ABSTRACT Smart healthcare systems are expected to grow exponentially in the 6G era, relying on their capacity to provide superior network services in terms of data rates and deployment scale. This generates vast amounts of health data that must be processed in real‐time. Therefore, computational offloading for smart healthcare systems is an inevitable trend and is widely applied in numerous medical applications, ranging from image diagnosis and clinical treatment to responding to viral pandemics. In this study, we provide a comprehensive overview of computational offloading for Health Internet of Things (HIoT) applications, examining various aspects. Then, we propose a deep reinforcement learning (DRL)‐based intelligent task offloading framework that performs decision‐making on local devices, accounts for real‐time resource constraints and provides detailed analyses. The results demonstrate that the DRL‐based offloading strategy outperformed Greedy and Threshold‐based methods by 20%, 10% and 45% in latency, energy consumption and failure rate, respectively. Finally, we identified challenges and open issues related to pervasive smart healthcare systems.

ABSTRACT Unmanned Aerial Vehicle (UAV) networks are increasingly deployed in dynamic environments where reliable and low‐latency communication is critical. However, high mobility, intermittent connectivity, spectrum limitations, and energy constraints make conventional static communication protocols inadequate for maintaining stable, dependable links. To address these challenges, this paper proposes TRC‐MAPPO, a topology‐aware reliability‐constrained multi‐agent deep reinforcement learning framework for adaptive UAV swarm communication. The routing problem is formulated as a constrained decision‐making task that jointly considers packet delivery reliability, end‐to‐end delay, link stability, bandwidth usage, and energy consumption. Unlike single‐agent DRL baselines, TRC‐MAPPO represents the swarm as a dynamic communication graph and combines local relay selection with centralized training, enabling cooperative routing decisions under time‐varying network conditions. Simulations are conducted in a controlled, dynamic UAV environment, with the same mobility and traffic settings used for all methods. The proposed framework is compared with DQN and PPO over different swarm densities. Results show that TRC‐MAPPO achieves a higher packet delivery ratio, lower end‐to‐end delay, and more efficient energy behaviour, particularly in dense deployments. These findings indicate that topology‐aware cooperative learning can provide a scalable and practical solution for reliable UAV communication in future intelligent aerial and 6G‐enabled networks.

ABSTRACT The sixth‐generation communications networks (6G) are expected to be deployed in the 2030s with integrated space‐aerial‐ground and undersea architecture to provide seamless global connectivity. In this architecture, unmanned aerial vehicles (UAVs) are one of the most unique characteristics and are becoming increasingly important. The flexibility, high speed, and infrastructure independence of UAV systems make them ideal for many applications. However, these advantages also create great challenges in effective communication between UAVs. To address these challenges, reinforcement learning (RL) algorithms such as Q‐Learning have been investigated. However, the traditional Q‐learning algorithm mainly relies on delay parameters in the reward function for decision‐making route selection. Aiming to optimise the selection of sustainable and efficient communication for UAVs, we propose an improved routing algorithm based on Q‐Learning for UAV communication. Our method integrates latency, energy consumption, and link quality parameters into the reward function to make smarter routing decisions. The simulation results show that Q‐Proposed achieves significant gains in terms of packet delivery ratio and end‐to‐end delay compared to other methods, paving the way for sustainable 6G UAV communications.

The development of real-time applications integrated with Intent-Based Networking (IBN) integrates an Internet of Things (IoT), providing interconnection between heterogeneous devices and physical objects for the formulation of smart cities. These systems provide seamless communication and maintain the adaptive network policies and infrastructure. Many existing schemes have proposed solutions for efficient routing with support from intelligent architectures; however, although most of them overlook the bounded and limited resources of IoT networks, they impose additional overhead while addressing unpredictable communications in IBN-IoT. Furthermore, security and trustworthiness are significant research challenges that must be addressed to prevent data breaches and allow only the use of authentic devices. This research presents a scalable model for an IBN-IoT environment that utilizes edge computing to enable trustworthy and fault-tolerant communication with energy efficiency. Firstly, Software-Defined Networking (SDN) is explored for load balancing and effective resource allocation in 6G Internet of Things (IoT) systems. Secondly, the proposed model explores artificial intelligence techniques to analyze the network environment and predict anomalies in the fault tolerance approach. Lastly, data is kept private and maintained in integrity using a private blockchain, providing a more reliable, distributed, autonomous system with minimal overhead. Using synthetic data, the proposed model is validated against QGA-ACO and MER-ODLADT solutions for energy consumption, anomaly detection, and time-to-failure metrics across dynamic scenarios.

Human Activity Recognition (HAR) has become a cornerstone in the dynamic development of the Industrial Internet of Things (IIoT). This study introduces an extensive framework aimed at embedding HAR functionality into industrial settings to promote workplace safety, streamline operations, and facilitate predictive maintenance. Through AI techniques, HAR problems can be achieved with high accuracy. However, traditional AI models require centrally trained data on remotely powerful cloud servers. This leads to issues with privacy and security of health records and increases latency. To address this problem, the Federated Learning (FL) technique was proposed. FL allows distributed training on the patient’s IoT devices and serves as the communication mechanism between local devices and the FL aggregator. Thanks to this architecture, the health data needs only to be stored locally on its devices without being uploaded to data centers, thus ensuring security and reducing service response times and computational costs. In this study, we implement FedANN and FedConvNN independently in a federated learning setting to address human activity recognition problems toward real-time applications. Finally, we evaluate the effectiveness of the based on the variation initiation of the number of different training clients. The results show that the FedConvNN solution improves accuracy and reduces model and communication complexity compared to FedANN and centralized training models, with the potential for real-time deployment in HAR tasks. Our code is available on our GitHub repository: https://github.com/itsminhcs/Fedavg-HAR.git.

The integration of blockchain into industrial environments promises secure and verifiable data exchange; however, existing permissioned blockchain (PBC) frameworks, such as Hyperledger Fabric and Quorum, impose overheads that are unsuitable for resource-constrained systems. This article introduces LowCapChain (LCC), a lightweight PBC designed for manufacturing nodes, such as programmable logic controllers, robotic arms, and smart sensors. LCC integrates Merkle ledger compression, elliptic-curve cryptography-based proof-of-membership, and delegated fault-tolerant consensus to enable secure operations without full ledger replication. Implemented in a robotic metal stamping facility using Raspberry Pi edge nodes, LCC achieves 964 transactions per second with 108 ms consensus latency and memory usage below 55 MB. Compared with Hyperledger Fabric, LCC reduces mean consensus latency by 61% and cryptographic overhead by up to 81%. Compared with Quorum, the reductions are 54% and 76%, respectively. Scalability tests confirmed near-linear throughput growth across 10–200 nodes, and fault tolerance experiments verified block finalization under validator failure. These findings establish LCC as an efficient architecture for embedded industrial systems, offering a pathway toward scalable Industry 4.0 adoption with energy savings inferred from reduced CPU cycles rather than directly measured power.

As next-generation communication networks increasingly rely on AI-driven automation, ensuring robust and secure intrusion detection becomes critical, especially under limited labeled data. In this context, we introduceMulti-Space Prototypical Learning(MSPL), a few-shot intrusion detection framework that improves prototype-based classification by fusing complementary metric-induced spaces (Euclidean, Cosine, Chebyshev, and Wasserstein) via a constrained weighting mechanism. MSPL further enhances stability through Polyak-averaged prototype generation and balanced episodic training to mitigate class imbalance across diverse attack categories. In a few-shot setting with as few as 200 training samples, MSPL consistently outperforms single-metric baselines across three benchmarks: on CICEVSE Network2024, AUPRC improves from 0.3719 to 0.7324 and F1 increases from 0.4194 to 0.8502; on CICIDS2017, AUPRC improves from 0.4319 to 0.4799; and on CICIoV2024, AUPRC improves from 0.5881 to 0.6144. These results demonstrate that multi-space metric fusion yields more discriminative and robust representations for detecting rare and emerging attacks in intelligent network environments.

The development of the healthcare industry is closely tied to the history of human development. The integration of sensing, communication, computing, and control technologies, along with cloud-based solutions, enables the realization of the Internet of Things concept and forms a series of Internet of Healthcare Things (IoHT) applications. However, providing realtime health services is one of the most important challenges for IoHT systems. To address this issue, a fog computing architecture (FC) is proposed as an additional computing layer to support the cloud computing layer, aiming to reduce service response time, computing costs, and energy consumption. In this study, we conduct a comprehensive evaluation to optimize resource allocation for hospitals under the constraints of scale and patient volume, as well as SLA thresholds. Then, we provide recommendations to optimize investment costs for computing infrastructure supporting real-time health services. Finally, we discuss challenges and open issues.

The rapid expansion of IoT and edge networks has increased vulnerability to security threats, notably Distributed Denial-of-Service (DDoS) attacks, which target the application layer and can overwhelm networks. This paper introduces a novel two-stage DDoS detection pipeline for IoT and edge networks that utilizes large language model (LLM) embeddings for textual protocol fields along with numeric features like TCP handshake metrics. The first stage accurately identifies traffic as Normal or DDoS, while the second stage classifies confirmed DDoS flows into specific types (UDP, ICMP, TCP, HTTP), correcting any initial false positives. Comprehensive testing on a specialized IoT dataset demonstrated perfect detection rates of 100% in the first stage and 99.91% in subclass classification, highlighting the effectiveness of combining LLM-based textual data with traditional numeric indicators for advanced intrusion detection.

The rapid proliferation of smart medical sensors and wearable healthcare devices has significantly increased the demand for low-latency computing services. The data generated by these monitoring devices must be processed efficiently by edge computing systems. However, the dynamic and unpredictable nature of incoming computational tasks requires real-time, robust, and priority-sensitive task allocation mechanisms. This study introduces an on-line task allocation algorithm for edge computing in healthcare systems, leveraging Deep Q-Learning to address these challenges. A multi-objective optimization framework is formulated, enabling the task allocator to enhance system efficiency without requiring prior knowledge of future tasks. The algorithm maximizes the number of allocated tasks while prioritizing critical ones and minimizing counterproductive operation times. A deep Q-Network is implemented to estimate Q-values for each state-action pair, empowering the task allocator to select actions with the highest Q-value consistently. The performance of the proposed algorithm is rigorously evaluated in a simulated environment. The results show that the algorithm effectively learns an optimal policy for the task allocation problem, producing an average improvement of 74.6% in the overall performance of the system.

Free-space optical (FSO) communication systems face significant challenges in maintaining image fidelity under atmospheric turbulence, particularly for high-bandwidth RGB applications. This paper introduces a novel dual-polarized (DP) FSO transmission framework enhanced by particle swarm optimization (PSO) to achieve turbulence-resilient RGB image recovery. By exploiting polarization-division multiplexing, the system doubles transmission capacity while employing a physics-informed PSO algorithm to dynamically optimize a multi-stage correction pipeline comprising Wiener deconvolution, gamma correction, histogram matching, and non-local denoising. The proposed method uniquely correlates optimized restoration parameters with propagation distance, enabling simultaneous image recovery and ranging under Gamma-Gamma turbulence modeling. Experimental validation across weak and strong turbulence conditions $\left( {C_n^2 = {{10}^{ - 17}} - {{10}^{ - 13}}{{\text{m}}^{ - 2/3}}} \right)$ demonstrates significant improvements in structure-similarity-index-measure (SSIM), particularly achieving 0.4397 SSIM recovery for Llama images at 2.9 km under strong turbulence and a 119% improvement over uncorrected results. Comparative analysis shows superior performance to deep learning baselines, with 11.2% higher SSIM at 2 km distances. By combining turbulence mitigation with accurate ranging, this framework advances high-fidelity FSO imaging for remote sensing and surveillance applications.

The healthcare industry has experienced remarkable growth in data generation and revenue, making it the most rapidly expanding sector. To enhance security measures, our study explores the adoption of blockchain technology, leveraging its various aspects such as decentralization, consortiums, Ethereum, and Hyperledger. This paper proposes a Secured Healthcare Framework utilizing blockchain technology, with a focus on securing Electronic Health Records (EHR) through smart contracts. This approach ensures end-to-end security and non-repudiation. By integrating IoT devices such as RFID and Arduino, our method not only enhances security but also stream-lines data management within healthcare settings. The results demonstrate significant efficacy in the secure transmission and management of patient medical records. Moreover, experimental findings indicate an 89.88 % difference in latency between the MetaMask and Ganache environments, and a 13.04 % difference in gas usage for transactions in the Remix and Ganache environments. These findings highlight the potential of our proposed framework to improve security and efficiency in healthcare data management.

