Search Results (1,690)

Direct winding places magnet wires in direct contact with soft magnetic composite (SMC) stator cores, making the winding-corner response important for component design. This study evaluated an Fe–5.0 wt.%Si SMC stator core through finite element analysis and corner-loading tests. Based on a wire tension of 1.44 N/turn and a 90° change in winding direction, the resultant load was 2.04 N per turn and 38.7 N for 19 turns. An engineering verification load of 177 N was defined from the mean nominal 0.2% proof load of the AI-EIW wire. The equivalent distributed model at 38.7 N and the representative one-turn circular-contact model at 2.04 N predicted elastic responses without effective strain. At 177 N, the maximum effective stress and strain were 876.9 MPa and 0.0132, and the maximum resultant displacement after unloading was 0.85 μm. Component-level tests showed continuous force increases to the prescribed loads. Optical microscopy at 20× magnification identified no indentation, cracking, edge chipping, particle detachment, or corner-profile change after unloading, although the FEA-predicted sub-micrometer localized deformation was not directly quantified by the optical evaluation. These results characterize the local mechanical response under the defined reference winding and engineering verification conditions. Full article

Rapid maritime incident response requires visual recognition near the point of observation, but practical deployment must balance detection accuracy, inference latency, hardware-conversion constraints, and limited communication capacity. This study benchmarks the visual-compute component of an Event-Centric Edge Intelligence (ECEI) architecture on two maritime edge tiers: a Jetson AGX Orin 64 GB Developer Kit and a reComputer AI R2000-12 based on Raspberry Pi 5. Three tasks were selected to represent distinct maritime event categories: Human Detection for person-overboard response, six-class Ship Detection for vessel monitoring and collision awareness, and eleven-class Marine Trash Detection for floating-debris and pollution monitoring. YOLO26, YOLO11, and YOLOv8 were evaluated across 54 model–dataset–device configurations. Repeated on-device validation used 100 deterministically selected images per task and 20 measured repetitions for each checkpoint, yielding 108,000 timed predictions. On the paired S/N subset, the median reComputer/AGX mean-latency ratio was 4.40×; mean latency ranged from 19.84–57.09 ms on AGX and 68.89–172.11 ms on reComputer. The highest AGX $mA{P}_{50–95}$ values were 0.331, 0.813, and 0.639 for Human, Ship, and Marine Trash Detection, respectively; the highest reComputer values were 0.318, 0.806, and 0.558. Across the evaluated event-message configuration, event-candidate JSON bytes were 98.20% lower than raw JPEG bytes. These results support deployment-aware visual-tier selection and an auditable detector-to-event interface for maritime edge intelligence. Full article

Pedestrian safety at urban intersections remains a major challenge in Intelligent Transportation Systems (ITSs). This study investigates whether crossing intention can be reliably inferred directly from temporal body-pose dynamics to drive real-time collision warnings on embedded edge platforms. Existing vision-based approaches that rely primarily on bounding-box proximity or scene-level spatial grids are often prone to false alarms in complex urban environments with motorcycles, stationary pedestrians, and background clutter. To overcome these limitations, we propose an end-to-end framework consisting of four sequential processing stages: (1) a perception layer integrating YOLOv8s, ByteTrack, a displacement filter, and rider suppression to generate reliable pedestrian trajectories; (2) a skeleton extraction layer utilizing YOLOv8s-pose to construct temporal sequences of 17 anatomical keypoints; (3) an ultra-lightweight Skeleton Graph Convolutional Network (SkeletonGCN, comprising 33.8 K parameters, <0.2 MB) that models body-joint kinematics and temporal motion dynamics; and (4) an image-space Time-to-Collision (TTC) risk-fusion module. While this fusion approach avoids explicit geometric camera calibration, it still relies on predefined scene-profile parameters and image-space motion assumptions. Furthermore, while the intention classifier is quantitatively evaluated, the risk-fusion module is procedurally defined, and its resulting four-level collision warnings are demonstrated operationally rather than validated against ground-truth hazard annotations. Evaluated on 49,948 valid sequences from the JAAD and PIE benchmark datasets under a strict video-level partitioning protocol, the unified SkeletonGCN achieves a macro-F1 score of 0.717 (with per-scene subset macro-F1 scores of 0.761 on JAAD/PIE urban and 0.895 on intersections), significantly outperforming baseline models. When deployed on an NVIDIA Jetson Orin NX edge device using TensorRT FP16, the full pipeline achieves an instrumented latency of 70.7 ms per frame (~14 fps) and a sustained wall-clock throughput of 7.4 fps on real-world urban dashcam video. System limitations include sensitivity to 2D printed human imagery and reduced prediction reliability under low-light nighttime conditions. Full article

Colles fracture is a common type of wrist fracture, typically resulting from falling onto an outstretched hand. Postoperative patients often require long-term self-rehabilitation to restore wrist function. However, traditional rehabilitation relies on medical personnel and lacks real-time feedback and progress tracking, which may lead to poor rehabilitation or even deterioration of the condition. This article proposes an intelligent wearable system based on edge computing to enhance the efficiency of self-rehabilitation, improve system portability, and ensure comprehensive recording of rehabilitation data. The system performs real-time data processing and feedback to assist patients and healthcare providers in monitoring rehabilitation progress and optimizing recovery outcomes. The proposed system integrates an STM32 microcontroller, NanoEdge AI, and a 9-axis inertial sensor to detect and evaluate the accuracy of hand rehabilitation movements, ensuring the precision of self-rehabilitation. All rehabilitation movements are schemed under the guidance of professional physicians to ensure correctness and minimize the risk of secondary injuries caused by improper rehabilitation. Data such as motion records, training duration, and count are transmitted via Bluetooth to the Local database for further analysis. A custom web interface allows healthcare providers to monitor and analyze collected data. This approach improves diagnostic accuracy, supports personalized rehabilitation recommendations, and improves treatment outcomes and patient recovery success rates. Full article

