ABSTRACT The Internet of Vehicles (IoV) is revolutionising transportation by connecting vehicles, infrastructure and devices, enabling more intelligent and safer mobility. One key challenge is ensuring efficient and secure communication among vehicles with varying capabilities, including different sizes, speeds and sensor configurations. This research introduces a Federated Learning‐Driven Deep Learning (FLDL) approach to intelligent collision avoidance, designed to address the heterogeneity of vehicles in the IoV ecosystem. The system integrates real‐time data from vehicle‐to‐vehicle (V2V) and vehicle‐to‐infrastructure (V2I) communications, while considering factors like vehicle type, road conditions, driver behaviour and Digital Twins. Our approach leverages multiple Federated Learning strategies, which enhance privacy protection, reduce communication overhead and enable real‐time decision‐making without the need for centralised data storage. Experimental results show that the GNN + FedGC model achieves the highest performance with an accuracy of 98.8%, outperforming other models such as MLP with FedLU (98.5%), DRL with FedPPO (98.3%) and LSTM with FedSGD (97.65%). The integration of Digital Twins further enhances model accuracy by simulating real‐time vehicle behaviour and environmental conditions. This FL‐based system not only improves collision prediction but also enhances safety, reduces accident rates and supports scalable decision‐making in smart city transportation systems.

The Internet of Natural Things (IoNT) is the extension of the Internet of Things (IoT) concept to natural environments, with the ability to monitor the environment in real-time using integrated sensor networks, Artificial Intelligence (AI), and remote sensing. IoNT can offer solutions to the ecological crisis that is facing the world today, such as climate change, loss of biodiversity, and depletion of resources. The survey addresses the technology behind the IoNT and its different applications in disaster management, forest conservation, biodiversity monitoring, and agriculture, among others. IoNT has been successfully applied to monitor the quality of water in remote river ecosystems and to irrigate precision agriculture. Furthermore, pilot projects have demonstrated that IoNT can be used to make decisions based on real-time data analytics so that it is possible to manage resources sustainably. The other aspect discussed in the survey is the consistency between the IoNT sensors and the commercial sensors, according to some case studies presented in the literature. In those works, it is also identified that the processes are experimental in sensor testing, data synchronization, and the functioning of standard metrics, which ensure the reliability and strength of IoNT sensor systems. Despite the fact that the IoNT has great strengths, it is still possible to encounter issues concerning data security, energy efficiency, and interoperability. The survey addresses these issues and identifies new trends, such as blockchain and autonomous systems, that can make IoNT applications more scalable and efficient to manage the sustainable ecosystem.

The hybrid model combines deep neural networks (DNN) and large AI models, such as large language models (LLM), for enhanced clinical information retrieval (CIR) from electronic clinical records (ECR). While LLMs show promise for encoding complex medical data, they face challenges in user-dependent information, such as patient reports with encoded knowledge, accessing real-time data, and requiring extensive fine-tuning for clinical decision-making in online communication systems. To overcome these limitations, we introduce a Transformer-based Sequence (TBS) multimodal method that integrates representation learning with human expertise to encode and analyze intricate relationships within clinical data. This model improves predictive tasks and medical search accuracy, achieving F1-scores of 0.83-0.80, and outperforms baseline methods. Integrating AI-driven methodologies in healthcare has the potential to transform medical record analysis and utilization, resulting in enhanced patient outcomes and more personalized healthcare solutions.