Reconfigurable Intelligent Surfaces (RIS) and Cognitive Radio (CR) offer solutions to improve spectral efficiency and mitigate spectrum scarcity. This article proposes a hybrid Dynamic Spectrum Management (DSM) paradigm combining Opportunistic Spectrum Access (OSA) and Concurrent Spectrum Access (CSA) models for spectrum allocation in RIS-supported CR environments. We frame the spectrum access issue as a game where each secondary user (SU) acts as a player to secure a communication channel. Our findings reveal that RIS technology significantly enhances spectrum allocation efficiency, quality of wireless communications, and spectrum sensing performance, providing high channel capacity and energy efficiency.

Software modeling relies heavily on UML class diagrams, essential tools for structuring a system’s entities, behaviors, and relationships. Yet, manually developing them from textual specifications remains a time-consuming task and subject to interpretation. This study proposes the creation of a corpus annotated according to a customized IOB schema, intended to train Named Entity Recognition (NER) models for the automatic extraction of UML elements from text. The schema integrates specific labels to accurately capture classes, attributes, methods, and relationships (association, aggregation, composition, inheritance), including their compound forms. The current corpus, built from 132 documents from various sources, includes more than 900 sentences and 11,000 manually annotated tokens. Particular attention was paid to the syntactic and semantic diversity of the texts, as well as to the linguistic quality, to ensure good generalization of the models. The empirical evaluation conducted with six Transformers models (BERT, RoBERTa, SpanBERT, XLNet, MiniLM and Electra) shows promising results, especially for classes and their relationships. This work thus lays the foundation for a reliable automation of UML class diagram generation from textual specifications, with strong potential for integration into software engineering environments and MDA processes.

Glaucoma, projected to impact over 110 million individuals by 2040, remains a significant challenge in low-resource settings due to limited access to diagnostic tools and physicians, delaying early detection. Automated glaucoma diagnosis systems have emerged as potentially scalable solutions but often rely on single-modality imaging and lack integration with real-world patient interactions. To bridge these gaps, DeepRet, an AI-based system, leverages multimodal data—retinal imaging, age, and intraocular pressure measurements—to achieve 96.34% accuracy, a 93% AUROC, and a 17.30% improvement over traditional methods. Designed for low-resource environments, DeepRet features an affordable, portable retinal imager with a manufacturing cost under $200, 3D printed materials, and low-cost electronics. Its intuitive design minimizes training requirements for health workers while offering real-time, patient-friendly feedback, enhancing accessibility and the patient experience.

Soil moisture prediction requires the integration of multisource data, including satellite observations, ground-based sensors, and airborne systems, each contributing critical information for modeling Earth’s hydrological cycles. The complexity of this task necessitates an analytical framework capable of reconciling general modeling principles with the intricate variability of climatic factors to ensure reliable predictions. This study explores the application of Spiking Neural Networks (SNNs) as an advanced approach beyond conventional methodologies, utilizing array-sensed data from the ERA5 dataset. SNNs are distinguished by their ability to merge computational efficiency with biologically inspired dynamics, employing Leaky Integrate-and-Fire neurons to process spatial and temporal information effectively. The model’s adaptability and precision in handling large-scale climatic datasets were evaluated using an 80-20 data split, achieving a Mean Squared Error (MSE) of 0.0003, an R 2 value of 0.8919, and a Pearson correlation coefficient of 0.9449, reinforcing its predictive capability and ability to capture intrinsic dependencies within soil moisture dynamics. This novel implementation of SNNs enhances prediction accuracy while offering a computationally efficient solution for soil moisture forecasting, addressing key challenges in environmental and agricultural applications. The findings provide a foundation for future research aimed at optimizing hydrological models through biologically inspired neural architectures.

This paper proposes a novel framework for climate anomaly detection by integrating antenna array-derived environmental data with Spiking Neural Networks (SNNs), a biologically inspired computational approach. The methodology systematically captures key climatic parameters, including temperature, wind, and moisture, allowing a precise and structured analysis. By encoding temporal and spatial dynamics, SNNs provide an advanced mechanism for detecting subtle patterns and anomalies that conventional methods may overlook. The study demonstrates the effectiveness of simulated antenna-based climate data in identifying temperature trend anomalies, reinforcing the potential of neural architectures in climate variability analysis. The results highlight the ability of SNNs to improve anomaly detection through efficient processing of time-sensitive and spatially complex datasets. The proposed pipeline not only addresses methodological gaps but also improves the broader understanding of climate disturbances by incorporating innovative technology-driven solutions. This research underscores the importance of integrating computational intelligence into climate studies, contributing to more accurate and scalable environmental monitoring systems.

The tourism sector has recently undergone a significant transformation with the integration of chatbots, enabling users to interact with services through natural language. At the heart of these systems lies the Natural Language Understanding (NLU) component, which processes user input through intent classification and entity extraction. A major challenge, however, is selecting the most effective machine learning method to build robust NLU systems tailored to tourism applications. This study evaluates the performance of various machine learning algorithms for intent classification in tourism-focused chatbots. The models under investigation include Support Vector Machine (SVM), LightGBM, XGBoost, and Random Forest. A tourism-specific dataset was developed for this comparative analysis, with evaluation based on metrics such as accuracy and weighted F1-score. The experimental results indicate that XGBoost, LightGBM, and Random Forest achieve the highest training accuracy in intent classification. These outcomes offer valuable insights for developing effective NLU components in tourism chatbots, improving their ability to interpret user queries accurately.

Distributed Denial of Service (DDoS) attacks continue to escalate in both frequency and sophistication, often overwhelming critical network infrastructures. While deep learning methods excel at recognizing malicious patterns, their lack of transparency undermines trust and hampers effective mitigation. This paper introduces a unified, on-premise pipeline that integrates an advanced flow based attack classifier with a local large language model (LLM) to deliver explainable, real-time DDoS defense. The proposed approach detects threats at the flow level, rapidly fags suspicious traffic, and then generates human-readable analyses and device specific countermeasures ranging from firewall rules to intrusion prevention system signatures all without transmitting data of-site. Through comprehensive testing on diverse, large scale network traces, we demonstrate that this framework not only achieves near-perfect detection accuracy but also considerably reduces operational costs and privacy risks associated with external cloud services. Furthermore, evaluators confirm the clarity and correctness of the automatically generated mitigation strategies, highlighting the system’s practicality in enterprise environments. Overall, our results validate on-premise, LLM-enhanced DDoS defense as a robust, transparent, and economical solution for safeguarding modern network ecosystems.

Chatbots are tools designed to interact with users through natural language. They are widely used in various sectors, such as education, tourism, and transportation. These systems perform several common tasks, such as enhancing customer service, providing information permanently, and answering frequently asked questions. They can be classified into two main categories: rule-based chatbots like Eliza, which rely on predefined rules and intents to handle specific tasks, and AIbased chatbots like ChatGPT, which use advanced technologies like deep learning and Natural Language Processing (NLP) to interpret and respond to user inputs dynamically. However, their development faces challenges due to constraints specific to their development tools, such as high-cost NLP services. In addition, the absence of a dedicated chatbot development platform for the transportation domain remains a significant limitation. To bridge this gap, this study conducts a comparative analysis of existing metamodels for chatbot development and identifies their concepts and relations. The outcome is the design of a unified metamodel specifically tailored to the transportation sector, serving as the abstract syntax for constructing a DomainSpecific Language (DSL) that accelerates chatbot development for the transportation domain and reduces costs associated with NLP services.

The need for high-bandwidth, low-attenuation communication in underwater environments is driving the development of new technologies. The Internet of Underwater Things (IoUT) addresses this need by enabling underwater devices to communicate, sense, and transmit data. Hybrid optical communication systems, combining Underwater Optical Wireless communication (UOWC), Single-Mode Fiber (SMF), and Free-Space Optics (FSO), have emerged as promising solutions to overcome these challenges and achieve high-speed, long-distance transmission. In this study, we propose a hybrid UOWC-SMF-FSO system utilizing a Photodetector, Remodulate, and Forward Relay (PRFR) for efficient wavelength conversion from 532 nm (visible spectrum) to 1550 nm (infra-red spectrum) at the water-fiber-air interface. System performance is evaluated under varying underwater distances, data rates, and water types. Key metrics such as Bit Error Rate (BER), Quality-factor (Q-factor), and eye diagrams are used to analyze the robustness and reliability of the proposed system. The results demonstrate that the hybrid system achieves at BER below threshold$(3.8\times 10^{-3})$an overall transmission range of up to 2035 m ($35\mathrm{m}$underwater range + 1000 m SMF length$+1000\mathrm{m}$FSO range) in Pure Seawater (PW) at 10 Gbps, with decreasing ranges for Clear Ocean (CL) and Coastal Ocean (CS) water types due to higher attenuation. These findings pave the way for the development of high-capacity underwater sensor networks and contribute significantly to the advancement of IoUT application.

This paper presents a study of Radio Frequency LoRa device classification performance under challenging channel's variation using a hybrid CNN-LSTM-Attention neural architecture. By addressing the temporal dynamics introduced by Doppler effects in mobile scenarios, combined with targeted data augmentation strategies, our method achieves 99.6% classification accuracy across 10 devices in stationary conditions and maintains robust performance of 85.8% even under high mobility conditions (100 Hz Doppler shift). The proposed hybrid architecture leverages convolutional layers for spatial feature extraction, LSTM layers for modeling temporal dependencies in RF emissions, and an attention mechanism to focus on the most discriminative temporal segments of the signal. Our experimental results, conducted using a dataset of 30 commercial LoRa IoT devices, demonstrate significant performance improvements over state-of-the-art approaches, particularly in challenging mobile environments where Doppler effects typically degrade classification reliability. The model maintains 93.1% accuracy in Line-of-Sight (LOS) mobile scenarios and 87.6% in Non-Line-of-Sight (NLOS) mobile environments, outperforming previous approaches by 3.1% and 2.6% respectively. This work contributes to the field of physical-layer security by demonstrating how temporal modeling techniques can enhance RF fingerprinting performance in realistic mobile deployment scenarios.

Smart grids are confronted with growing cybersecurity threats by Advanced Persistent Threats (APTs) targeting vulnerabilities of cyber-physical systems with stealthy, multi-stage attacks. Conventional signature-based rule-driven detection mechanisms cannot detect these advanced threats. This paper presents an active behavior detection system using Generative Adversarial Networks and Autoencoders for benign network behavior modeling and anomaly detection characteristic of APTs. Tested on actual smart grid data, our hybrid solution is 96.5 % accurate and has an$\text{F 1}$-score of$\text{96.59 \%}$, surpassing baseline MLPs and state-of-the-art techniques. The main innovations are adversarial training for generating attack patterns and Autoencoder reconstruction for anomaly detection. Experiments show the framework's robustness to stealthy APTs with few false positives. This research propels adaptive defense technologies for critical infrastructure, bridging the gaps in scalability and dynamic threat modeling.

ABSTRACT Chatbots have become widely adopted tools for improving user interactions across multiple platforms. They are advanced software applications designed to emulate human conversation across various platforms. Moreover, developing chatbots using existing platforms and frameworks presents challenges, such as the lock‐in of NLP services, and incurs substantial costs. Recently, research has introduced solutions to ease chatbot development. Many of these approaches utilise Model‐Driven Engineering (MDE) and Domain‐Specific Languages (DSLs) to automate processes and simplify implementation. Through the use of MDE and DSLs, these solutions enhance efficiency and make chatbot creation more accessible. This study aims to provide a comprehensive survey on MDE and DSLs in chatbot development, highlighting key research topics, opportunities, and challenges. The first contribution explores the primary application domains of DSLs in chatbot development and the associated challenges in their adoption. Second, this work examines the various ways in which DSLs are employed to model and develop chatbots, assessing their impact on automation and efficiency. Additionally, this study identifies the challenges and limitations of using DSLs in chatbot development. Atlast, it investigates the influence of DSL utilisation on user experience, both from the perspective of chatbot developers and end‐users, to determine how DSLs enhance the chatbot development process and interaction quality. To achieve this, a comprehensive search will be conducted across Scopus, Web of Science, and ScienceDirect for studies published between 2014 and 2024. A total of 306 publications were reviewed, of which 15 were identified as primary studies.

