Over the last few years, cloud computing has emerged as the best option for offering various applications. It can supply databases, web services, processing, storage, development platforms, and web services to help businesses swiftly expand their infrastructure and service offerings. However, massive amounts of data will severely burden the cloud computing environment. Due to this, load-balanced task scheduling has remained a crucial aspect of resource distribution from a data center, ensuring that each virtual machine (VM) has a balanced load to fulfill its full potential. Overloading or underloading a host or server can cause issues with processing speed or even cause a system crash. To prevent this, we need an intelligent way to schedule tasks. Therefore, the hybrid optimization algorithm called gazelle coati optimization algorithm (GCOA) is introduced in this paper to schedule tasks in a cloud environment. This algorithm integrates the coati optimization algorithm (COA) and the gazelle optimization algorithm (GOA) to enhance the GOA's exploitation process. The main objective of this hybrid approach is to optimize scheduling, maximize VM throughput and resource utilization, and establish load balancing between VMs based on makespan, energy, and cost. The performance assessment of the proposed approach is conducted on two real-world workloads, such as Google Cloud Jobs (GoCJ) and the heterogeneous computing scheduling problems (HCSP) datasets, using several performance metrics, and the results are compared with the previous scheduling and load balancing methods. The experiment results show that the suggested strategy produced significant gains in makespan, energy, cost, resource utilization, and throughput—up to 10% and 60%, respectively—making it appropriate for real-world cloud infrastructures.
Future network technology provides a new stage for industrial production systems, especially smart manufacturing system (SMS), which is a fully integrated and collaborative factory platform that responds in real time to meet changing demands and conditions in the factory. In this paper, we propose a new spectrum allocation scheme for the macro/small cell overlaid SMS. To improve the communication performance, we develop a new solution concept, called the multi-criteria bargaining solution ( MCBS ), according to the combination of multi-criteria decision strategy and cooperative bargaining ideas. Through an interactive two-step manner, the main feature of MCBS is to harness the full synergy of heterogeneous cell coexisting infrastructure while maximizing mutual advantages in the two-tier cellular network. In an indoor smart factory environment, our hierarchical approach can effectively handle the multi-agent multi-criteria resource sharing problem. Finally, the extensive simulation results are presented to illustrate the potential advantages of our spectrum sharing policy. Especially, our proposed approach increases the network throughput, service payoff, and operator fairness by about 10%, 10%, and 20%, respectively, than the existing baseline protocols.
The scalability, adaptability, and pay-per-use nature of cloud computing have contributed to its meteoric rise to prominence, enabling customers to access services regardless of their physical location. A major obstacle to effective resource management is the wide variety of services provided and the wide variety of user needs. Due to inadequate resource use and suboptimal scheduling tactics, cloud data centers, which consist of physical machines (PMs) hosting many virtual machines (VMs), frequently experience significant energy consumption. A task scheduling technique is introduced in this study to tackle energy efficiency in cloud environments. It integrates two meta-heuristic algorithms. A Slack-based classification algorithm is used first to cluster tasks and then rank them according to their criticality. In order to schedule vital work, we use the Remora Optimization Algorithm (ROA). For noncritical jobs, we use Particle Swarm Optimization (PSO). Researchers tested several configurations ofVMs and job counts in an experimental setting and then compared the outcomes to those of more conventional approaches like Genetic Algorithm (GA) and baseline PSO. The proposed approach shows promise as an efficient scheduling approach for environmentally conscious cloud computing, thanks to its substantial reductions in execution time and energy consumption. Evaluations were carried out in a simulated cloud environment, incorporating different task counts and VM configurations. The proposed mechanism underwent a comparative analysis with eight benchmark methods. The findings indicate that the proposed method shows a marked superiority over current techniques, realizing a 33.5% decrease in execution time (168.57 s compared to 253.47 s) and an 11%-52% enhancement in energy efficiency (0.653 kWh vs. a maximum of 0.852 kWh). The results validate the efficacy of the scheduling strategy in improving energy efficiency and performance within cloud computing environments.