Architectural Distortion (AD) is a common abnormality in digital mammograms, alongside masses and microcalcifications. Detecting AD in dense breast tissue is particularly challenging due to its heterogeneous asymmetries and subtle presentation. Factors such as location, size, shape, texture, and variability in patterns contribute to reduced sensitivity. To address these challenges, we propose a novel feature fusion-based Vision Transformer (ViT) attention network, combined with VGG-16, to improve accuracy and efficiency in AD detection. Our approach mitigates issues related to texture fixation, background boundaries, and deep neural network limitations, enhancing the robustness of AD classification in mammograms. Experimental results demonstrate that the proposed model achieves state-of-the-art performance, outperforming eight existing deep learning models. On the PINUM dataset, it attains 0.97 sensitivity, 0.92 F1-score, 0.93 precision, 0.94 specificity, and 0.96 accuracy. On the DDSM dataset, it records 0.93 sensitivity, 0.91 F1-score, 0.94 precision, 0.92 specificity, and 0.95 accuracy. These results highlight the potential of our method for computer-aided breast cancer diagnosis, particularly in low-resource settings where access to high-end imaging technology is limited. By enabling more accurate and timely AD detection, our approach could significantly improve breast cancer screening and early intervention worldwide.

Embedded resources have been becoming part of the Internet of Things networks, where they are increasingly taking part in various kinds of decision-making using Tiny Machine Learning (TinyML) models. Although offloading the TinyML model for these devices includes removing many layers that have less impact on the overall performance, they often lead to a sacrifice on the overall performance of the model. In this paper, we propose a novel device-aware training strategy to customize the training based on the resources on which the model will be applied. We proposed AlphaEnhancer, a resource-aware game agent for medical image super-resolution. We baseline our approach on the Residual Feature Distillation Model (RFDN) and propose a device efficacy metrics, which is based on the learned actions of the agent. The model with the highest efficacy is deemed appropriate for that particular device. Our preliminary results show that our methods performed significantly well with respect to the baseline and other recent state-of-the-art.

ABSTRACT Brain‐computer interface (BCI) is a growing area of research in human‐computer interaction (HCI), where its potential ranges from medicine to entertainment. It intends to manage various assistive technologies through the utilization of brain signals. This technology acquires and interprets brain signals before sending them to a connected device, which generates controls based on the obtained signals. Emotion‐based mental state categorization employing electroencephalogram (EEG) signals is an emerging method of BCI application. However, EEG signals comprise artifacts and redundant or noisy information from the subject, equipment, and external environment. Also, the EEG signals have a low spatial resolution (physical location of the activity within the brain) but a high temporal resolution (millisecond level). Therefore, artifact removal, feature extraction, and classification of EEG signals are challenging. This work proposed a novel approach called Extended Independent Component Analysis (E‐ICA) for artifact removal from EEG signals. A Multi‐class Common Spatial Pattern (M‐CSP) is proposed for feature extraction. A Bidirectional long short‐term memory (BiLSTM) network is proposed to improve the classification of EEG signals and fine‐tune its parameters. This study leverages the Database for Emotion Analysis using the Physiological Signals (DEAP) dataset to validate the model's performance. This dataset includes EEG recordings annotated with emotional attributes such as valence, arousal, dominance, and liking. After conducting several experiments, the proposed approach achieves a high classification accuracy of 94.61% and outperforms state‐of‐the‐art works. The proposed approach can be successfully integrated into BCI systems for real‐time emotion identification in healthcare and user engagement detection in gaming environments.

Abstract In the dynamic domain of healthcare data management, safeguarding sensitive information while ensuring data efficiency is always of the highest priority. Healthcare data are frequently mishandled, posing significant risks. This research offers a new network that assesses the quality of visual data using robust features‐driven Mutual Entropy Gain (MEG). The proposed network addresses a critical gap in healthcare data management, significantly enhancing patient data security and operational efficiency in medical institutions. Our method begins with a thorough empirical investigation to find the optimal intermediate features for network input. We incorporate both distance entropy and probability entropy adopted and normalized in MEG, resulting in a comprehensive healthcare data quality evaluation. The results show that the network can distinguish between high‐quality and low‐quality data based on information content. Furthermore, our assessment reveals a large performance discrepancy between high and low‐quality data, even with variable datasets. Notably, using only half of the data achieves commendable accuracy when compared with using the complete dataset, demonstrating possible efficiency gains. This breakthrough has far‐reaching implications for healthcare providers, potentially reducing data storage costs, accelerating data processing times, and minimizing the risk of data breaches. In essence, our proposed network enhances efficiency and security in healthcare data and adapts to the evolving landscape of convergence ICT, paving the way for more robust, cost‐effective, and secure healthcare information systems that can significantly improve patient care and operational outcomes.

