smart contract applications in IoT for privacy and security in supply chain transactions. The study highlights the robustness of combining homomorphic encryption and blockchain to protect sensitive data throughout its lifecycle.

In conventional lossless transmission (CLT)-based real-time video transmission (RTVT), the user-perceived quality of the transmitted video frames decreases significantly even when there exists 1% packet loss. To improve the quality of experience (QoE) of RTVT with lossy channels, we propose a semantic communication (SeCom)-based deep lossy transmission (DLT) paradigm by using a deep video semantic coding (DeepVSC) model to achieve end-to-end deep joint source-channel coding (DeepJSCC) in RTVT, such that the quality of the recovered video frames can be significantly improved in lossy transmission scenarios by leveraging the strong data compression and error correction capabilities of DeepVSC. We present the basic framework of DLT, compare it with the CLT system, and an illustrative test shows that DLT can recover the image when packet loss rate (PLR) is 80%, while in CLT the images failed to be reconstructed when the PLR is 10%. We also summarize the research challenges of DLT to motivate more future research efforts in this area.

Healthcare 5.0, driven by the Internet of Medical Things (IoMT), introduces transformative changes in the medical field but also exposes systems to growing cybersecurity threats. While Deep Learning (DL) offers high accuracy in attack detection, its effectiveness is often limited by data imbalance and difficulty in identifying key features dynamically. Additionally, DL models are often criticized for their lack of interpretability, as their internal decisionmaking remains obscure. To overcome these limitations, this paper presents an explainable and adaptive DL-based security framework. It integrates a Generative Adversarial Network (GAN) to balance the dataset by generating realistic samples for underrepresented attack classes, and employs Bidirectional Long Short-Term Memory (BiLSTM) to identify temporal patterns and critical features. To enhance transparency, SHapley Additive exPlanations (SHAP) and Permutation Feature Importance (PFI) are used for interpreting the model's decisions. Experiments conducted on the NSL-KDD dataset demonstrate the effectiveness of the proposed method, achieving 93.81% accuracy and an F1-score of 82.95%.