Cyber security is very important in Wireless Sensor Networks (WSNs) for securing the transfer of files from attackers. Cyber-physical systems (CPs) are essential to monitor and observe the location of data in WSN. CPs are essential to monitor and track the location of data in a WSN. Many researchers have implemented different mechanisms to improve cybersecurity in WSN-enabled CP. These mechanisms are effectively performed based on the mobile anchor node or mobility of the head node. This algorithm suffers from computational complexity. Some traditional cybersecurity systems suffer from data loss, important data theft, and information leakage. In addition, the CPs also suffer from service interference issues. Black holes, scheduling, gray holes, and flooding are some of the examples of common WSN attacks that damage the entire security system in WSN. The WSN has disadvantages such as low identification rates, high computing overhead, and increased false alarm rates. Conventional cybersecurity systems are required to decrease data redundancy and increase the data correlation for better data transformation. In this paper, a new cybersecurity system in WSN is developed to detect WSN intrusions effectively to enhance adaptability and security. The normal and anomalous information is gathered from online resources. Initially, the gathered information is given to the Adaptive and Attention serial Cascaded Ensemble Network (A-ASCENet) for detecting various intrusions. Here, the variational autoencoder, Convolutional Neural Network (CNN), and extreme learning are integrated into a cascaded form to develop an A-ASCENet model. Here, the parameters are optimized using the Revised Fitness-based Lyrebird Optimization Algorithm (RF-ILOA) from A-ASCENet to enhance the performance of cybersecurity. At last, various WSN attacks like gray, scheduling, flooding, black holes, and holes are effectively detected. The performance of cybersecurity in WSN is compared over different traditional methods with some performance metrics.
Vehicular ad hoc networks (VANETs) enable the exchange of safety-critical messages, but their open and dynamic nature makes them vulnerable to message falsification and denial-of-service attacks. Federated learning (FL) offers a distributed defense mechanism, yet conventional approaches such as FedAvg degrade severely under non-IID client data and are highly sensitive to adversarial updates. To address these limitations, we propose FedCross-VAN, a novel FL framework that incorporates cross-domain behavioral priors with a similarity-weighted aggregation scheme. On the client side, FedCross-VAN employs a dual-objective loss that balances anomaly classification with alignment to external priors, improving generalization under heterogeneous data. On the server side, updates are aggregated proportionally to their embedding similarity with the priors, suppressing the influence of poisoned or noisy clients. Experiments on two benchmark datasets show that FedCross-VAN achieves up to approximately 2%-4% higher accuracy on HAR and 13% on KWS, converges in fewer rounds, and exhibits stronger robustness than FedAvg and No-Transfer FL. These findings establish FedCross-VAN as a practical and resilient framework for anomaly detection in next-generation intelligent transportation systems.
In this paper, the big data classification is done using the hybrid Squeeze-EfficientNet with Feature fusion in the spark framework. At first, the partitioning of big data is executed by employing Bayesian Fuzzy Clustering (BFC). Later, big data classification is accomplished in the spark architecture. At the slave node, the subsequent process is accomplished and the partitioned data are applied to the slaves, where they are pre-processed by Quantile normalization. Further, the feature fusion is carried out based on the Deep Kronecker Network (DKN) and Matusita Distance measure. At the Master node, all the features from the slave nodes are fused and the big data are classified by employing the Hybrid Squeeze-EfficientNet. The Hybrid Squeeze-EfficientNet is generated by integrating SqueezeNet and EfficientNet. The evaluation results show that the Squeeze-EfficientNet attained an accuracy of 0.904, a sensitivity of 0.916, and a specificity of 0.925. The high performance obtained by the devised model improves the big data classification task in real-time scenarios like targeted marketing, agriculture, smart transportation, efficient resource allocation, and personalized health monitoring systems. Thus, the integration of big data classification within daily life provides more informed decisions thereby improving the life quality.