Understanding complex traffic patterns has become more challenging in the context of rapidly growing city road networks, especially with the rise of Internet of Vehicles (IoV) systems that add further dynamics to traffic flow management. This involves understanding spatial relationships and nonlinear temporal associations. Accurately predicting traffic in these scenarios, particularly for long‐term sequences, is challenging due to the complexity of the data involved in smart city contexts. Traditional ways of predicting traffic flow use a single fixed graph structure based on the location. This structure does not consider possible correlations and cannot fully capture long‐term temporal relationships among traffic flow data, making predictions less accurate. We propose a novel traffic prediction framework called Multi‐scale Attention‐Based Spatio‐Temporal Graph Convolution Recurrent Network (MASTGCNet) to address this challenge. MASTGCNet records changing features of space and time by combining gated recurrent units (GRUs) and graph convolution networks (GCNs). Its design incorporates multiscale feature extraction and dual attention mechanisms, effectively capturing informative patterns at different levels of detail. Furthermore, MASTGCNet employs a resource allocation strategy within edge computing to reduce energy usage during prediction. The attention mechanism helps quickly decide which services are most important. Using this information, smart cities can assign tasks and allocate resources based on priority to ensure high‐quality service. We have tested this method on two different real‐world datasets and found that MASTGCNet predicts significantly better than other methods. This shows that MASTGCNet is a step forward in traffic prediction.

The Internet of Health Monitoring (IoHM) plays a vital role in Emergency Medical Package Delivery (EMPD) by enabling real-time monitoring and transmission of critical health data through interconnected devices in 6G networks. UAVs act as Aerial Base Stations (ABSs), facilitating data collection and transmission between GAI-IoHM devices and edge servers. This is crucial for efficient communication in 6G-enabled Space-Air-Ground Integrated Networks (SAGIN). However, UAVs supporting EMPD face challenges related to limited energy and computational capacity, especially during task offloading to edge servers. To address these constraints, this paper proposes the integration of Generative Artificial Intelligence (GAI) into UAVs for intelligent policy learning, enabling adaptive decision-making, real-time diagnostics, and efficient path planning under uncertainty. We present a novel multi-agent deep reinforcement learning (MADRL)-based joint optimization framework for cooperative task allocation, trajectory planning, and power management (CTATP) in a 6G-enabled SAGIN architecture. The problem is modeled as a Partially Observable Markov Decision Process (POMDP) to capture dynamic and uncertain operational conditions. To solve this, we introduce the GAI-based Deep Deterministic Double Policy Gradient (GAI-DD3PG) algorithm, which leverages a generative actor-network to learn adaptive, energy-efficient control policies from latent action spaces. Simulations in urban emergency scenarios with 10–20 UAVs demonstrate GAI-DD3PG’s efficacy. Compared to Proximal Policy Optimization (PPO), Deep Deterministic Policy Gradient (DDPG), Multi-Agent Federated Reinforcement Learning (MAFRL), and Greedy heuristics, GAI-DD3PG achieves a 20% energy reduction, 15% higher delivery success rate,25% shorter trajectories, and 30% improved resource utilization. These results highlight its potential for reliable EMPD in complex 6G SAGIN environments.

