Salient object detection in RGB-D imagery remains challenging due to inconsistent depth quality and suboptimal cross-modal fusion strategies. This paper presents a novel dual-stream architecture that integrates contextual feature refinement with adaptive attention mechanisms for robust RGB-D saliency detection. We extract two features from the ResNet-50 backbone for both the RGB and depth streams, capturing low-level spatial details and high-level semantic representations. We introduce a Contextual Feature Refinement Module (CFRM) that captures multi-scale dependencies through parallel dilated convolutions, enabling hierarchical context aggregation without substantial computational overhead.... More >

Fire incidents cause devastating environmental damage and human casualties, necessitating robust automated detection systems. Existing fire recognition methods struggle with visual ambiguities, illumination variations, and computational constraints, while current attention mechanisms lack hierarchical integration for comprehensive feature refinement. We propose a cascaded multi-attention architecture that combines Multi-Scale Strip Attention (MSSA), Optimized Spatial Attention (OSA), and the Convolutional Block Attention Module (CBAM) to enhance fire detection. MSSA employs three-scale orthogonal strip pooling to capture fire patterns across varying spatial extents through horizontal and ver... More >

RGB–thermal (RGB-T) salient object detection exploits complementary cues from visible and thermal sensors to maintain reliable performance in adverse environments. However, many existing methods (i) fuse modalities before sufficiently enhancing intra-modal semantics and (ii) are sensitive to modality discrepancies caused by heterogeneous sensor characteristics. To address these issues, we propose PACNet (Pyramid Attention Collaboration Network), a hierarchical RGB-T framework that jointly models multi-scale and global context and performs refinement-before-fusion with cross-modal collaboration. Specifically, Dense Atrous Spatial Pyramid Pooling (DASPP) captures multi-scale contextual cues... More >

Video summarization (VS) aims to generate concise representations of long videos by extracting the most informative frames while maintaining essential content. Existing methods struggle to capture multi-scale dependencies and often rely on suboptimal feature representations, limiting their ability to model complex inter-frame relationships. To address these issues, we propose a multi-scale sensing network that incorporates three key innovations to improve VS. First, we introduce multi-scale dilated convolution blocks with progressively increasing dilation rates to capture temporal context at multiple levels, enabling the network to understand both local transitions and long-range dependencie... More >

Fire detection in smart cities requires intelligent visual recognition systems capable of distinguishing fire from visually similar phenomena while maintaining real-time performance under diverse environmental conditions. Existing deep learning approaches employ attention mechanisms that aggregate spatial information isotropically, failing to capture the inherently directional characteristics of fire and smoke patterns. This paper presents DirFireNet, a novel fire detection framework that exploits directional fire dynamics through Strip Pooling Coordinate Attention (SPCA). Unlike conventional attention mechanisms, DirFireNet explicitly models vertical flame propagation and horizontal smoke d... More >

In response to the challenges of insufficient accuracy in face detection and missed small targets under low-light conditions, this paper proposes a detection scheme that combines image preprocessing and detection model optimization. Firstly, Zero-DCE low-light enhancement is introduced to adaptively restore image details and contrast, providing high-quality inputs for subsequent detection. Secondly, YOLOv11n is enhanced through the following improvements: a P2 small-target detection layer is added while the P5 layer is removed, addressing the original model's deficiency in detecting small targets and streamlining the computational process to balance model complexity and efficiency; the P2 up... More >

As the core component of the copper top-blown smelting, the service life of the injection lance critically affects production stability. To monitor the operating condition of the injection lance, a data-driven model is proposed to predict the Remaining Useful Life (RUL) or service life, namely, the DKT-LSSVM model. Firstly, to reduce noise interference, the Daubechies wavelet with four vanishing moments (DB4) denoising is used to process the raw data. Then, the Kernel Principal Component Analysis (KPCA) method is utilized to extract the principal components from the denoised data, which retains at least 90% information content (18 principal components are obtained). These principal component... More >