The emergence of cloud computing has revolutionized business operations by providing effective scalability and flexibility. Security concerns have intensified due to the vast amount of data processed and stored in the cloud; hence protecting cloud infrastructure from cyber threats is crucial. Intrusion detection system plays a pivotal role in seamless monitoring of network traffic for exhibiting unauthenticated or malicious attempts. Recent advancements in IDS highlight certain issues such as low classification accuracy, high false positive rate, as well as overfitting when processing various network data. The feature extraction uses graylevel radial component analysis (GRCA) to extract salient features, while dimensionality reduction is performed by introducing the radial basis function principal component analysis. In this work, the crossover boosted dynamic cheetah optimization algorithm is employed in the feature selection process, which integrates Cheetah Optimization with dynamic evolutionary strategies to improve the overall search efficiency and tackle local optimal issues. The detection and classification of intrusion are performed by proposing a novel threshold-based kernel extreme learning machine, which uses different thresholds to enhance generalization capability. Extensive experimental and statistical analysis is carried out, and the results exhibit that the proposed framework achieves a classification accuracy, precision, recall, F 1 score, and security rate of 98.84%, 97.22%, 97%, 97.2%, and 98.85%, respectively, compared to all other existing models. Finally, the classified data is stored in cloud infrastructure that allows third-party monitoring services to assess and analyze critical intrusions and also provide threat analysis.
Digital twin technology has emerged as a key innovation in digitalization, gaining significant attention for its wide applicability across space and manufacturing industries. Its primary goal is to enable efficient command execution and secure data access, empowering users within a virtual environment. Digital twins support various functions, such as real-time monitoring, data analysis, and synchronized operations. However, despite their growing adoption, critical issues related to data privacy and security within digital twin systems remain underexplored. To address this, the article introduces an advanced optimization algorithmic technique, Chronological_Fossa Optimization Algorithm_Secure Key Generation (CFOA_Seckeygen), for generating an optimal key to improve the security and privacy of data stored in a digital twin environment with a blockchain framework. Towards this, different entities, like the twin manager, data owner, database server, and data user, are involved in the authentication process, which is executed by considering different functions, like Exclusive OR (XOR) operations, cryptographic hashing, encryption, and keys. Following this, a secret key is generated using CFOA_Seckeygen to increase security as well as the privacy of digital twin data. Furthermore, the CFOA_Seckeygen model demonstrates superior performance, achieving a communication cost of 3007.556, memory usage of 43.876 MB, a normalized variance of 0.885, and a conditional privacy score of 0.886.
The Internet of Things is growing tremendously due to new technologies, advancements, and big data. With the digitization of data and continuous technological progress, network data traffic has seen a significant increase. This growth makes IoT networks more vulnerable to attacks because of the rising number of devices and the massive amount of data they generate. One of the emerging topics in the research field is security in IoT. The enormous volume of data poses significant challenges to privacy and cybersecurity, and the frequency of attacks is directly proportional to Internet usage. Intrusion Detection Systems (IDS) have proven effective in detecting various attacks, malicious activities, and unauthorized access in IoT networks, helping to prevent intrusions. Furthermore, advanced AI technologies such as machine learning, deep learning, ensemble learning, and transfer learning have shown promising results in efficiently identifying intrusions, attacks, and malicious actions. This paper presents the development of an effective Intrusion Detection System using Machine and Deep Learning algorithms, compares their performance, and identifies the most effective algorithm for securing IoT data while preserving privacy. Random Forest, Convolutional Neural Networks, and Deep Neural Networks are implemented, tested, and compared with other machine learning algorithms, including Decision Trees, Gaussian Na & iuml;ve Bayes, and XG-Boost. The implementation is carried out in Python, using the benchmark KDD dataset. This paper covers the processes of data generation, preprocessing, analysis, and intrusion detection. The experimental results are compared with other state-of-the-art methods to evaluate overall performance. The performance metrics such as accuracy, precision, recall, and F1 score have been computed for the case of deep learning and machine learning for given IoT network.