ABSTRACT The tourism sector is adopting smart solutions to offer visitors more personalised and sustainable experiences. By leveraging urban infrastructure and new technologies, tourist destinations aim to enhance the interaction between travellers and their environment. Artificial intelligence (AI) and natural language processing (NLP) play a key role in this transformation, particularly through chatbots. They are AI‐driven applications designed to simulate human‐like conversations, enabling users to interact with digital services through text or voice interfaces. In the tourism sector, they facilitate real‐time access to information and services, improving the visitors' experience. These applications typically rely on intent recognition APIs, which may be proprietary, requiring access fees and potentially leading to high implementation costs. This study explores the use of a domain‐specific language (DSL) dedicated to chatbot development for smart tourism. The first contribution comprises various research topics and emerging technologies used to improve smart tourism experiences and their impact on key tourism components such as attractions, accessibility, amenities, activities, available packages, and ancillary services. Second, this work aims to present the key concepts of model‐driven engineering involved in constructing a DSL and to introduce our approach to building a DSL, with a focus on presenting the DSL metamodel. Third, this study identifies the challenges and limitations of using DSLs in chatbot development.

Abstract Conversational assistants (CAs) and Task‐oriented ones, in particular, are designed to interact with users in a natural language manner, assisting them in completing specific tasks or providing relevant information. These systems employ advanced natural language understanding (NLU) and dialogue management techniques to comprehend user inputs, infer their intentions, and generate appropriate responses or actions. Over time, the CAs have gradually diversified to today touch various fields such as e‐commerce, healthcare, tourism, fashion, travel, and many other sectors. NLU is fundamental in the natural language processing (NLP) field. Identifying user intents from natural language utterances is a sub‐task of NLU that is crucial for conversational systems. The diversity in user utterances makes intent detection (ID) even a challenging problem. Recently, with the emergence of Deep Neural Networks. New State of the Art (SOA) results have been achieved for different NLP tasks. Recurrent neural networks (RNNs) and Transformer architectures are two major players in those improvements. RNNs have significantly contributed to sequence modelling across various application areas. Conversely, Transformer models represent a newer architecture leveraging attention mechanisms, extensive training data sets, and computational power. This review paper begins with a detailed exploration of RNN and Transformer models. Subsequently, it conducts a comparative analysis of their performance in intent recognition for Task‐oriented (CAs). Finally, it concludes by addressing the main challenges and outlining future research directions.

ABSTRACT Chatbots have emerged as ubiquitous tools for enhancing user interaction across various platforms, from customer service to personal assistance. They are computer programs that simulate and process human conversation, either written, spoken or both. However, developing efficient chatbots remains a challenge, primarily due to the intricate nature of critical components of chatbots like natural language understanding (NLU) requiring a subscription from intent recognition providers like Dialogflow and Amazon Lex. This makes chatbots closely linked to NLP services and can be locked in. Recently, various research studies have provided solutions to reduce the workload of developers and designers. These approaches have proposed model‐driven development via domain‐specific languages (DSLs), which make the chatbot development process more accessible and more automated. This advancement aims to enhance effectiveness in chatbot development by leveraging DSLs. This study aims to provide a comprehensive overview of DSLs for developing chatbots, with the first contribution comprising various research topics, tools, approaches, and technologies employed to implement DSLs. Second, this work aims to assess and contrast the primary DSLs currently available for chatbot development, focusing on presenting the key elements used in constructing these DSLs. Third, this study identifies the challenges and limitations of using DSLs in chatbot development.

In the rapidly advancing field of intelligent transportation systems, integrating artificial intelligence (AI) with edge computing presents a promising way to enhance the safety and efficiency of the Internet of Vehicles (IoV). This study explores and presents a deep learning-based object detection model within an edge computing framework which aims to facilitate real time object detection in self driving cars. Using an urban traffic scenarios-based dataset, our research shows the ability of the model to accurately detect and classify various objects important for autonomous driving. The YOLOv8 model is used in this work due to its optimal balance between accuracy and computational efficiency. This model has also demonstrated its worth by achieving good performance results, including an average precision of 0.79, a recall of 0.62, and an F1-score of 0.69. The results are demonstrated by a detailed confusion matrix, highlighting the model’s effectiveness in complex driving environments and underscoring its reliability for in-vehicle deployment. By implementing AI directly on edge devices within vehicles, our approach might be helpful in significantly reducing latency, boosting decision-making speed, and enhancing data privacy by minimising dependence on cloud processing. The findings not only support the model’s capabilities but also illustrate the practical benefits of edge intelligence in autonomous vehicles. These benefits, such as faster decision making and improved data privacy, contribute effectively to the IoV infrastructure. This study marks a substantial step toward recognizing the possibility of AI-enhanced edge computing in driving the next generation of autonomous vehicle technology.

This study explores the effectiveness of Convolutional Neural Networks (CNNs) in automatically classifying skin cancer for e-health applications. The trained model showcases impressive performance by leveraging the HAM10000 dataset, which includes a wide range of skin lesion images from seven different classes. The parameters and architecture of the CNN model are presented in a systematic manner, providing valuable insights into the reasoning behind its design. The model is optimized using the Adam optimizer and annealing techniques to ensure efficient convergence. The model’s performance is assessed on validation and test datasets, showcasing an accuracy of 78.55% and 76.49%, respectively, for skin cancer classification. This study highlights the significant potential of CNN as a powerful tool for automating the diagnosis of skin cancer, which is in line with the growing trend of using deep learning for medical image analysis.

The rapid proliferation of intelligent Internet of Health Things (IoHT) applications within the context of the COVID-19 pandemic has exerted significant strain on the backhaul network infrastructure. This paper aims to introduce a novel framework that leverages software-defined networking (SDN) to enhance edge computing capabilities. This framework will expect to facilitate dynamic and adaptable communication between edge and cloud servers, specifically designed to support real-time Internet of Health Things applications. Through the establishment of a connection between servers and the Software-Defined Networking controller, the system is expected to facilitate load balancing, network optimization, and the utilization of resources in an efficient manner. This, in turn, enables the provision of real-time healthcare services. Ultimately, the efficacy of the suggested framework is assessed by analyzing its impact on service response time. The research results indicate that the proposed framework significantly benefits IoHT systems’ service response times across various workloads and traffic.

This study presents and investigates a free space optics (FSO) transmission technique that utilizes polarization division multiplexing (PDM) and orbital angular momentum (OAM) multiplexing. The aim is to provide an effective and reliable backhauling solution. The transmission system uses four-level-pulse amplitude modulation (PAM-4) signals achieving a high capacity of 400 Gbps. Two orthogonal polarized beams of 1550 nm Laser diode are used. For each polarized beam, 4 OAM beams are further generated, and each OAM beam transports 50 Gbps data using PAM-4 modulation over the free space channel. The impact of varying FSO ranges on the system performance under clear weather sky (sunny), light rain (LR), medium rain (MR), and heavy rain (HR) is considered. The bit error rate (BER) metric is used to evaluate the proposed system's performance. Results from the simulation indicate that transmission of 400 Gbps is achieved successfully with a range of 1800 m, 900 m, 775 m, and 537 m under clear weather, LR, MR, and HF, respectively, with an acceptable Log(BER) of -2.42 for ultra-FEC limit.

Skin cancer diagnosis, a critical task in the medical domain, can be revolutionized through the application of advanced deep-learning techniques. This work investigates the efficacy of Convolutional Neural Networks (CNNs) in the automated classification of skin cancer. The process begins with a comprehensive explanation of key CNN layers: Conv2D, MaxPool2D, Dropout, and Dense. The Conv2D layers employ learnable filters that transform localized image segments, while MaxPool2D contributes to downsampling, effectively reducing computational cost and overfitting risk. Integrating these layers enables the network to capture local and global characteristics, which is crucial for accurate classification. Adding Dropout layers enhances generalization and mitigates overfitting by introducing randomness during training. ReLU activation functions infuse non-linearity, and the Flatten layer facilitates the transition to fully connected layers. The proposed CNN architecture is meticulously designed considering filter counts, kernel sizes, and pooling dimensions. The trained model demonstrates promising performance by utilizing the HAM10000 dataset, encompassing diverse skin lesion images across seven classes. The CNN model’s parameters and architecture are systematically presented, offering insights into its design rationale. The model undergoes optimization with the Adam optimizer and annealing techniques to facilitate convergence. The model’s effectiveness is evaluated on validation and test datasets, demonstrating an accuracy of 78.55% and 76.49%, respectively, for skin cancer classification. Data augmentation strategies are introduced to enhance model generalization further. The results underscore CNN’s potential as a robust tool for automating skin cancer diagnosis, aligning with the broader trend of leveraging deep learning for medical image analysis

This work addresses the challenge of classifying visually similar bird species, a task complicated by subtle interspecies variations. We focused on ten bird species, assembling a dataset of approximately 8000 images from Google Images. These species were specifically chosen for their high degree of similarity, presenting a unique challenge for classification algorithms. To enhance our dataset and improve classification accuracy, we employed Generative adversarial networks (GANs), a state-of-the-art generative adversarial network, to augment our original dataset with synthetic yet realistic images. This augmentation aimed to provide a more prosperous, diverse training environment for our deep learning model. Subsequently, we developed a specialized multi-classification model tailored to recognize and differentiate these closely related bird species. Integrating GANs like StyleGAN3-augmented data into our training process represents a novel approach to ecological image analysis, potentially setting a new standard for accuracy and efficiency in classifying highly similar species. This study demonstrates the effectiveness of advanced generative models in complex classification tasks and contributes a valuable methodology to ecological research and species identification.

This study delves into the challenge of classifying visually similar bird species, an area of significant interest in the field of fine-grained image classification. Utilizing a substantial dataset comprising images of ten bird species which was selected carefully to challenge the model to classify species of extreme similarities. To achieve this, we were keen to collect the data with subtle visual dissimilarities and of different positions taken for these birds. The research explores the potential of deep learning techniques to differentiate species based on subtle inter-species variations. This task is particularly demanding due to the minimal yet critical differences between these closely related species. Our research leveraged a unique deep learning model using convolutional neural networks (CNNs) to accurately classify birds with minimal visual differences. This innovative approach marks a significant step forward in machine learning for biological classification, with implications for biodiversity and ecological conservation. Our study demonstrates the effectiveness of our deep learning model in accurately classifying bird species, showcasing the potential of advanced techniques in complex Classification tasks. This research enhances the use of computational methods in biodiversity and ecological conservation. Additionally, it underscores the importance of birds as indicators of environmental changes, such as climate shifts, aiding in early detection of potential ecological issues.

The utilization of Artificial intelligence (AI) and other cutting-edge techniques in the field of medical image analysis has exhibited significant potential. Nevertheless, a significant obstacle that impedes the extensive implementation of these models in the healthcare sector is their restricted interpretability. The concept of explainability is a subject of extensive discussion and debate within the context of utilizing Artificial intelligence in the healthcare domain. Notwithstanding the empirical evidence demonstrating the superior performance of AI-driven systems compared to humans in certain analytical tasks, particularly in the field of medical image computing, these systems still encounter challenges due to their limited explainability. The present study provides a comprehensive assessment of the significance of explainability in the field of medical Artificial intelligence and performs an ethical analysis of the influence of explainability on the incorporation of AI-driven tools in data engineering in medicine and health care. The paper examines various subjects including data security, confidentiality, privacy, fairness, and discrimination, among others.

This research paper presents a novel formal framework designed for piloting human resources performance and fostering agile HR management within organizations. The framework facilitates the systematic collection of HR data, enabling the computation of sensitive Key Performance Indicators (KPIs) essential for predictive analytics and data-driven decision-making. Through this framework, top management gains a clear understanding of recruitment and training strategies, as well as the ability to distinguish between easily-replaceable and critical resources. The framework empowers organizations to optimize resource allocation, enhance operational efficiency, and mitigate risks associated with human capital management. The integration of predictive analytics enables the development of comprehensive dashboards, providing actionable insights to guide strategic HR initiatives and ensure organizational success in dynamic environments.

Zero-shot learning (ZSL) is a machine learning technique that identifies the target classes without any observed data, using semantic information from some source classes as the basis of knowledge transfer. ZSL has emerged as a new paradigm in machine learning to solve the constraints of classical supervised learning. This study explores applying ZSL techniques to a plant species dataset containing rich attribute descriptions for each class. Our study uses the multimodal capabilities of Contrastive Language–Image Pre-training to predict unseen plant species classes without needing labeled examples during training. Our objective is to contribute to advancing biodiversity conservation efforts and promoting environmental sustainability by facilitating the automated recognition of plant species based on semantic attributes.