Federated learning is a collaborative machine learning paradigm that trains models across multiple computing nodes while aiming to preserve the privacy of local data held by participants. However, because of the open network environment, federated learning faces severe privacy and security challenges. Studies have shown that attackers can reconstruct original training data by intercepting gradients transmitted across the network, thereby posing a serious threat to user privacy. One representative attack is the Deep Leakage from Gradients (DLG), which iteratively recovers training data by optimizing dummy inputs to match the observed gradients. To address this challenge, this paper proposes a... More >

This study briefly discusses the primary AI’s roles in enhancing control engineering education (CEE), which has the potential to revolutionise the teaching-learning framework by making complex concepts and methodologies more intuitive, interactive, and application-driven. While understanding the potential benefits of these AI tools, such as assisting with problem-solving in education, some of the concerns about their use are summarised. An example is discussed how AI enhances CEE in MATLAB \& Simulink. The centre point in the brief paper is that AI should be a tool to enhance teaching-learning, rather than a shortcut to avoid it. More >

A fixed-time adaptive optimal parameter estimation (FxT-AOPE) scheme is proposed to address the difficulties in estimating dead zone parameters and slow convergence speed of tracking errors in permanent magnet synchronous motor systems. First, the continuous piecewise linear neural network is used to model the nonlinear dead zone dynamics. Second, an auxiliary filter is constructed to extract estimation errors, and this filter is used to drive an adaptive law with time-varying gain, minimizing the cost function of estimation errors and achieving adaptive optimal parameter estimation (AOPE). Then, the AOPE method is introduced into the fixed-time non-singular terminal sliding mode control (Fx... More >

Strain sensors have become fundamental to contemporary sensing technology driven by the growing demand for flexible, sensitive, and durable sensors, transforming across a broad range of applications including medical care, robotics, structural monitoring, human-machine interface and robotics. The swift progress in the fields of materials and nanotechnology has facilitated the fabrication of very flexible, resilient, and ultra-sensitive strain sensors, enabling the emergence of next-generation electronic devices. This mini review covers sensors, strain sensors' fundamentals, classifications, innovative materials utilized, applications overall provide an in-depth analysis of the latest develop... More >

The Internet of Vehicles (IoV) is a core component of smart transportation systems, making it feasible to exchange information among vehicles, infrastructure, and central systems in real time. However, the effective use of resources and the efficient distribution of tasks in these dynamic environments is a challenging task. This paper presents a blockchain-based collaborative task allocation framework method that can solve these problems by using a greedy algorithm for general task allocation and adopting a dynamic collaboration scheduling algorithm for emergent tasks. Employing the blockchain-based reward mechanism, the transparency, fairness, and security in dynamic mobile crowdsensing (MC... More >

This study explains the main idea and structure of a What-How intelligent control (WH-I-control) system and a universal control (U-control) system. The system has two control layers. The bottom layer uses the U-control framework to manage 'How' to control within a universal framework. The top layer uses intelligent control (I-control) to coordinate and guide 'What' to achieve both global and local control goals. This study also reviews the configurations, functions, and integration of these two control layers in analysis, design, and applications. More >

Ensuring the freshness of food products is essential for both acute and chronic health outcomes. However, significant health risks can be triggered by dietary resources subjected to improper storage protocols. Current methods are often unreliable and unfeasible for detecting food freshness. This research proposes an IoT-based food freshness detection system that uses biosensors and gas sensors to monitor perishable items like meat, produce, and dairy. The system is integrated with a mobile application that allows users to analyze food quality in real-time, based on predefined degradation thresholds. This study assists in providing valuable insights for future research and improving food safe... More >

To address evolving security challenges in cloud computing, this study proposes a hybrid deep learning architecture integrating Bidirectional Long Short-Term Memory (BiLSTM) and Bidirectional Gated Recurrent Units (BiGRU) for cloud intrusion detection. The BiLSTM-BiGRU model synergizes BiLSTM's long-term dependency modeling with BiGRU's efficient gating mechanisms, achieving a detection accuracy of 96.7% on the CIC-IDS 2018 dataset. It outperforms CNN-LSTM baselines by 2.2% accuracy, 3.3% precision, 3.6% recall, and 3.6% F1-score while maintaining 0.03% false positive rate. The architecture demonstrates operational efficiency through 20% reduced computational latency and 15% lower memory foo... More >