Industrial countries are moving toward digitizing the manufacturing processes in their factories by integrating the expected next-generation technologies such as software-defined networking (SDN), cloud computing, and industrial Internet-of-Things (IIoT). However, developing smart factories that combine these physical and cyber components faces critical challenges, particularly regarding the efficiency and security domains. For example, Distributed Denial of Service (DDoS) attacks in industrial environments could impact the progress of the automated processes and the availability of SDN-based networks. In this paper, we present a novel collaborative intrusion detection system (CIDS) approach for SDN-based industrial environments that integrates edge computing techniques to enhance security and operational efficiency. Our model optimizes resource utilization across dispersed industrial sites by uniquely combining three different IDSs: centralized-based IDS, edge-based Anomaly IDS (AIDS), and signature-based IDS (SIDS). The proposed approach establishes consistent, network-wide security policies to accommodate the varying processing capabilities. Moreover, the use of edge computing techniques minimizes the overhead introduced by the SDN controller located in the cloud layer and addresses scalability challenges in large-scale networks with heavy traffic loads. Evaluation is performed using the Mininet emulator, and the results reveal a detection accuracy of up to 98%. Furthermore, profiling outcomes of the centralized controller indicate a 50% reduction in traffic monitoring function calls, highlighting the efficiency and superiority of the proposed methodology, particularly for geographically dispersed industrial sites.
Voice over Internet Protocol (VoIP) has emerged as a game-changing communication technology given that it allows for low-cost long-distance conversations with plenty of additional benefits. In this era of cloud computing, VoIP can offer even cheaper calls and scalable services with the help of virtualized telephone infrastructure. The integration of virtualized telephone infrastructure with VoIP is known as “ cloud-based VoIP .” In this paper, we investigate a cloud-based VoIP under the advanced persistent threat (APT) attack. An APT attack is a sophisticated type of cyberattack that tries to steal personal information by staying in the infected system for an extended period of time, thereby impacting the system dependability. “Dependability is a measure of a system's availability, reliability, maintainability, and in some cases, other characteristics such as durability, safety and security”. Hence, we develop a robust mechanism for mitigating APT attack in a cloud-based VoIP phone system and investigate its dependability to minimize the aftermaths of the attack. We employ a semi-Markov process (SMP) model to study the dependability as it gives consideration to the non-Markovian nature of the holding times of various system states. The SMP model is then used to analyze both the time-dependent behavior and the long-term (stationary) performance characteristic of the cloud-based VoIP system, specifically in terms of availability, reliability, and confidentiality. Numerical results are displayed graphically, and the proposed dependability model is supported by stochastic simulation. It has been established from the numerical results that the cloud-based VoIP is the most sensitive and critical when it is exploited by cyberattacks, and the lifetime of the system can be extended if the weaknesses of the system are discovered before it is exploited by the attackers.
Vehicular ad hoc networks (VANETs) enable real-time communication but are vulnerable to security threats, particularly distributed denial of service (DDoS) attacks, that cause delays and network failures. Traditional static detection systems struggle to adapt to dynamic traffic conditions. To address this problem, we propose ELITE, a lightweight and intelligent DDoS detection framework designed for secure VANETs. ELITE employs a three-layer architecture featuring a random fuzzy tree (RFT) classifier, which combines the speed of decision trees with adaptive fuzzy reasoning for efficient anomaly detection. It also includes a latency-aware scheduling system that ensures the urgent traffic is handled, while a few essential requests are sent to nearby edge servers or to the cloud. This work has three distinct contributions: integration of X and Y into one intelligent smart-environment architecture, like edge-cloud optimization model 96% stability of delay-sensitive edge-cloud optimization, and development of a lightweight threat detection module with improved accuracy and real-time capability. Experimental results demonstrate that ELITE achieves a high detection accuracy of 95.7%, effectively adapts to traffic changes, reduces false positives, and improves latency performance.