The creation of software documentation is widely recognized as a critical and demanding undertaking within the rapidly changing realm of software development. This study introduces a novel method for generating software documentation by leveraging Large Language Models (LLM). The paper presents a novel system that extracts use cases from UML Use Case Diagrams and employs a Generative AI Model to generate descriptive text for each extracted use case. This approach aims to reduce the amount of time dedicated to documentation and encourage uniformity in the description of software functions. The results suggest that the level of manual labor and time needed can be substantially decreased by upholding elevated levels of clarity and comprehensiveness in software documentation. This study presents a use-case scenario that showcases the practical application of our methodology in real-world situations. The purpose of this example is to demonstrate the practicality and effectiveness of the method.

In today’s world, chatbots have become a significant advancement in Artificial Intelligence (AI). They are extensively utilized to provide users convenient access to 24/7 services using natural language. The development of these conversational applications is evolving rapidly and necessitating specific knowledge and practical experience to successfully exploit all the functionalities of chatbot development platforms and frameworks. The heterogeneity of chatbot development tools and their need for NLP services makes it challenging to build chatbots. Thus, one possible solution to these problems is to construct a domain-specific language (DSL) to accelerate the development of Chatbots. A Domain Specific Language (DSL) is a programming language that provides expressive power within a specific problem domain by using appropriate abstraction notations. Abstract syntax, concrete syntax, and semantics are the three components that describe it. Furthermore, it is necessary to utilize generation templates to construct a chatbot for an already established platform. Through the use of a Model-Driven Architecture (MDA), which is an approach that focuses on modeling software systems at different levels of abstraction, from high-level requirements to platform-independent designs, this work aims to define a sketch of an independent language of the chatbot development platform by providing the components needed for our DSL, like metamodel for modeling conversations and developing transformations between models to generate the source code for a chatbot conforming to a specific implementation platform. This will facilitate the automatic generation of code.

Each new wave of technological progress sparks debates about the effects of automation on the future of employment. Current debates on artificial intelligence (AI) and employment are reminiscent of those raised by mechanization in the 19th century, the generalization of electricity, and the introduction of computers in the 20th century: some consider new technologies as a way to relieve workers of the most challenging tasks, and others are alarmed by the imminent threat to employment. This article aims to contribute to the ongoing debate on the potential changes that may arise from the recent emergence of Generative AI in job markets. It is based on a historical analysis of technological revolutions and a literature review of technology’s impact on employment. The purpose of this study is not to gather general statistics but rather to analyze potential changes and help design suitable policy responses. This analysis will also consider the possible impact on job quality. The study emphasizes the potential implications for various professional categories but does not predetermine the outcomes of technological transition. The decision to incorporate such technologies is driven by humans, and it is their responsibility to guide the transition process.

Facial Expression Recognition (FER) is a crucial aspect in various domains, given its significance in understanding human emotions. However, designing efficient FER systems entails addressing challenges in feature extraction and selection. While previous studies have primarily focused on static feature selection methods, these approaches often struggle with spontaneous expressions due to the unique facial characteristics of each individual. To address this challenge, we implemented a Facial Morphology-Guided Feature Selection Method that combines texture features using Local Binary Pattern histograms and geometric features employing linear and eccentricity features. Subsequently, we employ Recursive Feature Elimination (RFE) and Binarized Genetic Algorithm (BGA) algorithms for feature selection, combining their outputs to identify the optimal subset of features tailored for each face. Experimental validation using the CK+ and DISFA datasets demonstrates the effectiveness of our approach in enhancing facial expression recognition accuracy.

The 6th generation mobile networks (6G) are expected to be launched in the 2030s. The architecture of 6G will be the integration of heterogeneous mobile networks. One of the indispensable components of 6G is unmanned aerial vehicle (UAV) communication networks. Powerful support capabilities with high mobility and flexibility are the main driving forces behind the breakout of UAV networks into countless domains to enhance the quality of human life such as disaster rescue, agriculture, military, and smart cities. However, high mobility and real-time services also pose significant challenges for communication protocols. In this work, we consider in detail the requirements of the modern 6G UAV communications and then develop a more realistic simulation environment based on NS3, to evaluate the performance of traditional routing protocols consisting of AODV, DSR, and OLSR on aspects of delay, packet delivery ratio, and routing overhead. We hope that these quantitative results will be an important guide for selecting suitable routing protocols in different 6G UAV scenarios.

This paper examines the critical function of Field-Programmable Gate Arrays (FPGAs) in speeding Spiking Neural Networks (SNNs) for real-time edge neuromorphic computing. Our work systematically evaluates the integration of FPGA technology for the optimization and speeding of SNN models. The analysis covers the power efficiency, low latency processing, and parallelism that are intrinsic benefits of FPGAs, emphasizing their relevance for edge computing applications. We discuss the smooth transfer of trained SNN models to FPGA platforms. Using an extensive analysis of state-of-the-art architectures, we demonstrate the efficiency benefits of using FPGA to accelerate SNNs. We derive more insights into the real-world applications of this FPGA-SNN integration in various fields. The analysis supports advances in edge computing and neuromorphic processing paradigms by adding to the collective knowledge of how FPGA enhances the real-time processing capabilities of Spiking Neural Networks.

Enhancing the security of wireless networks involves implementing a user authentication method when the fingerprint of a network device is unknown or considered a potential threat. This technique is known as radio frequency (RF) fingerprinting. This paper presents a novel method for RF fingerprinting of Internet of Things (IoT) devices, addressing the challenges of the radio frequency spectrum. The proposed architecture integrates a feature generator module that transforms time-series I/Q samples into a multi-dimensional matrix and a deep learning module inspired by the ResNet-50-1D model. We assess the effectiveness of our approach by analyzing a real-world dataset of BT emissions obtained from 10 commercial IoT devices in two challenging indoor environments. The datasets, made publicly accessible on IEEE Dataport, were gathered using a USRP X300 software-defined radio (SDR) in both line-of-sight (LoS) and rich multipath propagation scenarios. Our method showcases excellent results in the TTS scenario and shows promise in the challenging TTD scenario, considering the complex nature of frequency hopping. The evaluation results emphasize the significance of evaluating RF fingerprinting models in various scenarios and offer valuable insights into the strengths and limitations of our approach in handling radio frequency waveforms.

Denial of Service (DoS) attacks pose a significant threat to AI systems security, causing substantial financial losses and downtime. However, AI systems’ high computational demands, dynamic behavior, and data variability make monitoring and detecting DoS attacks challenging. Nowadays, statistical and machine learning (ML)-based DoS classification and detection approaches utilize a broad range of feature selection mechanisms to select a feature subset from networking traffic datasets. Feature selection is critical in enhancing the overall model performance and attack detection accuracy while reducing the training time. In this paper, we investigate the importance of feature selection in improving ML-based detection of DoS attacks. Specifically, we explore feature contribution to the overall components in DoS traffic datasets by utilizing statistical analysis and feature engineering approaches. Our experimental findings demonstrate the usefulness of the thorough statistical analysis of DoS traffic and feature engineering in understanding the behavior of the attack and identifying the best feature selection for ML-based DoS classification and detection.

Abstract Image fusion plays a significant role in computer vision since numerous applications benefit from the fusion results. The existing image fusion methods are incapable of perceiving the most discriminative regions under varying illumination circumstances and thus fail to emphasize the salient targets and ignore the abundant texture details of the infrared and visible images. To address this problem, a multiscale aggregation and illumination‐aware attention network (MAIANet) is proposed for infrared and visible image fusion. Specifically, the MAIANet consists of four modules, namely multiscale feature extraction module, lightweight channel attention module, image reconstruction module, and illumination‐aware module. The multiscale feature extraction module attempts to extract multiscale features in the images. The role of the lightweight channel attention module is to assign different weights to each channel so as to focus on the essential regions in the infrared and visible images. An illumination‐aware module is employed to assess the probability distribution regarding the illumination factor. Meanwhile, an illumination perception loss is formulated by the illumination probabilities to enable the proposed MAIANet to better adjust to the changes in illumination. Experimental results on three datasets, that is, MSRS, TNO, and RoadSence, verify the effectiveness of the MAIANet in both qualitative and quantitative evaluations.

Integrating Internet of Things (IoT) and artificial intelligence (AI) techniques have found widespread application in various fields, including smart cities, agriculture, and environmental monitoring. With the increasing availability of satellite imagery and other remote sensing data, deep learning algorithms can be used and trained to detect, classify, and segment flood regions in real time. In addition, deep learning techniques, such as convolutional neural networks (CNNs), have been successful in this field, enabling the automated analysis of vast amounts of satellite imagery. By combining AI-based flood detection with other data sources, such as meteorological forecasts and ground-based sensors, comprehensive flood monitoring systems that provide early warning of flood events and facilitate effective emergency response can be developed. In this article, we developed an image-based flood segmentation system called DeepLab that uses a deep learning algorithm to detect and segment the presence and extent of floods with high accuracy and speed. The neural network was trained on an extensive collection of satellite images, which were complemented by ground truth labels that indicated the presence of flooded areas. The trained DeepLabv3 model is applied to new satellite images during inference to forecast the likelihood of each pixel belonging to a flooded area. To do this, a binary flood map was generated from the pixel-level forecasts by incorporating a threshold into the output probabilities. The proposed system’s accuracy was high compared to the state-of-the-art methods, as evidenced by segmentation and experimental results. The segmentation accuracy achieved an overall score of 87%.

Millions of individuals around the world have been impacted by the ongoing coronavirus outbreak, known as the COVID-19 pandemic. Blockchain, Artificial Intelligence (AI), and other cutting-edge digital and innovative technologies have all offered promising solutions in such situations. AI provides advanced and innovative techniques for classifying and detecting symptoms caused by the coronavirus. Additionally, Blockchain may be utilized in healthcare in a variety of ways thanks to its highly open, secure standards, which permit a significant drop in healthcare costs and opens up new ways for patients to access medical services. Likewise, these techniques and solutions facilitate medical experts in the early diagnosis of diseases and later in treatments and sustaining pharmaceutical manufacturing. Therefore, in this work, a smart blockchain and AI-enabled system is presented for the healthcare sector that helps to combat the coronavirus pandemic. To further incorporate Blockchain technology, a new deep learning-based architecture is designed to identify the virus in radiological images. As a result, the developed system may offer reliable data-gathering platforms and promising security solutions, guaranteeing the high quality of COVID-19 data analytics. We created a multi-layer sequential deep learning architecture using a benchmark data set. In order to make the suggested deep learning architecture for the analysis of radiological images more understandable and interpretable, we also implemented the Gradient-weighted Class Activation Mapping (Grad-CAM) based colour visualization approach to all of the tests. As a result, the architecture achieves a classification accuracy rate of 0.96, thus producing excellent results.

This study examines the variability of upwelling along eastern boundaries, specifically focusing on two upwelling systems in the northern hemisphere: the Moroccan Atlantic coast and the Californian coast. In this study, we use a specialized archive that is both highly efficient and robust. The archive is specifically designed to tackle the challenge of detecting upwelling phenomena in satellite data. We ensure accurate and reliable results by prioritizing cloud processing in upwelling zones and questionable pixels. The description and analysis of upwelling dynamics from 2000 to 2019 include the interannual and seasonal variability, which are examined using different upwelling indices. Furthermore, this study proposes a novel index to evaluate the agreement between Sea Surface Temperature (SST) and Chlorophyll-a concentration (Chl-a) as indicators of upwelling. This index successfully captures the relationship between biological and physical components at various spatio-temporal scales, indicating its potential for use in other upwelling systems.

With the development of society, technological progress, and new needs, autonomous driving has become a trendy topic in smart cities. Due to technological limitations, autonomous driving is used mainly in limited and low-speed scenarios such as logistics and distribution, shared transport, unmanned retail, and other systems. On the other hand, the natural driving environment is complicated and unpredictable. As a result, to achieve all-weather and robust autonomous driving, the vehicle must precisely understand its environment. The self-driving cars are outfitted with a plethora of sensors to detect their environment. In order to provide researchers with a better understanding of the technical solutions for multi-sensor fusion, this paper provides a comprehensive review of multi-sensor fusion 3D object detection networks according to the fusion location, focusing on the most popular LiDAR and cameras currently in use. Furthermore, we describe the popular datasets and assessment metrics used for 3D object detection, as well as the problems and future prospects of 3D object detection in autonomous driving.