This paper focuses on the implementation of a control framework for a fixed wing aircraft system and the simulation demonstrations. The aim is to develop several Proportional Integral Derivative (PID) controllers to stabilise the altitude and attitude in a 2D environment by regulating the engine power, the pitch angle, and height in flight operation. In technique, a dynamic mathematical model is established by considering the degrees of freedom and the dynamics of motion of a fixed wing aircraft, which provide a foundation for design and simulation. A simplified aircraft dynamic model is tailored for testing the formed control systems, which can be flexibly modified with different aircraft c... More >

3D holographic displays on vehicles are a new out-of-home (OOH) advertisement technology offering an interactive and dynamic means of engaging consumers. They enhance advertisement visibility and viewer interaction through the projection of colorful, three-dimensional images in public areas. An Internet of Things (IoT) based holographic advertisement system is presented, integrating remote content management and personalized advertising control. The system includes an Android application that allows users to select and schedule ads displayed on moving vehicles. A GPS module provides real-time vehicle tracking, supporting security and targeted advertisements. Cloud-based storage ensures remot... More >

The Tuned Liquid Damper (TLD) method offers a practical and cost-effective solution for seismic design. Accurate modeling of the TLD system’s dynamic behavior is crucial for optimizing its performance. In this study, the nonlinear dynamics of the TLD system are characterized using the Housner model, with parameters estimated via a nonlinear state estimation approach. To address challenges associated with model discretization and unknown noise processes, we introduce a Robust Extended Kalman Filter (REKF) that incrementally incorporates uncertainties to more accurately capture system dynamics. The proposed method is evaluated through real-time hybrid simulation, employing seismic input sign... More >

This study presents an intelligent automated system for real-time detection and classification of tomato diseases using a Convolutional Neural Network (CNN) integrated within an Internet of Things (IoT) based unmanned ground vehicle (UGV). The CNN was trained and evaluated using a dataset comprising over 20,000 images of tomato leaves categorized into ten distinct diseases—Late Blight, Early Blight, Septoria Leaf Spot, Tomato Yellow Leaf Curl Virus, Bacterial Spot, Target Spot, Tomato Mosaic Virus, Leaf Mold, Spider Mites Two-Spotted Spider Mite, Powdery Mildew—and healthy leaves. The developed CNN architecture, optimized for lightweight deployment on edge devices like Raspberry Pi 4, ac... More >

The rapid urbanization and technological advancements have driven the development of smart cities, envisioned as sustainable, efficient, and interconnected urban spaces. The integration of sixth-generation (6G) wireless technology in smart cities promises unprecedented opportunities in connectivity, low-latency communication, and data management, which transforms urban living. However, this evolution raises critical ethical concerns related to privacy, inclusion, transparency, accountability, and environmental sustainability. This paper explores the ethical considerations inherent in designing smart cities with 6G, emphasizing data governance, equity, and human-centric approaches. It delves... More >

Osteoarthritis (OA) is a degenerative joint disease that primarily affects the knee, causing cartilage deterioration and discomfort. Early diagnosis is crucial for effective management, as it can slow disease progression and improve the quality of life. This study proposes a deep learning approach to automatically classify knee OA severity from X-ray images using Convolutional Neural Networks (CNNs) and the VGG16 model. The models were trained on a dataset of knee X-ray images, and performance was evaluated using accuracy, precision, recall, and F1-score. The proposed CNNs model achieved 99% training accuracy and 80% testing accuracy after 50 epochs, while the VGG16 model, after fine-tuning... More >

Brain tumor detection (BTD) is a crucial task, as early detection can save lives. Medical professionals require visual intelligence assistance to efficiently and accurately identify brain tumors. Conventional methods often result in misrecognition, highlighting a critical research gap. To address this, a novel BTD system is proposed to predict the presence of a tumor in a given MRI image. The system leverages a convolutional neural network (CNN) architecture, combined with a multi-layer perceptron (MLP) for feature extraction and understanding complex pixel patterns. An extensive ablation study was conducted to empirically analyze and identify the optimal model for the task. The findings dem... More >