To address the challenges of false negatives and false positives of small objects and the difficulty of fine-grained behavior recognition in complex traffic scenarios, this paper constructs a hybrid deep learning framework based on image detection to synergistically improve multi-object localization accuracy and semantic understanding capabilities. The framework first uses a combination of Gaussian and bilateral filtering for denoising, enhancing input quality and improving detection sensitivity for small objects. In the detection phase, the YOLOv5s (You Only Look Once 5s) model is used as the baseline. The Convolutional Block Attention Module (CBAM) attention mechanism is applied to enhance the representation of key features. K-means clustering is used to adaptively generate prior anchor boxes that match the scale distribution of objects in traffic scenarios. The CIoU (Complete Intersection over Union) loss function is also used to optimize bounding box regression accuracy, improving small object detection performance while maintaining model lightweight. To achieve fine-grained semantic understanding, a two-branch classification network is designed. The attribute branch uses the ConvNeXt-Tiny (Convolutional Next-Generation Tiny) structure to extract static appearance features, while the event branch utilizes the nonlocal operations module to capture dynamic contextual dependencies. Weighted fusion of these two features enables joint recognition of attributes and behaviors. A GNN-CNN (Graph Neural Network-Convolutional Neural Network) hybrid classification module is also constructed. The GNN models the spatiotemporal interactions between vehicles, while a lightweight CNN extracts local texture features. These features are adaptively fused using the Squeeze-and-Excitation (SE) attention mechanism, and a softmax classifier performs traffic behavior judgment. Experiments show that the YOLOv5s-CBAM model achieves a mean average precision (mAP) of 0.55 for detecting extremely small objects (< 16 x 16). In the overloaded vehicle detection task, the GNN-CNN module achieves accuracy and recall of 0.92 and 0.90, respectively. This hybrid deep learning framework provides reliable technical support for automated traffic inspections. It improves the accuracy and stability of small object detection and fine-grained event recognition in complex traffic scenarios. Its modular design and strong scalability make it widely applicable and conducive to promoting intelligent transportation towards higher levels of automation.
The rapid expansion of digitalization has intensified cybersecurity risks, exposing critical network vulnerabilities despite significant advances in encryption and intrusion detection systems (IDS). Many existing deep learning-based IDS still struggle with high false-positive rates, misclassification, and limited adaptability, reducing their effectiveness in real-time defense scenarios. To address these limitations, this study proposes a Polymorphic Graph Gudermannian Neural Network integrated with Adaptive Chaotic Satin Bowerbird Optimization (PG-GNN-AC-SBO), complemented by a lightweight encryption mechanism. The framework incorporates a Fuzzy K-Top Matching Value (FKTMV) module for robust preprocessing and normalization, along with a Hybrid Cat Hunting Sea-Horse Optimizer (H-CHO-SHO) for efficient and interpretable feature selection. The PG-GNN classifier employs graph-based learning and a Gudermannian nonlinear activation function to effectively capture complex traffic behavior, while AC-SBO dynamically tunes hyperparameters to enhance stability and classification accuracy. To ensure data confidentiality, a Synchronously Scrambled Diffuse Encryption (SSDE) scheme is applied, delivering strong security with low computational overhead. Experimental evaluations on the NSL-KDD and CICIDS2017 datasets demonstrate the superiority of the proposed approach, achieving up to 99.82% accuracy and outperforming state-of-the-art methods. The encryption and decryption times of 3.50 and 3.55 ms further confirm the model's lightweight design. Overall, the proposed system provides high throughput with minimal latency, demonstrating strong potential for real-time and large-scale cybersecurity deployments.
Healthcare data holds immense potential to improve diagnostics and treatment, but centralized collection risks patient privacy and suffers from limited labeled samples. Existing methods fall short in securely leveraging distributed data while reducing annotation costs. To address this, we propose FTAL-QNC—a novel Federated Transfer Active Learning framework with Quantized Neural Cryptography—that enables privacy-preserving, communication-efficient, and label-efficient collaborative model training. FTAL-QNC integrates four core components in a unified architecture: (1) Federated learning for decentralized model training without data sharing, (2) Transfer learning to handle data heterogeneity across hospitals, (3) Active learning to minimize manual labeling by prioritizing informative samples, and (4) Quantized neural cryptography to ensure secure, low-overhead exchange of encrypted model updates. Empirical evaluations on real-world datasets, including Lung CT and ChestX-ray14, demonstrate that FTAL-QNC enhances segmentation and classification accuracy to 97.3% and 97.0% recall for Lung CT, respectively, while significantly reducing annotation effort compared to standard federated learning and other sampling methods. Our contributions include a privacy-preserving and communication-efficient collaborative framework, an integrated active learning mechanism for efficient data labeling, and a secure aggregation protocol via quantized neural cryptography. These results demonstrate FTAL-QNC's potential to advance safe, collaborative medical research and improve patient outcomes.