Object counting, defined as the task of accurately predicting the number of objects in static images or videos, has recently attracted considerable interest. However, the unavoidable presence of background noise prevents counting performance from advancing further. To address this issue, we created a group and graph attention network (GGANet) for dense object counting. GGANet is an encoder-decoder architecture incorporating a group channel attention (GCA) module and a learnable graph attention (LGA) module. The GCA module groups the feature map into several subfeatures, each of which is assigned an attention factor through the identical channel attention. The LGA module views the feature map as a graph structure in which the different channels represent diverse feature vertices, and the responses between channels represent edges. The GCA and LGA modules jointly avoid the interference of irrelevant pixels and suppress the background noise. Experiments are conducted on four crowd-counting datasets, two vehicle-counting datasets, one remote-sensing counting dataset, and one few-shot object-counting dataset. Comparative results prove that the proposed GGANet achieves superior counting performance.

Free space optical (FSO) communication offers an excellent opportunity to develop energy-efficient, secure, and ultrafast data links for 5G and beyond applications, including heterogeneous networks with massive connectivity and wireless backhauls for cellular systems. However, the effect of an optical beam's pointing inaccuracy combined with the impact of climate factors must be considered in the FSO communication system. In this paper, we first evaluate the performance reliability and availability of NRZ-based mode division multiplexing (MDM)-FSO backhaul. In particular, a single wavelength laser is used to transmit four different optical beams, using four different wavelengths. It also explores and classifies four beams used for capacity enhancement in mode division multiplexed MDM-FSO backhaul links. Several Machine Learning (ML) models are used to classify the four optical modes. Results indicate successful transmission of 80 Gbps. Furthermore, the primary findings indicate that the ML model exhibits an impressive accuracy rate of approximately 97% in classifying four distinct beams.

The development of medicine hand-in-hand with the history of humans. The advent of 5th-generation communication networks have realized the Internet of Things concept and formed a series of smart applications in almost domains such as health-care, agriculture, transportation, retails, etc. In these contexts, the Internet of Medical Things (IoMT) is one of the most attended domains, where the service response time is a key design factor. In this study, we consider the effectiveness of this framework and compare it with the cloud-based computing framework under varying changes in the arrival rate of service requests by queuing models. The simulation results have demonstrated that the proposed fog-to-cloud based computing scheme outperforms cloud-based computing schemes in terms of response time, and meets SLA requirements for real-time IoMT systems. Finally, we discuss challenges to realising real-time IoMT systems in the Internet of Things Era.

Precision agriculture plays a crucial role in optimizing crop yield, reducing environmental impact, and ensuring sustainable agricultural practices. Early detection and accurate diagnosis of leaf diseases are essential for preventing significant losses in crop production and maintaining food security. However, the inherent challenge of class imbalance in leaf disease datasets poses a significant obstacle for machine learning algorithms. In this paper, we explore and compare different techniques for handling class imbalance in leaf disease detection to improve the accuracy and reliability of machine learning models in the context of precision agriculture. We investigated the performance of different methods for leaf disease detection using the challenging New Plant Diseases Dataset (NPDD), which consists of image-based plant leaves. Our experiments reveal promising results, particularly with the hierarchical approach, achieving an accuracy of 97.17%. The outcomes of our study contribute to the growing body of knowledge in precision agriculture by providing a comprehensive analysis of techniques for handling class imbalance in leaf disease detection. Furthermore, our findings serve as a valuable resource for researchers and practitioners in the field, offering guidance on selecting and implementing the most effective approaches to tackle class imbalance challenges and improving the overall performance and reliability of machine learning models in the domain of precision agriculture.

A dialogue state tracker is a component in a task-oriented dialogue system that monitors the current state of a conversation and gives information about its context and history to other system components. The dynamic and open-ended character of human interactions is one of the primary obstacles in dialogue state monitoring, necessitating robust and adaptable models to keep up with the quick context changes. Recently, numerous deep learning-based algorithms have been developed for this purpose. Still, these models are typically heavily-engineered and conceptually sophisticated, making them challenging to deploy, debug, and maintain in a production environment. To overcome these challenges, we offer the BERT-SIAM-DST model, a unique way to dialogue state monitoring employing a Siamese network with BERT as the base network. This model uses the robust representation capabilities of BERT and the ability of Siamese networks to record correlations between inputs to make accurate predictions regarding the current state of the discussion. In addition, the number of parameters does not increase proportionally with the size of the ontology, and the model is adaptable to alterations in the domain ontology. We test the performance of the BERT-SIAM-DST model on the standard WoZ 2.0 dataset of annotated dialogues and compare it to other approaches. Compared to numerous baseline models, the BERT-SIAM-DST model is effective at tracking the state of discussions, demonstrating the promise of BERT-based Siamese networks for this purpose.

This paper discusses an Internet of Things (IoT)-based energy meter for photovoltaic systems (PV) to forecast energy production at industrial locations. Based on an ESP-32 card, the proposed IoT device collects energy consumption data from the sub-meter and delivers it to the cloud. The data is used to monitor the values of a typical PV system installed in Benguerir (Morocco). The method involves conducting an analysis of the solar resource available at the site in Benguerir, as well as conducting an investigation, evaluation, and selection of the components of the solar station using simulation software such as the PVSYST tool. The proposed solution will improve the management of a PV system. Additionally, the method involves the development of a datalogger that is used for monitoring solar panels’ energy production, storing data in the cloud, and displaying results on a web interface. Lastly, we apply an artificial neural network for solar energy forecasting to future production.

As a rapidly evolving paradigm, digital transformation (DT) remains one of today's most significant challenges. Consequently, numerous businesses devote a substantial portion of their resources to this captivating transition. However, this investment's return is difficult to quantify or even unexplored by others. This observation constituted the foundation for Solow's productivity paradox. Even if they are aware of how the KPI measurement affects their performance, managers continue to disregard this evaluation. In terms of academic literature, this field of study is still inadequately developed. To address this limitation, this study uses the NVivo software to conduct a systematic literature review highlighting the KPIs, the methodology used in the corpus, the evolution of related works, and their type. According to our analysis, several works cite this term without describing it or defining a general approach to DT metrics. In addition, the financial and commercial performance KPIs are the most specific. Following a summary of the primary KPIs in the literature, we classified them into two prominent families: generic and specific. This classification represents the originality of this paper. The generic KPIs are multidimensional and applicable to a variety of businesses, while the specific ones pertain to a particular function or industry.

Link prediction is a challenging issue in practical applications such as recommendation systems. The purpose of such applications is to predict the presence of links between distinct objects based on examining structured data within a network. In this paper, we develop a graph convolutional neural network (GCNN) model to address this problem while incorporating interaction relationships and content information on various elements. The suggested GGCN improves prediction accuracy by constraining the consistencies of the graph embedding from multiple perspectives, in contrast to existing strategies that directly combine hybrid approaches based on interaction and content information into a single display. Experimental results were tested on three datasets, including Facebook, Google+, and Twitter, using various hyperparameters.

This research paper presents a deep learning-based predictive model for classifying currency trends using technical indicators. The model is trained on a dataset generated from three technical indicators: relative strength index (RSI), moving average convergence divergence (MACD), and stochastic. The dataset consists of historical currency data along with the corresponding values of the technical indicators. The deep learning model can accurately classify the trends of a given currency based on the importance of these indicators. The model's performance is evaluated using standard metrics, and the results demonstrate its effectiveness in classifying currency trends. The proposed model provides a valuable tool for traders and investors in the foreign exchange market by helping them make informed decisions about the direction of currency prices.

The retina is affected by glaucoma and diabetic retinopathy (DR). Glaucoma must be detected early because it is irreversible and one of the leading causes of blindness. A delayed diagnosis will result in permanent vision loss. It is characterized primarily by ganglion cell dysfunction, which changes the thickness of the retinal nerve fiber layer and the shape of the optic nerve head. As a result, early detection of glaucoma is critical for preventing vision loss. This study employs a hybrid approach to glaucoma diagnosis by combining its powerful Multichannel Gabor filtering and Principal Component Analysis (PCA) capabilities with various transfer learning architectures such as MobileNet, MobileNetV2, and NASNet. These classifiers divide the source retinal images into two groups: glaucoma and non-glaucoma. The suggested approaches are used and evaluated on a dataset of Retinal Fundus Images. For the glaucoma diagnostic method, this strategy yields 99% Precision, 97% Recall, and 98% Accuracy.

Cognitive radio (CR) is considered a primary technology for spectrum usage efficiency and dynamic spectrum management. Spectrum sensing is an essential cognitive radio cycle function. Recently, reconfigurable intelligent surfaces (RIS) technology has emerged as a promising new enabler of 6G wireless communication, with the ability to control signal propagation and increase signal coverage and spectrum management. This paper investigates how integrating RISs with CR can improve spectral efficiency. First, we optimize the RIS parameters by identifying the optimal transmit powers and number of RIS antennas. Then, we compare the novel RIS technology with the traditional decode-and-forward (DF) relaying technology. Next, we investigate how adopting RIS can enhance spectral efficiency. Finally, we demonstrate how integrating the novel technology of reconfigurable intelligent surfaces with CR can significantly increase spectrum detection, the most crucial step in the cognitive radio process. Simulation studies indicate that RIS is more energy-efficient than SISO and DF-relay communications in various setups. Moreover, by boosting spectrum sensing capability in CR situations, RISs can significantly enhance the efficiency of frequency spectrum usage.

Recent advancement in biomedical imaging technologies has contributed to tremendous opportunities for the health care sector and the biomedical community. However, collecting, measuring, and analyzing large volumes of health-related data like images is a laborious and time-consuming job for medical experts. Thus, in this regard, artificial intelligence applications (including machine and deep learning systems) help in the early diagnosis of various contagious/ cancerous diseases such as lung cancer. As lung or pulmonary cancer may have no apparent or clear initial symptoms, it is essential to develop and promote a Computer Aided Detection (CAD) system that can support medical experts in classifying and detecting lung nodules at early stages. Therefore, in this article, we analyze the problem of lung cancer diagnosis by classification and detecting pulmonary nodules, i.e., benign and malignant, in CT images. To achieve this objective, an automated deep learning based system is introduced for classifying and detecting lung nodules. In addition, we use novel state-of-the-art detection architectures, including, Faster-RCNN, YOLOv3, and SSD, for detection purposes. All deep learning models are evaluated using a publicly available benchmark LIDC-IDRI data set. The experimental outcomes reveal that the False Positive Rate (FPR) is reduced, and the accuracy is enhanced.

Integration of advanced signal processing, image processing, deep learning, edge computing, and the Internet of Things (IoT) into vehicles allows intelligent automated vehicles to navigate autonomously in different environments. It is crucial for reliable and safe driving that an autonomous vehicle can accurately, effectively, and efficiently recognize, perceive, and observe the surrounding environments. Autonomous vehicles comprise advanced sensor technologies such as RGB cameras and LiDaR that produce an extensive data set in the form of RGB images and 3D measurement points, also recognized as a point cloud. It is necessary to understand and interpret collected data information efficiently and to identify other road users, such as pedestrians and vehicles. Thus, we introduced a smart IoT-enabled deep learning based end-to-end 3D object detection system that works in real-time, emphasizing autonomous driving situations. The detection model is based on YOLOv3; firstly, the model is utilized for 2D object detection and then modified for 3D object detection purposes. The presented model uses point cloud, and RGB image data as input and outputs detected bounding boxes with confidence scores and class labels. Experiments are carried out on the Lyft data set; results reveal that the YOLOv3 model achieves high accuracy and outperforms from other state-of-the-art detection models in terms of effectiveness and accuracy. The overall accuracy of the model is 96% and 97% for 2D and 3D object detection, respectively.

With the rapid evolution of security technology, small field-size elliptic curve-based point multiplication (PM) has gradually become obsolete, leading to the implementation of PM with large field sizes. From this perspective, in this article, through a novel algorithm-architecture co-design strategy, we propose an efficient implementation of the PM on the elliptic curve over GF($2^{m}$) (particularly targeting large field sizes). To achieve an area-time-efficient elliptic curve cryptography (ECC) processor implementation on the field-programmable gate array (FPGA) platform, we have proposed a bottom-up approach based on three coherent interdependent layers of efforts. First, we proposed an efficient digit-serial versatile multiplier (DSVM) based on polynomial representation. The system is built using the four-way overlap-free Karatsuba algorithm (OFKA) and a modified radix-n interleaved multiplication (mRnIM) technique (for area and time complexities reduction). Of course, the efficiency of the proposed multiplier is demonstrated by the complexity analysis and comparison with the existing reported designs. Second, we have adopted the López–Dahab (LD) Montgomery PM algorithm to avoid data dependency and enhance signal control in the ECC design. Meanwhile, a series of resource optimization techniques have also been adopted for the proposed ECC processor to optimize the overall design efficiency further. Third, the proposed ECC PM architecture is then implemented on the FPGA platform, showing that the proposed ECC crypto-processor obtains the least area-delay product (ADP) among all the existing structures for the large field sizes.