Accurate and timely detection of wheat diseases remains crucial for sustainable agriculture, particularly in major wheat-producing regions. Wheat diseases pose a significant threat to global food security, need precise and timely detection to promote sustainable agriculture. Existing approaches consistently employ single-scale features with shallow-layered convolutional neural networks (CNNs). To bridge the research gaps, we introduce a novel Multi-Scale Wheat Disease Network (MSWDNet) with feature collaboration for wheat disease recognition supported by a comprehensive dataset collected from wheat fields. This study fills research gaps by introducing a novel technique to improve detection a... More >

The main goal of this study is to design and develop a half-bridge inverter architecture specifically for high-voltage power supply applications. An effective, small, and affordable system that converts direct current (DC) to alternating current(AC) can be built, thanks to the IR2151 chip’s dependable characteristics and performance. To get the desired output voltage, the transformer first increases the voltage and then the voltage is increased with a voltage-doubling rectifier (VDR) circuit. The study emphasizes how crucial it is to choose components carefully and simulate the circuit design and implementation process to guarantee dependable performance. The experimental results validate... More >

In recent years, vehicular ad hoc networks (VANETs) have faced growing security concerns, particularly from Denial of Service (DoS) and Distributed Denial of Service (DDoS) attacks. These attacks flood the network with malicious traffic, disrupting services and compromising resource availability. While various techniques have been proposed to address these threats, this study presents an optimized framework leveraging advanced deep-learning models for improved detection accuracy. The proposed Intrusion Detection System (IDS) employs Convolutional Neural Networks (CNN), Long Short-Term Memory (LSTM), and Deep Belief Networks (DBN) alongside robust feature selection techniques, Random Projecti... More >

In engineering applications, high-precision tracking control is crucial for robotic manipulators to successfully complete complex operational tasks. To achieve this goal, this study proposes an adaptive tunable predefined-time backstepping control strategy for uncertain robotic manipulators with external disturbances and model uncertainties. By establishing a novel practical predefined-time stability criterion, a tunable predefined-time backstepping controller is systematically presented, allowing the upper bound of tracking error settling time to be precisely determined by adjusting only one control parameter. To accurately address lumped uncertainty, two updating laws are designed: a fuzz... More >

Smart Environment is rapidly growing with the inclusion of Artificial Intelligence of Things (AIoT) when it connects to future communication and social media networks. Security and privacy are significant challenges, including data integrity, account hijacking, cybersecurity, and cyberbullying. To mitigate these challenges, Social Media 3.0 is utilized with advanced emerging technologies such as Blockchain, Federated Learning (FL), and others and offers solutions in existing research. This article comprehensively reviews and proposes Next-Generation Technologies for Secure Future Communication Service Scenario for Smart Environment and Social-Media 3.0. We discuss existing attacks with their... More >

Humanoid robots play a significant role in numerous fields, where efficient and intuitive control inputs are essential, particularly in applications requiring remote operation. In this paper, we investigate the potential advantages of inertial sensors as a key component for generating command signals in humanoid robot control systems. The objective is to accurately detect user motion through inertial sensing, thereby enabling precise control commands. Finger gestures are first captured as signals from the inertial sensor, and movement commands are extracted through filtering and recognition processes. These commands are then translated into corresponding robot actions based on the sensor’s... More >

Collective motion has been a pivotal area of research, especially due to its substantial importance in Unmanned Aerial Vehicle (UAV) systems for several purposes, including path planning, formation control, and trajectory tracking. UAVs significantly enhance coordination, flexibility, and operational efficiency in practical applications such as search-and-rescue operations, environmental monitoring, and smart city construction. Notwithstanding the progress in UAV technology, significant problems persist, especially in attaining dependable and effective coordination in intricate, dynamic, and unexpected settings. This study offers a comprehensive examination of the fundamental principles, mod... More >

The efficient extraction and fusion of video features to accurately identify complex and similar actions has consistently remained a significant research endeavor in the field of video action recognition. While adept at feature extraction, prevailing methodologies for video action recognition frequently exhibit suboptimal performance in the context of complex scenes and similar actions. This shortcoming arises primarily from their reliance on uni-dimensional feature extraction, thereby overlooking the interrelations among features and the significance of multi-dimensional fusion. To address this issue, this paper introduces an innovative framework predicated upon a soft correlation strategy... More >