In the context of rapidly growing city road networks, understanding complex traffic patterns and implementing effective safety monitoring through advanced Transportation Cyber-Physical Systems (T-CPS) has become increasingly challenging. This involves understanding spatial relationships and non-linear temporal associations. Accurately predicting traffic in such scenarios, particularly for long-term sequences, is challenging due to the complexity of the data. Traditional ways of predicting traffic flow use a single fixed graph structure based on location. This structure does not consider possible correlations and cannot fully capture long-term temporal relationships among traffic flow data, thereby limiting the system ability to ensure safety and reliability. To address this challenge, we propose a novel traffic prediction framework called Attention-based Spatio-temporal Multi-scale Graph Convolutional Recurrent Network (ASTMGCNet). This study introduces a novel framework designed to improve prediction accuracy in dynamic urban traffic systems by effectively capturing complex spatio-temporal correlations through multi-scale feature extraction and attention mechanisms. ASTMGCNet records changing features of space and time by combining Gated Recurrent Units (GRU) and Graph Convolutional Networks (GCN). Its design incorporates multi-scale feature extraction and dual attention mechanisms, effectively capturing informative patterns at different levels of detail. This strategic design allows ASTMGCNet to effectively capture complex spatio-temporal correlations within traffic sequences, enhancing prediction accuracy. We have tested this method on two different real-world datasets and found that ASTMGCNet predicts significantly better than other methods, demonstrating its potential to advance traffic flow prediction and improve safety and reliability in T-CPS applications.

The Internet of Vehicles (IoV) enables technology that allows IoV and vehicles to connect everything. IoV has become an essential component of modern life. This exponential growth of IoV technology has introduced significant security and privacy issues, which pose potential threats to different types of attacks and cause different threats to the normal operation of vehicles. To prevent intelligent vehicle accidents and identify malicious attacks within IoV networks, various researchers have focused on machine learning (ML)-based methods to detect attacks. Intrusion detection systems (IDS) are a prominent solution for cyber attacks in IoV using ensemble learning. To achieve higher accuracy and detection rate, designing an improved detection framework using ensemble learning is a challenging task. The design of an ensemble-based IDS depends on two main challenges: selecting base classifiers and their combination methods. Therefore, in this study, we propose a hybrid ML model to detect various attacks in IoV. We have used different ML algorithms to develop an enhanced algorithm that can efficiently detect attacks in IoV networks. To evaluate the performance of the proposed system, we have used two well-known datasets, (CIC-IDS2017) and (UNSW-NB15). The proposed algorithm shows outstanding performance from the performance results, with an average attack detection accuracy of 99.75% and 100% and an F1 score of 99.74% and 100%, respectively, for both datasets. Further performance scores, that is, recall, precision, and F1 score metrics, validate the exceptional effectiveness of the proposed framework.

This paper proposes a novel radiation-hardened high-reliability SRAM cell, namely SHRCO, with 12 transistors for robust value storage as well as 6 transistors for parallel access operations. Using separated and error-interceptive feedback paths, the proposed cell has a complete self-recoverability from single-node upset (SNUs) at all single nodes and an excellent self-recoverability from double-node upsets (DNUs) at a part of node pairs. In addition, the proposed cell has superior access operation speed due to the inclusion of extra parallel access transistors. Simulation results show that the proposed cell has the largest number of node pairs that can self-recover from DNUs. Moreover, compared to the existing radiation-hardened SRAM cells, the proposed cell saves 28% of read time and 3% of write time on average.