In today's digital world, with increasing threats of cyberattacks and unauthorized data access, protecting confidential information demands more robust security mechanisms. Cryptography and steganography are two prominent techniques employed to secure data. However, conventional steganography suffers from reduced embedding capacity and the risk of image distortion. To overcome these challenges, this research proposes a novel hybrid framework to enhance data security and embedding efficiency. The approach comprises two phases including the embedding phase and the extraction phase. In the embedding phase, both the cover and secret images undergo a 3-level discrete wavelet transform (DWT) using the Daubechies wavelet. Region selection in the transformed cover image is optimized by extracting and leveraging various features, including color, shape, deep features, and local Gabor transitional pattern (LGTrP) features. These features are processed via a modified Bidirectional Long Short-Term Memory (Bi-LSTM) model, enhanced with architectural improvements, which boost feature learning. Simultaneously, the secret image undergoes transformation using a modified Arnold map integrated with a Bernoulli map allows faster execution. The modified Arnold function's outcome is subjected to an encryption process. The embedding process is done after the encryption process; the modified Blowfish algorithm is used for the decryption process. Subsequently, the inverse Bernoulli map is utilized, with the resultant output given to the inverse Arnold map. Finally, an inverse 3-level DWT reconstructs the original secret image. Comparative evaluations demonstrate the proposed framework attains lower KPA and KCA rates of 0.12 and 0.15, respectively, which underscores the innovation of integrating a steganography-cryptography model in securing sensitive data against sophisticated attacks.
The rise in network intrusions has led to significant consequences, including privacy violations, financial losses, and unauthorized data transfers. Attackers exploit vulnerabilities in network systems, compromising security and disrupting services. Traditional intrusion detection systems (IDS) often face challenges such as false positives, delayed threat identification, and poor detection of minority attack classes. To address these issues, this research proposes advanced deep reinforcement learning with deep learning-based NIDS for improved threat detection and mitigation. Data from IoT-2023, BoT-IoT, CIC-IoT 2023, and RT-IoT 2022 datasets are preprocessed through null value handling, data cleaning, one-hot encoding, and Min-Max normalization. To enhance the detection of minority attacks, the Tabular Auxiliary Classifier Generative Adversarial Network (TACGAN) is employed for synthetic data augmentation. Feature extraction is performed using the Graph Sample and Aggregate Attention Network (GSAAN), which captures basic, content, and traffic-based features. Significant features are selected using the Mountaineering Team-Based Optimization (MTBO). Attack classification is carried out using a novel ensemble of the Improved Double Deep Q-Network (IDDQN) and Deep Autoregression Feature Augmented Bidirectional LSTM (DAF-BiLSTM), which is termed the OptIDQDBiLSTM approach, ensuring robust learning of spatial and temporal dependencies. Hyperparameter tuning is optimized using the Boosted Wild Horse Optimization Algorithm (BWHOA). Experimental results show that the proposed approach outperforms existing IDS methods, achieving higher detection rates, improved accuracy, and a reduced false alarm rate while maintaining computational efficiency. While comparing with existing state of the art approaches, the proposed approach surpasses existing methods with over 99.64% accuracy, 99.34% precision, and 99.42% recall. These findings demonstrate the effectiveness of deep reinforcement learning in enhancing network security against evolving cyber threats.