The next generation of the Internet of Things will rely heavily on RFID and sensor technology. Compared to existing systems, the development of various RFID and sensor technologies is expected to increase IoT applications, especially in the field of healthcare. This research summarizes the fundamentals of the Internet of Things (IoT) based RFID technology. RFID sensors are critically compared by separating them into near-field and far-field configurations. The two categories are then analyzed for their operating mechanism. RFID sensing via tags equipped with chips is now a sophisticated technological device that continuously increases its presence in the market and several application scenarios; therefore, adopting RFID tag sensors for healthcare applications is discussed in this work. The advantages and limitations of several IoT-based RFID configurations for healthcare are reviewed. In addition, a summary of the most appropriate application scenarios of the RFID sensors is then illustrated. Finally, a look at overall machine learning (ML) satisfactory solutions for RFID antenna design is highlighted.

Conversational Systems are increasingly substituting humans in many service industries. They aim to provide human-like interaction with users for task completion or chitchat in a conversation style. User sentiment analysis is an important task that can help better understand users’ behavior and satisfaction in conversations. Although some researchers have studied the problem of sentiment analysis, most of the existing methods are oriented toward general felds. To overcome the challenges of sentiment analysis, we propose a BE-Att-BiLSTM, which stands for an attention-based Bidirectional Long Short-Term Memory (BiLSTM) model. The proposed model uses pre-trained BERT, contextual embeddings and a combination of BiLSTM and attention mechanism for efficient sentiment analysis in conversations. In addition, text-augmentation techniques are leveraged to enhance the performance of the proposed model. Experimental results on a public benchmark dataset show an improved accuracy of 68.00% and an F1-score of 67.50%.

The need to preserve the environment has become a real societal and economic challenge. There are many sources of pollution, and identifying these sources, the different polluting substances, and their effects on ecosystems is complex. They can arise from natural disasters or as a result of human activity. The need to preserve the environment has become a real societal and economic challenge. The preservation of the environment can also be done by better management of human activities. However, these solutions require several human and technological resources, and they are expensive. However, such applications are usually deployed in large areas, not accessible and without pre-existing infrastructure, or that would require substantial redevelopment costs) In addition, they have to work for months (or even years) without human intervention. Internet of Things (IoT) is widely used to implement such applications. In this context, we introduce not only the notion of IoT but also Edge/Cloud Computing for environmental monitoring. The proposed solution represents a solution that is both financially and energy-efficient. This paper also focuses on accessing issues to IPv6 over Low Power Wireless Personal Area Networks (6LoWPAN) sensors using Constrained Application Protocol (CoAP).

The smart city has become a persistent need and is no longer just a concept. The concept of smart cities heavily relies on collecting enormous amounts of data. This paper proposes a data-management-based solution for smart city, which is labeled Smart Systems Oriented Big Data Architecture. Big data technologies have become essential to the functioning of cities. The architecture includes complex components to be implemented based on the architectural requirements. A data migration strategy was proposed to handle the various data sources such as IoT devices, video cameras, and drones. The proposed approach also takes into account data processing and data storage. The technical constraints related to data processing in a big-data environment are also studied. We also consider data modeling from a business intelligence point of view and a data science perspective. Our main goal is to favor the facilitation of the daily life practices in the context of a smart city by providing the city administrators with a solution that helps them maintain their city smartly and effectively.

Spectrum sensing is the first step in the cognitive cycle and represents the most critical function in cognitive radio-based dynamic spectrum management. Recently, a new technology termed reconfigurable smart surfaces has emerged as a promising enabler of smart radio environments to control the signal propagation further and improve signal coverage and spectrum management capabilities. This paper investigates how the adoption of reconfigurable intelligent surfaces (RISs) can increase spectral efficiency in a cognitive radio environment. To this end, we optimize the parameters of the new RIS technology by determining the optimal transmit powers and the optimal number of these elements. Next, we derive expressions for the false alarm and detection probabilities of the cognitive radio (CR) node and the transmission probability and throughput. Finally, we demonstrate how the detection phase of the CR spectrum can be improved by employing RIS technology at the UP. Simulation results show that RIS achieves higher energy efficiency than SISO communication for different configurations. In addition, the RISs can significantly improve the wireless communication quality and spectrum sensing performance in a CR environment.

This paper presents a deep artificial neural network approach based currency market volatility based recommendation engine. Since deep learning classification needs labeled data set that we don't have, an approach is designed specially for that point in order to generate labeled data set from non labeled one. This is a major innovative aspect in this contribution in addition to the recommendation service. It is based on Gaussian kernel density and Monte Carlo simulation. The main goal of the proposed approach is to predict - for each hour of the day - the volatility behaviour of the selected currency pair. The proposed model has a range of applications in financial market specially the algorithmic trading. Deep neural network was trained and evaluated and testing process gave good convergence rate.

Nowadays, the Internet of things (IoT) plays a vital role in various sectors, including smart cities, health care, industries, and agriculture. With the advancement of IoT and machine vision-based technologies, such a manual system can be replaced with an automated weed control system. The main objective of this research work is to develop an automatic IoT-based system that detects and identifies weeds with much high accuracy and with a low computational time. Furthermore, we proposed two simple, fast, and effective methods to discriminate broad, narrow, and little weeds. The primary building blocks of the proposed system consist of a pre-processing system followed by Circular Mean Intensities and Discrete Fourier Transform, which are applied to extract useful features from images. A threshold value is then set based on these features to distinguish different classes of weeds. The experimental results produced by the proposed algorithms achieved a classification rate of 96% using 350 images.

With the rapid development of artificial intelligence (AI), many companies are moving towards automating their services using automated conversational agents. Dialogue-based conversational recommender agents, in particular, have gained much attention recently. The successful development of such systems in the case of natural language input is conditioned by the ability to understand the users’ utterances. Predicting the users’ intents allows the system to adjust its dialogue strategy and gradually upgrade its preference profile. Nevertheless, little work has investigated this problem so far. This paper proposes an LSTM-based Neural Network model and compares its performance to seven baseline Machine Learning (ML) classifiers. Experiments on a new publicly available dataset revealed The superiority of the LSTM model with 95% Accuracy and 94% F1-score on the full dataset despite the relatively small dataset size (9300 messages and 17 intents) and label imbalance.

The Internet of Things or IoT describes the network of physical terminals, the “objects” that integrate sensors, software, and other technologies to connect to different terminals and systems on the Internet and exchange data with them. These terminals can be simple household appliances and industrial tools of great complexity. Augmented Reality (AR) provides operators with information directly in their field of vision, using a helmet or adapted glasses. The obstacles to the adoption of Augmented Reality are related to several factors. First, the devices do not yet offer efficient and discreet ergonomics and are still expensive. While IoT and sensors are focused on the productivity of machines, Augmented Reality makes it possible to increase the performance of a human. Industrial Augmented Reality aims to improve the factory’s economic performance. This article proposes a hybrid solution between mixed Reality and IIoT for industrial applications.

Machine learning (ML) and deep learning (DL) techniques have been used to study the changes in human physiological and non-physiological properties. DL has proven his efficiency when perceiving positive emotions (joy, surprise, pride, emotion) and negative emotions (anger, sadness, fear, disgust). Furthermore, the DL is used to identify the emotions accordingly. First, this paper describes the different DL and ML algorithms applied in the emotion recognition field. Then, as a perspective, it proposes a three-layered emotion recognition architecture that leverages the massive data generated by IoT devices such as mobile phones, smart homes, and health monitoring. Finally, the potential of emerging technologies, such as 5G and 6G communication systems in a parallel Big Data infrastructure, were discussed.

Nowadays, there is an important need for fault-tolerant and energy-efficient self-organization systems, especially within smart cities. Internet of Things (IoT) proved capable of observing and examining the environment, generating & processing data. IoT is now applicable to almost every industry, including transportation and logistics, utilities, agriculture, smart cities, and more. In these industries, various types of meters, sensors, and trackers are used to constantly monitor activities, automate processes and optimize tasks. With the help of big data analytics, they can drive decision-making systems based on observations. As a result, the cities-management challenges are growing. The smart cities requirements are increasing to remedy the challenges, which requires a self-organized network composed of a sizeable number of nodes distributed across an area of interest. The traditional communication systems show limitations, especially when dealing with massive data rates, latency, the explosive growth of vehicular communication, and dynamic mobility. In this study, we explore a way to leverage the capabilities of wireless communication and big data analytics in favor of Smart Cities.

Smart city applications are using extensively artificial intelligence for decision-making. Among the fields of application are facial recognition and intrusion detection. The subject is old, but processing techniques and hardware are constantly evolving. This paper will review the most widely known practices and apply them to a smart parking and intrusion detection system using the “JetsonNano” board. Nowadays, quality assurance for machine learning systems is becoming increasingly important. This article focuses on detecting bugs in implementing two classical face recognition algorithms: Eigenface (EF) and Local binary pattern histogram (LBPH). We tested the efficiency of our system using metamorphic testing depending on many factors: weather conditions, pixel noise, and distortion.

This paper reports the characterization procedure of a radiofrequency identification tag powered by meandered L-matching configuration and placed directly on the planar layered anatomical model of a human arm. The tag antenna component and its matching system interface to the RFID chip are designed with the help of electromagnetic simulators. A new optimal tag structure is combining multiconductor strips and a meandering schema used to achieve the required inductance. The folded configuration of the proposed tag adds stretchability and more reduction of the antenna size, especially when attached to the non-uniform as the human body. It is demonstrated that the tag can communicate with a reader. The simulated performances indicate the robustness of the proposed tag structure and its ability to be deployed in several healthcare sensing applications.

The Industrial Internet of Things (IIoT) is expected to enable Industry 4.0 through the extensive deployment of low-power devices. However, industrial applications require, most of the time, high reliability close to 100% and low end-to-end delays. This corresponds to very challenging objectives in wireless (lossy) environments. This delay can be disastrous in time-sensitive industrial IoT deployments where immediate detection and actions impact security, safety, and machine failures. With an efficient MAC protocol, data will be provided quickly to enable the IoT to be fully effective for mission-critical applications. Efficient medium sharing is even more difficult in IIoT due to ultra-low latency, high reliability, and high quality of service (QoS) compared to best-effort for IoT. This article does not survey all existing MAC protocols for IoTs, which was already done in other works. The goal of this paper is to analyze existing MAC protocols that are more suitable for IIoT.

Since the arrival of the internet and affordable access to technologies, digital technologies have occupied a growing place in industries, propelling us towards a 4th industrial revolution: Industry 4.0. In today’s era of digital upheaval, enterprises are increasingly undergoing transformations that are leading to their digitalization. The traditional manufacturing industry is in the throes of a digital transformation that is accelerated by exponentially growing technologies (e.g., intelligent robots, Internet of Things, sensors, 3D printing). Around the world, enterprises are in a frantic race to implement solutions based on IoT to improve their productivity, innovation, and reduce costs and improve their markets on the international scene. Considering the immense transformative potential that IoTs and big data have to bring to the industrial sector, the adoption of IoT in all industrial systems is a challenge to remain competitive and thus transform the industry into a smart factory. This paper presents the description of the innovation and digitalization process, following the Industry 4.0 paradigm to implement a successful enterprise IoT strategy.

Reconfigurable Intelligent Surfaces (RISs) have been recently considered in communication environments for focusing signal reflections to create smart radio environments. As it can create favorable propagation conditions by controlling the phase shifts of the reflected signals at the surface to enhance the communication quality. However, the current proofs of concept require complex operations for the RIS configuration, which are mainly realized via wired control connections. When used in wireless networks, RISs may contribute enhance wireless communications. The present paper, compares the RIS technology with the SISO case and evaluates the performance of RIS assisted communication system by giving an approximation of the achievable data rate and the energy efficiency, and investigates the effect of the achievable data rate and distances on the energy efficiency. In this work, based on a work by Emil Björnson [18], we have studied the impact of RIS on the wireless sensor networks by modifying some parameters and giving more interpretations. In particular, we illustrate numerical results that highlight the spectral efficiency gains of RISs when their size is sufficiently large as compared with the wavelength of the radio waves. In addition, we discuss key open issues that need to be addressed for unlocking the potential benefits of RISs for application to wireless communications and networks.