Abstract Single image super‐resolution (SISR) has gained significant attention in image processing and computer vision, driven by deep learning‐based models like convolutional neural networks (CNN). Yet, the resource‐intensive nature of these models poses challenges when deploying them on edge devices. To address this issue, resource‐constrained models need to be developed. While recent models like the information distillation network (IDN), the information multi‐distillation network (IMDN), the residual feature distillation network (RFDN), and so on, have attempted to reduce parameters and computational complexity, further optimization remains vital. Therefore, this paper presents an approach to enhancing the efficiency and lightweight nature of the SISR. We introduce a novel lightweight SR model by building upon the RFDN architecture, the winner of the AIM2020 and NTIRE2022 SR challenges. The proposed depthwise channel attention network (DWCAN) model makes some key changes to RFDN. First, it replaces the main residual feature distillation block (RFDB) with a depthwise channel attention block (DWCAB). Additionally, DWCAN includes a shallow residual block (SRB) with depthwise separable convolution (DW) and a channel attention (CA) block. The primary goal of our work is to significantly reduce model parameters, computational operations, inference time, and memory size while maintaining or improving a peak signal‐to‐noise ratio (PSNR) of 29 dB. The experimental results demonstrate the effectiveness of the proposed model. By applying our modifications, we achieve a notable reduction in model complexity, leading to an improved PSNR of 29.07 dB, up from RFDN's 29.04 dB on a diverse 2 K resolution (DIV2K) dataset. This underscores the potential of our lightweight model to balance computational efficiency and SR quality. Additionally, the proposed work is essential for the metaverse for two key reasons: (1) Enhancing visual quality by adding complex details to textures and objects, making the digital world feel more like reality. (2) Ensuring device compatibility across a range of gadgets, from smartphones to VR headsets, optimizing the metaverse experience for all users. In short, a lightweight single image super‐resolution model for image reconstruction is proposed in this paper.

Many problems related to natural language processing are solved by neural networks and big data. Researchers have previously focused on single-task supervised goals with limited data management to train slogan classification. A multi-task learning framework is used to learn jointly across several tasks related to generating multi-class slogan types. This study proposes a multi-task model named slogan generative adversarial network systems (Slo-GAN) to enhance coherence and diversity in slogan generation, utilizing generative adversarial networks and recurrent neural networks (RNN). Slo-GAN generates a new text slogan-type corpus, and the training generalization process is improved. We explored active learning (AL) and meta-learning (ML) for dataset labeling efficiency. AL reduced annotations by 10% compared to ML but still needed about 70% of the full dataset for baseline performance. The whole framework of Slo-GAN is supervised and trained together on all of these tasks. The text with the higher reporting score level is filtered by Slo-GAN, and a classification accuracy of 87.2% is achieved. We leveraged relevant datasets to perform a cross-domain experiment, reinforcing our assertions regarding both the distinctiveness of our dataset and the challenges of adapting bilingual dialects to one another.

Epilepsy is a neurological disorder characterized by abnormal neuronal discharges that manifest in life-threatening seizures. These are often monitored via EEG signals, a key aspect of biomedical signal processing (BSP). Accurate epileptic seizure (ES) detection significantly depends on the precise identification of key EEG features, which requires a deep understanding of the data's intrinsic domain. Therefore, this study presents an Advanced Multi-View Deep Feature Learning (AMV-DFL) framework based on machine learning (ML) technology to enhance the detection of relevant EEG signal features for ES. Our method initially applies a fast Fourier transform (FFT) on EEG data for traditional frequency domain feature (TFD-F) extraction and directly incorporates time domain (TD) features from the raw EEG signals, establishing a comprehensive traditional multi-view feature (TMV-F). Deep features are subsequently extracted autonomously from optimal layers of one-dimensional convolutional neural networks (1D CNN), resulting in multi-view deep features (MV-DF) integrating both time and frequency domains. A multi-view forest (MV-F) is an interpretable rule-based advanced ML classifier used to construct a robust, generalized classification. Tree-based SHAP explainable artificial intelligence (T-XAI) is incorporated for interpreting and explaining the underlying rules. Experimental results confirm our method's superiority, surpassing models using TMV-FL and single-view deep features (SV-DF) by 4% and outperforming other state-of-the-art methods by an average of 3% in classification accuracy. The AMV-DFL approach aids clinicians in identifying EEG features indicative of ES, potentially discovering novel biomarkers, and improving diagnostic capabilities in epilepsy management.