Obstructive sleep apnea (OSA) becomes a sleep disease caused by recurrent cessation of breathing during sleep; it also leads to various health complications. Despite the availability of diagnostic methods, there are challenges in accurately identifying and classifying OSA severity. This research addresses the need of an efficient and reliable automated system for OSA detection using deep learning techniques. Existing problems include the complexity of OSA diagnosis, reliance on manual scoring, and variability in interpretation. The proposed OSA grading network (OSAG-Net) encompasses several steps: preprocessing of raw Electrocardiogram (ECG) data to extract relevant features, application of self-feature controllable-black window optimization (SFC-BWO) for feature selection to enhance classification performance, and utilization of bidirectional gated recurrent neural network (BGRNN) architecture with recurrent neural networks (RNN) and bidirectional gated recurrent units (Bi-GRU) for OSA severity classification. Preprocessing involves filtering noise and artifacts from ECG signals, followed by segmenting data into smaller windows to extract informative features. The SFC-BWO technique optimally selects the features by iteratively refining feature subsets based on classification performance, effectively reducing dimensionality and enhancing model interpretability. The RNN architecture with Bi-GRU units is employed to capture temporal dependencies of sequential data, such as ECG recordings, enabling more accurate classification of OSA severity levels. Finally, the performance of the system is validated with different metrics. Hence, the proposed OSAG-Net model achieves a high accuracy value of more than 4.77% of SVM, 3.67% compared to Grad-CAM, and 2.54% of CNN-LSTM, respectively. This results in improvement in the system proves that it rapidly and effectively diagnoses the disease and treats the patients accordingly.
The Internet of Things (IoT) is particularly vulnerable in this new era, as IoT devices often rely on lightweight encryption and security measures due to their limited processing capabilities and power constraints. Existing signature-based intrusion detection strategies are inadequate against the advanced and adaptive nature of quantum attacks, which can exploit the polymorphic and metamorphic behavior of quantum-enhanced malware. This research addresses the challenges posed by a possible attack scenario named the “Quantum-Enhanced Cloak Malware (QECM) Attack” and aims to provide enhanced protection to IoT systems against this sophisticated threat. The proposed “Intelligent Swift Scan Quantum-Resilient Intrusion Detection System (ISS-QR-IDS)” integrates advanced techniques to enhance detection speed and accuracy, mitigate risks associated with polymorphic and metamorphic malware behaviors, and secure communication channels against quantum threats. The model incorporates the Parallel Vario-Isolation Detector, which combines Variational Autoencoders (VAEs) and Parallelized Isolation Forests to detect quantum-enhanced malware, and Hypergraph Attention Networks (HGA-Net), leveraging Hypergraph Neural Networks (HGNNs) and Graph Attention Networks (GATs) to detect the critical interactions and improve anomaly detection accuracy. Additionally, postquantum cryptographic algorithms like NTRU Encrypt and FALCON ensure secure communication channels and data integrity. By combining these advanced techniques, ISS-QR-IDS aims to provide a robust defense mechanism against sophisticated cyber threats targeting IoT networks, ensuring their security and resilience in the face of quantum computing advancements.
The Internet of Vehicles (IoV) collects real-time data on traffic, environmental conditions, and vehicle behavior through vehicle interconnection and interaction with infrastructure, providing support for the of Machine Learning (ML) in intelligent decision-making. However, centralized learning approaches suffer from issues like privacy leakage and high communication costs. Federated Learning (FL) addresses these issues by sharing local model updates, but in IoV environments, challenges such as data heterogeneity result in slow convergence, limited communication resources, and security threats like gradient leakage. To tackle these challenges, this paper proposes Adaptive Blockchain-based Hierarchical Federated Learning with Gradient Alignment (ABHFL). ABHFL groups vehicle nodes and RSUs into a hierarchical structure to perform local training, gradient alignment, and model aggregation at different levels. The proposed Adaptive Gradient Alignment (AGA) mechanism aligns the update directions of nodes towards the global optimal direction through multiple rounds of alignment after local gradient computation, accelerating model convergence and ensuring that the gradients uploaded contribute positively to global optimization. In addition, a lightweight Proof-of-Gradient-Alignment (PoGA) consensus mechanism is designed, which performs two-stage verification of the uploaded gradients and integrates reputation scores and blockchain storage to guarantee gradient reliability and protect against attacks. Extensive experiments demonstrate that ABHFL significantly improves model convergence, communication efficiency, and security reliability, providing an effective and robust solution for FL in IoV scenarios.