People hope that computers can be in constant intelligence development. Just like humans, they can ”see” the world and ”recognize” a visual event. We propose an approach based on computer vision methods to recognize Human-Object interaction(HOI). The technique stands on aggregating significant contextual features Human-Object interactions and scene recognition. We design a branch architecture consisting of the main branch for HOI detection and a supplementary branch for scene recognition. We explore the deep learning models through the knowledge distillation method and the Cross Branch Integration mechanism for encoding models into graph neural network architecture. We construct a knowledge graph to merge between high-level context information. When trained collaboratively, those models allow computing efficiency, strong context knowledge.

Several video coding standards and techniques have been introduced for multimedia applications, particularly the h.26x series for video processing. These standards employ motion estimation processing to reduce the amount of data that is required to store or transmit the video. The motion estimation process is an inextricable part of the video coding as it removes the temporal redundancy between successive frames of video sequences. This paper is about these motion estimation algorithms, their search procedures, complexity, advantages, and limitations. A survey of motion estimation algorithms including full search, many fast, and fast full search block-based algorithms has been presented. An evaluation of up-to-date motion estimation algorithms, based on several empirical results on several test video sequences, is presented as well.

The main goal ofNew Radio 5G (NR) mobile technology is to support three generic service categories, each with very specific requirements. The first category is enhanced Mobile Broadband (eMBB), the second category relates to massive Machine-Type Communications (mMTC), and the third category relates to ultra-Reliable Low Latency Communications (URLLC). The slicing of the radio part of 5G network access network has greatly contributed to the emergence of these three categories of service with different qualities of service. This division therefore enabled the network to reserve the necessary resources for each category of services, orthogonally, and according to the performance required. In this article, we have dealt with the problem of Call Admission Control (CAC) in 5G networks where we have considered the case of the only two categories eMBB and uRLLC, which their users are served by a single cell. We calculated the maximum eMBB users admitted into the system with guaranteed data rate, while allocating power, bandwidth, and beamforming directions to all uRLLC users whose latency requirements and reliability are always guaranteed. We only considered the downlink communication, and we used the case of the multiple-input single-output (MISO) system. This CAC problem is formulated as a minimization problem l0 which is known as NP-hard problem. We therefore chose to use Sequential Convex Programming (SCP) to find a suboptimal solution to the problem.

Summary With the rapid growth in wireless data networks and increasing demand for multimedia applications, the next generation of wireless networks should be able to provide services for heterogeneous traffic with diverse quality of service (QoS) requirements. Multiuser diversity refers to a type of diversity present across different users in a fading environment. This diversity can be exploited by scheduling transmissions so that users transmit when their channel conditions are favorable. Hence, scheduling algorithms that support QoS and maintain a required throughput to ensure users' satisfaction are crucial to the development of these wireless networks. In this paper, different scheduling techniques have been evaluated using OFDM with different scenarios. The goal is to analyze the properties of networks such as throughput, fairness, and delay. Experimental results indicate that the PFS approach outperforms the other techniques in terms of fairness, throughput, and delay.

Summary Optimizing cell locations of cellular networks is one of the most fundamental problems of network design. However, in order to meet a growing appetite for mobile data services, a large number of base stations are being deployed, which leads to tremendous energy consumption in cellular networks. This augmentation increases not only the system's capital and operational expenditure (CAPEX/OPEX) for mobile operators but also CO2 emissions. Besides the issue of meeting overwhelming traffic demands, network operators around the world now realize the importance of managing their cellular networks in an energy‐efficient manner. In this paper, we develop a self‐organizing framework for energy saving in orthogonal frequency‐division multiple‐access–based cellular access networks. We consider three different objectives, namely, coverage maximization, overlap minimization, and power consumption minimization, which is different from all existing works on energy saving in cellular networks.

Computational intelligence techniques are traditionally adopted in several different application domains such as industry, healthcare, decision making, and gaming, to name but a few. Despite this growing diffusion, there are still many possible areas where computational intelligence application is partial or could be extended and improved due to the actual limitations in terms of computational power or strict requirements in terms of assurance of the results. \n \nThis special issue aims to investigate the impact of the adoption of advanced and innovative computational intelligence techniques in emerging application fields like IoT and big data. This edition of the special issue is focused primarily on industrial and health applications with special emphasis on real-time systems grounded on big data ecosystems. \n \nThe special issue will bring together researchers on different disciplines from academia and industry with a common objective: go beyond the frontiers of today's applications of computational intelligence techniques.

Real-time social multimedia level threat monitoring is becoming harder, due to higher and rapidly increasing data induction. Data induction through electric smart devices is greater compared to information processing capacity. Nowadays, data becomes humongous even coming from the single source. Therefore, when data emanates from all heterogeneous sources distributed over the globe makes data magnitude harder to process up to a needed scale. Big data and Deep learning have become standard in providing well-known solutions built-up using algorithms and techniques in resolving data matching issues. Now, with the involvement of sensors and automation in generating data obscures everything, predicting results to overcome a current era of ever enhancing demands and getting real-time visualization brings the need of feature like human behavior mode extraction to overcome any future threats. Big data analytics can bring the opportunity of predicting any misfortune even before they happen. Map reduce feature of big data supports massive data oriented process execution using distributed processing. Real-time human feature identification and detection can occur through sensors and internet sources. A behavioral prediction can further classify the information collected for introducing enhanced security extents. Real-time sensor devices are producing 24/7-hour data for further processing recording each event. IoT-based sensors can support in behavioral analysis model of a human. Real-time human behavioral monitoring based on image processing and IoT using big data analytics.

Abstract Embedded systems and wireless sensor networks (WSN) are found today in increasingly critical areas of applications. They have become integrated and embedded in nearly all aspects of everyday life, including manufacturing, healthcare, education, critical infrastructure, and entertainment. The number of connected devices continues to grow, and due to the insecure nature of these devices, the amount of risk continues to grow as well. These risks, however, can be mitigated with the creation and adoption of WSN security standards developed to create an environment of safety, security, and confidence in the technology. Designing the security policy for WSNs requires asking some preliminary questions. These questions are particularly important in the case of WSNs because their use is highly decentralized. Blockchain's ability on governing decentralized networks makes it especially suitable for designing a self‐managing system on WSN devices. This article proposes a routing protocol that uses Blockchain technology to offer a shared memory between the network's nodes. The simulation results have shown that this solution could be applicable and could resolve the issues cited above.

Summary The International Telecommunications Union defines in its IMT‐2020 recommendations three types of use of 5G services: mMTC (massive Machine‐type Communications), eMBB (enhanced Mobile Broadband), and uRLLC (ultra‐Reliable Low Latency Communications). The mMTC service allows a considerable number of machines and devices to communicate while guaranteeing a good quality of service. The eMBB service allows very high data throughput, even at the cell border. The uRLLC service is used for ultra‐reliable communication for critical needs requiring very low latency. These services are provided separately in a given cell. However, the number of connected objects is starting to increase rapidly as well as the bit rates and energy consumption. The 5G network must make it possible to provide access to a vast number of users of its different service categories. Call admission control (CAC) techniques focus more on availability in terms of bit rate and coverage. In this article, we suggest an algorithm for modeling CAC in an area served by the three categories of services in a 5G access network, mainly based on minimum energy consumption. This technique will allow connected objects that consume low energy to connect to the network with an adequate quality of service and enable the development of the Internet of Things.

The new concept of Industry 4.0 has been developed: it includes both Internet of Things (IoT) structure and the local networks that are still needed to carry out real-time tasks. Genetic algorithms are successfully used for decoding some classes of error correcting codes, and offer very good performances when solving large optimization problems. This article proposes a decoder based on parallel Genetic Algorithms (PGAD) for Decoding Low Density Parity Check (LDPC) codes. The proposed algorithm gives large gains over the Sum-Product decoder, which proves its efficiency, the best performances are obtained for Ring Crossover (RC) as a type of crossover and the tournament as a type of selection. Furthermore, the performances of the new decoder are improved using Multi-criteria method. For the LDPC code, simulation results showed that our Proposed PGAD exceeds the sum-product by a gain of 1.5 dB at BER = 10-4, and the PGAWS exceeds the sum-product by 2.5 dB.

Automatic fault detection using machine learning has become an exciting and promising area of research. This because it accurate and timely way to manage and classify with minimal human effort. In the computer vision community, deep-learning methods have become the most suitable approaches for this task. Anodes are large carbon blocks that are used to conduct electricity during the aluminum reduction process. The most basic function of anode rod inspection is to prevent a situation where the anode rod will not fit into the stub-holes of a new anode. It would be the case for a rod containing either severe toe-in, missing stubs, or a retained thimble on one or more stubs. In this work, to improve the accuracy of shape defect inspection for an anode rod, we use the Fast Region-based Convolutional Network method (Fast R-CNN), model. To train the detection model, we collect an image dataset composed of multi-class of anode rod defects with annotated labels. Our model is trained using a small number of samples, an essential requirement in the industry where the number of available defective samples is limited. It can simultaneously detect multi-class of defects of the anode rod in nearly real-time.

The massive growth of wireless communications in recent years is mostly due to new connectivity demands and advances in technology development of low power) transceivers. An example of the unique demands is the increasing exchange of data in Internet services, which has led to wireless network deployment for data transmissions. The coordination of the IoT devices, smart systems, and agriculture can contribute directly to the development of the farmer’s practices by building their farm more intelligent and digital. However, enhancing farming practices requires inspecting farm equipment and farmer’s experiences, which can be analyzed through the interconnectedness of IoT objects to collect farm data over the Internet to launch smart digital agriculture. It is challenging to control all farming processes (especially in real-time), this remaining as the main limitation of traditional farming. In this work, we focus on how wireless sensors can play a vital role in smart farm systems and allow processing the large amount of data generated in batches or real-time to analyze it, retrieve insights from it, and create a Smart Digital Farm. This paper proposes hierarchical-logic mapping and deployment algorithms to tackle the problem of poor network connectivity and sensing coverage in random IoT deployment.

Recent advances in sensing modules and radio technology will enable small but smart sensors to be deployed for a wide range of environmental monitoring applications. They collect data from different environment or infrastructures in order to send them to the cloud using different communications platforms. These data can be used to provide smarter services. However, they are various issues and challenges related to the ubiquitous sensors that should be solved. In this paper we interest on analysis of wireless sensor network from an energy management perspective. The idea behind the energy-efficiency wireless sensor networks is that each node can only transmit to a limited number of other nodes directly. The limited resources of nodes imply that the transmission range is limited. In order to transfer the data to the final destination, the traffic must be relayed using intermediate nodes, creating a multi-hop route. The total energy consumption associated with an end-to-end transmission over such a route can be significantly reduced if the nodes are correctly configured. In this paper, underground mine monitoring system is presented with an overview of the related issues and challenges such as reliability, cost, and scalability.

The dramatically development of wireless technologies in the last few decades, leads to the growth of channel resources demand in a limited spectrum with inextensible character. Cognitive radio network (CR) is a promising technology that provides solutions for the spectrum management and optimization problems via dynamic spectrum management. The spectrum resources management and optimization are an important part of the future network performances. In this paper, we propose an efficient algorithm to examine the design specification issues regarding the choice of optimal power, optimal speed, and optimal amount of information in a wireless network along with studying the effect of different parameters on the obtained results. Our objectives are to guarantee the protection on licensed users (Primary users ‘PU’) from harmful interference caused by the unlicensed users (Secondary users ‘SU’), more especially, to optimize the quality of communication link, Transmission levels, and battery life of the wireless devices. Results show that our proposed work leads to an efficient utilization of radio spectrum and strongly contributes to alleviating the spectrum scarcity problem.

Spectrum sensing is an essential step in cognitive radio-based dynamic spectrum management. Spectrum sensing to detect the presence of the licensed signals in a particular frequency band is one of the most important research topics in cognitive radio. To identify primary user (PU) presence, we propose a low cost and low power consumption implementation of spectrum sensing operation based on real signals. These signals are generated by smart embedded devices at 433 MHz wireless transmitter using ASK (Amplitude-Shift Keying) and FSK (Frequency-Shift Keying) modulation type. The reception interface is constructed using an RTL-SDR dongle connected to MATLAB software. The signal detection is done by using four techniques: the artificial neural network (ANN), support vector machine (SVM), Decision Trees (TREE), and k-nearest neighbors (KNN). This article comparatively analyzed the performance of the classifiers to identify the best method for spectrum sensing between the three techniques. The performance evaluation of our proposed model is the probability of detection (Pd) and the false alarm probability (Pfa). Results show also that the sensing is susceptible to signal to noise ratio value. This comparative study has been demonstrated that the spectrum sensing operation by ANN and SVM can be more accurate than KNN, TREE, and some other classical detectors.