Abstract Automatic modulation classification is a task that is essentially required in many intelligent communication systems such as fibre‐optic, next‐generation 5G or 6G systems, cognitive radio as well as multimedia internet‐of‐things networks etc. Deep learning (DL) is a representation learning method that takes raw data and finds representations for different tasks such as classification and detection. DL techniques like Convolutional Neural Networks (CNNs) have a strong potential to process and analyse large chunks of data. In this work, we considered the problem of multiclass (eight classes) classification of modulated signals, which are, Binary Phase Shift Keying, Quadrature Phase Shift Keying, 16 and 64 Quadrature Amplitude Modulation corrupted by Additive White Gaussian Noise, Rician and Rayleigh fading channels using 3D‐CNN architectures in both frequency and spatial domains while deploying three approaches for data augmentation, which are, random zoomed in/out, random shift and random weak Gaussian blurring augmentation techniques with a cross‐validation (CV) based hyperparameter selection statistical approach. Simulation results testify the performance of 10‐fold CV without augmentation in the spatial domain to be the best while the worst performing method happens to be 10‐fold CV without augmentation in the frequency domain and we found learning in the spatial domain to be better than learning in the frequency domain.

For more than two decades, the issue of simultaneous localization and mapping (SLAM) has gained more attention from researchers and remains an influential topic in robotics. Currently, various algorithms of the mobile robot SLAM have been investigated. However, the probability-based mobile robot SLAM algorithm is often used in the unknown environment. In this paper, the authors proposed two main algorithms of localization. First is the linear Kalman Filter (KF) SLAM, which consists of five phases, such as (a) motionless robot with absolute measurement, (b) moving vehicle with absolute measurement, (c) motionless robot with relative measurement, (d) moving vehicle with relative measurement, and (e) moving vehicle with relative measurement while the robot location is not detected. The second localization algorithm is the SLAM with the Extended Kalman Filter (EKF). Finally, the proposed SLAM algorithms are tested by simulations to be efficient and viable. The simulation results show that the presented SLAM approaches can accurately locate the landmark and mobile robot.

Mobile robot localization has attracted substantial consideration from the scientists during the last two decades. Mobile robot localization is the basics of successful navigation in a mobile network. Localization plays a key role to attain a high accuracy in mobile robot localization and robustness in vehicular localization. For this purpose, a mobile robot localization technique is evaluated to accomplish a high accuracy. This paper provides the performance evaluation of three localization techniques named Extended Kalman Filter (EKF), Unscented Kalman Filter (UKF), and Particle Filter (PF). In this work, three localization techniques are proposed. The performance of these three localization techniques is evaluated and analyzed while considering various aspects of localization. These aspects include localization coverage, time consumption, and velocity. The abovementioned localization techniques present a good accuracy and sound performance compared to other techniques.

Spectrum sensing is one of the key technologies in the field of cognitive radio, which has been widely studied. Among all the sensing methods, energy detection is the most popular because of its simplicity and no requirement of any prior knowledge of the signal. In the case of low signal-to-noise ratio (SNR), the traditional double-threshold energy detection method employs fixed thresholds and there is no detection result when the energy is between high and low thresholds, which leads to poor detection performance such as lower detection probability and longer spectrum sensing time. To address these problems, we proposed an adaptive double-threshold cooperative spectrum sensing algorithm based on history energy detection. In each sensing period, we calculate the weighting coefficient of thresholds according to the SNR of all cognitive nodes; thus, the upper and lower thresholds can be adjusted adaptively. Furthermore, in a single cognitive node, once the current energy is within the high and low thresholds, we utilize the average energy of history sensing times to rejudge. To ensure the real-time performance, if the average history energy is still between two thresholds, the single-threshold method will be used for the end decision. Finally, the fusion center aggregates the detection results of each node and obtains the final cooperative conclusion through “or” criteria. Theoretical analysis and simulation results show that the algorithm proposed in this paper improved detection performance significantly compared with the other four different double-threshold algorithms.