New generations of mobile telephony succeed every decade, each bringing an evolution or even a revolution. Nowadays, the Internet of Things and the tactile Internet are starting to grow, and 5G technology is there to enable these services. 5G technology has introduced three types of services, namely eMBB (for services requiring very high bit rates), mMTC (for massive connection of user equipment), and uRLLC (for critical services requiring very high reliability and extremely reduced latency). In this paper, we have dealt with some issues encountered by uRLLC services for tactile Internet services. In this article, we have studied the transmission of very small packets as required by the 5G uRLLC services. We also examined the probability of transmission error and its variation concerning the transmission delay and the length of the packet transmitted. This study was conducted considering its application in the Tactile Internet.

As power line communication (PLC) technology does not require dedicated network setup, it can be used to connect a multitude of sensors and IoT (Internet-of-Things) devices. Those IoT devices could be deployed in homes, streets, or industrial environments for sensing and control related applications. The key challenge faced by future IoT-oriented narrowband PLC networks is to provide a high quality of service (QoS). In fact, the power line channel has been traditionally considered too hostile. Combined with the fact that spectrum is a scarce resource and interference from other users, this requirement calls for means to increase spectral efficiency radically and to improve link reliability. However, the research activities carried out in the last decade have shown that it is a suitable technology for a large number of applications. Game theory is a set of analytical tools to model the complex interactions between rational agents. Motivated by the relevant impact of PLC on IoT, this paper investigates cooperative spectrum allocation based on game theory.

The mining industry has been acting as a leader in the development of fully autonomous vehicles. Mining equipment manufacturers have been developing and testing autonomous vehicle technologies for many years. There are many ongoing innovations in autonomous vehicle technology. One of these is the Dedicated Short Range Communications (DSRC). The DSRC is a one-way or two-way short-to-medium-range wireless communications capability. In this paper, we present a review of the DSRC large-scale fading channel at 5.9 GHz in the tunnels and underground mines. The requirements for DSRC receiver performance for VANET- Vehicular Ad-hoc Networks applications in an underground mine is calculated. This paper also reports the overall performance evaluation of three existing routing protocols, namely, Emergency Message Dissemination for Vehicular Environments (EMDV), Enhanced Multi-Hop Vehicular Broadcast (MHVB), and Efficient Directional Broadcast (EDB) for active safety applications. Finally, a comparative study of these three routing protocols for cooperative collision warning in underground mining galleries was evaluated.

The demand for wireless services is becoming much more significant than the currently available spectrum could accommodate. Therefore, improvements at the physical layer alone cannot sustain such high data rates. Extreme densification of wireless networks (e.g., small cells) and the use of additional bandwidth are necessary to meet this increasing demand. On the other hand, multiple-input-multiple-output (MIMO) technology was also introduced to overcome the saturated data rate capabilities of conventional single antenna systems given the limited bandwidth and power levels. However, among various performance aspects and design problems regarding MIMO, precoding is one of the most crucial function components to ensure reliable communication. Without precoding, inter-stream interference can be severe even with cross-pol antenna design. However, the precoding could help to reduce the inter-stream interference and increase the signal strength along the desired signal subspace. In this work, a low complexity design a PHY-MAC precoder for sub-6 GHz backhaul small cell is proposed. The simulations were performed using an SUI-3 channel model for fixed wireless applications.

In an asynchronous environment, optical code division multiple access (OCDMA) is an advanced technique. However, this technique shows limitations in terms of the low spectral density and inefficient bandwidth utilization when implemented with one-dimensional (1D) codes. Thus, this study presented a novel two dimensional (2D) spectral/spatial multi-wavelength code to overcome these limitations. The proposed code is formulated using a 1D balanced incomplete block design (BIBD) technique. It is designed and implemented for spectral amplitude coding (SAC) based OCDMA networks and constructed using a 1D BIBD code matrix. Optisystem software-based simulation results indicate that the proposed code provided improvement in the number of simultaneous users, code construction, cross-correlation, and minimize the noises. Due to its practical code design approach, the proposed code family yields large cardinality with optimal code length. Moreover, system performance illustrates that the system with the proposed code maintains required optical transmission property by supporting six clients for source power -10 dBm with the data transmission rate of 1 Gbps.

Summary Recent technology developments have produced small and smart biomedical sensors, which can be worn or implanted in the human body. These biosensors create closed wireless networks named Wireless Body Area Networks (WBAN). The WBAN will continuously observe the physiological state of patients for both diagnosis and prevention. Those include on‐body measurements such as the Electrocardiogram (ECG), Electroencephalogram (EEG), temperature, and blood pressure. Ultra‐Wide‐Band (UWB) is a technology that has received a lot of attention due to several unique features such as secure transmission, low noise, and low energy consumption. Given the fact that the patients' well‐being might be dependent on the accurate realization of such networks, a high level of design and implementation accuracy are maintained throughout the system. In this paper, we proposed an Impulse‐Radio Ultra‐Wideband system, which is composed of static biomedical nodes mounted on a patient's body to collect vital data and send it wirelessly to a central node or subsequent analysis by healthcare professionals. The performance of this network, such as the effect of node location, the number of transmitted symbols, multiuser interference, and intersymbol interference, is evaluated. We also study the physical layer and quality of service of this proposed architecture.

Vehicular communication is characterized by a dynamic environment and high mobility. In this paper we present a shadow fading model targeting system simulations based on real measurements performed in underground gallery. As first results, we present in this paper the delay spread statistics for each investigated environment. We also study the large-scale, small-scale fading and extract some time channel parameters such as root-mean-square (RMS) delay for a realistic underground propagation environment at 5.9 GHz. Since there are so far few published results for these confined environments, the results obtained can be useful for the deployment of vehicular-to-vehicular (V2V) and vehicular-to-infrastructure (V2I) communication systems inside underground mines galleries.

Vehicular Ad-Hoc NETworks (VANETs) have received considerable attention in recent years, due to its unique characteristics, which are different from Mobile Ad-Hoc NETworks (MANETs) such as rapid topology change, frequent link failure, and high vehicle mobility. VANET will provide several applications and services such as cooperative collision avoidance, emergency warning messages, cooperative intersection collision avoidance, and traffic management. VANET’s routing objective is to conduct packets through a path in the network to their final destinations. Currently, most of the proposed VANET routing protocols focus on urban or highway environments. The contribution of the paper falls within the study of VANET routing protocol in extreme and complex environments such as underground mines. This paper addresses the need for a bio-inspired adaptive routing protocol in VANETs which can tolerate low-density network traffic with little throughput and delay variation.

The uses of a vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) communications in the mining industry are expected to open significant opportunities for collecting and exchanging data. However, besides enabling the service of these technologies, radio channel propagation should be investigated. In this paper, we present extensive channel measurement and characterization at the 5.9 GHz dedicated short-range communications (DSRC) frequency band. The measurements were performed in a real underground mine gallery. The primary purpose of this study is to characterize the large-scale and small-scale fading. We provide results for the root-mean-square (RMS) delay spread and Kurtosis of the received power for both stationary and moving car scenarios. We conclude the paper by identifying the under-researched aspects of the vehicular propagation and channel modeling in underground mines.

Energy consumption in the residential sector is a considerable source of a wide range of environmental problems. It has been mentioned that the energy consumption in this sector represents 46% of total energy consumption and 23% of greenhouse gas emissions. Technology and services can help consumers manage the energy generated by their renewable or traditional electric grid. Service oriented architecture (SOA) has recently become very popular. SOA is an approach to build distributed systems that deliver application functionality as services to end-user applications or to build other services We adopt an approach to formulate the problem of managing energy in the building in a general form of linear programming by optimizing three criteria: environmental, economic and user comfort. First, we formulate this problem into an integer linear programming (ILP) problem. A round-up heuristic algorithm which is based on linear programming relaxation is presented.

The mobile networks are continually evolving in order to support more users, to achieve higher data rates, and to provide new (multimedia) services. The next generation networks must be able to service heterogeneous traffic with diverse quality of service (QoS) requirements. Orthogonal frequency division multiple access (OFDMA) is important techniques for high data rate wireless multiuser communication systems, such as 3GPP Long Term Evolution (LTE) and IEEE 802.16 Worldwide Interoperability for Microwave Access (WiMAX), not only because of its flexibility in resource allocation, but also because of its ability to exploit multiuser diversity. The scheduling algorithms that both support the QoS and maintain the throughput required to ensure users' satisfaction are essential to the development of the next network generation. In this paper, different scheduling techniques were evaluated using OFDMA in several different scenarios. The goal is to analyze the properties of networks such as throughput an fairness.

Ultra-wideband (UWB) has attracted a lot of attention in the past few years. UWB offers several advantages over traditional narrow band. Transmitted reference (TR) receivers have been known for many decades, but there is a renewed interest for applications of TR receivers as a suboptimal solution for UWB communications because of the difficulty in estimating the channel accurately for an optimal solution. In this paper, the performance of a UWB-TR receiver in an UWB underground mine channel is evaluated. The performance at higher data rates in the case of inter-symbol interference (ISI) is also investigated.

Fingerprinting is chosen as the localization approach as fingerprinting has a higher accuracy than other approaches such as time-of-arrival or angel-of arrival. This paper introduces a positioning system based on IEEE802.15.4/ZigBee-based sensor networks. The system uses fingerprinting and employs Support Vector Machines (SVMs) to estimate node position. The system is cost-effective since it works with real deployed IEEE 802.15.4/ZigBee sensors nodes. The whole system requires minimal setup time, which makes it readily available for real-world applications.

Since several years, communications between fixed and mobile units have been studied in areas in underground tunnels. The leaky feeder (LF) is one of the most useful results of these studies. The LF technology in combination with other communication systems allows us to reduce the transmission power of the mobile entities without sacrificing quality of service. Because the leaky feeder cables are able to transfer both power and RF signal; these technologies are considered as particular case of line power communication. In this paper we present and analyze the results of narrowband and wideband radio channel characterization. The measured frequency bands are selected for uses in several applications (WPAN, CDMA, video and analog telephony). The measurement were performed in underground mine gallery.

Wireless Sensor networks (WSN) have become a focus of research in the last few years. WSN is composed of small battery-powered devices that has sensors and wireless communication capabilities. Energy management is one of the key issues in WSNs because it directly impacts the network life-time. In order to overcome this restriction, several energy-efficient approaches for different layers have been investigated. In this paper, energy optimization on physical layer is analyzed. The node's power consumption is optimized through scaling the modulation scheme used in node communications. Results show that an optimal modulation scheme can lead to the minimum power consumption over the whole wireless sensor network.

Underground mine galleries can be considered as complex transmission lines where multipath, attenuation, reflection, diffraction and scattering effects are dominants. However, some companies have started to deploy modern wireless system networks in mine galleries with the objective of increasing safety and productivity. In the last decade, ultra-wideband (UWB) technology has gained much interest for its application to wireless communications. This paper reports on experimental results of UWB channel propagation in an underground mine. Eigen-decomposition and subspace-based statistical signal processing on the autocorrelation matrix of the channel impulse response are used. We apply information theoretic criteria to estimate the number of significant eigenvalues. This result is then used to calculate the RMS delay spread of the channel.

The necessity for wireless communications in underground mines is well understood. Some companies have started to deploy modern wireless system networks in mine galleries with the objective of increasing safety and productivity. In the last decade, ultra-wideband (UWB) technology has gained much interest for its applications in wireless communications. A number of UWB channel measurement have been published in the literature. However, all these works treated environments such as office buildings, residential or industrial. This paper reports on experimental results of UWB channel characterization in underground mines. The communication channel is still not well modelled in these environments. Important channel parameters such as path loss exponent, shadow fading, spatial correlation, small-scale fading, RMS delay spread and mean excess delay are investigated. This work has been carried out at the underground communications research laboratory LRCS, and at the experimental mine CANMET (Canadian Center for Minerals and Energy Technology) in Val-d'Or, Canada.

A procedure of measurement and analysis for the UWB channel in underground mines is presented. The measured data is saved in the frequency-domain via a vector network analyzer (VNA). We first compare two methods for the analysis of the path loss dependence on frequency. Then, we present results from an autoregressive modeling technique. We show that a two pole model is sufficient to represent the characteristics of the UWB channel in underground mines. This work was carried out by the underground communications research laboratory LRCS1, and the CANMET (Canadian Center for Minerals and Energy Technology) experimental mine in Val-d'Or, Canada.