Smart systems, the Industrial Internet of Things, and cyber-physical networks increasingly make decisions on the devices where data is generated, on nodes short of memory, compute, energy, and bandwidth, and exposed to real adversaries. Two research currents have grown to meet this: one makes artificial intelligence small and distributed (quantization, pruning, distillation, TinyML, federated and split learning), the other makes it safe (defenses against poisoning, backdoors, inversion, and evasion). This review argues that the two are entangled rather than parallel. Operators that shrink a model or scatter it across nodes also redraw its attack surface, each carrying a security dividend and a security liability, and because a node’s resources are finite and shared, model capacity and defense strength compete for one multi-dimensional budget. We formalize this as an efficiency-decentralization-security (EDS) design tension, explicitly a tension and not an impossibility, and show with published measurements that the coupling is non-monotonic. Around this thesis we build three artifacts, following an explicit design-science research process: an evidence-graded scoring matrix that separates each operator’s security dividend from its liability across seven axes and reports the direction of every effect separately from the confidence in the evidence behind it; a resource-aware threat model that judges attack and defense feasibility against a tiered device, gateway, network, and server budget with stated units; and a co-design framework whose decision workflow terminates in a defense-selection program and a verification step under adaptive attack. We work the framework through an industrial predictive-maintenance scenario with the resource arithmetic computed line by line, and evaluate it retrospectively against six published edge-AI systems. The result is a decision-support guide for building edge AI that is efficient, decentralized, and secure at once. Full article

Industrial water treatment plants are increasingly dependent on cyber–physical systems (CPS) and automated control processes for their operational safety and efficiency. However, the embedding of digital control networks exposes these critical infrastructures to sophisticated cyber–physical attacks, including malicious tampering with chemical dosing units and physical actuators. This paper proposes a robust, AI-enhanced anomaly detection framework designed to identify multi-stage malicious activities in water treatment systems using real-world industrial datasets. The proposed system is developed and validated on the Secure Water Treatment (SWaT) dataset, which contains multivariate sensor and actuator time-series data collected from a fully operational physical testbed under both normal operations and targeted cyber–physical attacks. First, high-frequency sensor noise is filtered, and cross-channel measurement reliability is maximized using a Kalman filter-based sensor fusion module. Subsequently, the fused-state vector is analyzed using an unsupervised Isolation Forest algorithm optimized for high-dimensional boundary isolation. To eliminate false negatives caused by stealthy, low-amplitude data injections that bypass purely statistical models, a deterministic, rule-based verification layer derived from physical process control logic is integrated. By integrating a discrete linear Kalman filter with an unsupervised Isolation Forest and deterministic physical rules, the framework effectively suppresses high-frequency sensor noise, achieving a 67.8% reduction in root mean square error (RMSE), while maintaining high detection accuracy across complex industrial attack scenarios. Experimental results demonstrate that the proposed hybrid framework yields superior detection capability, achieving a Precision of ≈95%, a Recall of ≈93%, a scenario-level F1-score of 94.1 % (alongside a sample-level F1-score of 21.5 %) and an edge inference latency of 0.6 ms, effectively demonstrating its suitability for deployment within simulated real-time industrial edge computing environments. The findings further confirm that combining statistical machine learning, state-space sensor fusion, and invariant physical process logic provides a resilient defense paradigm for securing critical industrial infrastructure against modern cyber–physical threats. Full article

This systematic review examines how artificial intelligence (AI) technologies relevant to human-factor risk reduction can be integrated into international and Spanish maritime safety frameworks. The formal PRISMA corpus comprises 23 core sources (13 peer-reviewed studies and 10 regulatory, institutional or technical documents), while a separate supplementary search provides recent independent technical and regulatory evidence up to 31 August 2026. The analysis covers computer vision, thermal and near-infrared sensing, multimodal fusion, behavioral and fatigue analysis, and onboard edge processing, with particular attention to precision, recall, false alarms, latency, computational requirements and operational robustness. The evidence shows that high detection performance can be achieved in specific maritime datasets, but the reported values depend strongly on the task, sensor, dataset and hardware and cannot be treated as a universal accuracy threshold. A system architecture is therefore proposed in which heterogeneous sensors feed synchronized edge processing, event verification, alarm management, VDR-compatible event logging, and human confirmation with defined fail-safe behavior. On the regulatory side, the study proposes staged adaptations of SOLAS, the ISM Code, STCW, MLC and Spanish inspection frameworks. The 2026 IMO MASS Code, considered as supplementary regulatory evidence, provides a relevant precedent for goal-based approval, risk assessment and progressive operational experience. Fixed tonnage and implementation-date thresholds are consequently treated as illustrative parameters rather than validated requirements; any mandatory carriage provision should be supported by formal safety assessment, type approval and operational evidence. The resulting framework links electronics implementation with a short-, medium- and long-term regulatory roadmap for safer merchant-ship operations. Full article

The rapid growth of renewable energy sources, particularly solar and wind power, has increased the demand for efficient and reliable battery energy storage systems (BESSs) to address intermittency, enhance grid stability, and improve energy management. In recent years, artificial intelligence (AI) has emerged as a transformative technology for optimizing the operation, control, monitoring, and maintenance of battery storage systems. This review provides a comprehensive overview of AI-driven BESS technologies for renewable energy applications. The study examines recent advances in machine learning, deep learning, reinforcement learning, and hybrid intelligent algorithms applied to battery state estimation, energy management, fault diagnosis, predictive maintenance, thermal management, and lifetime prediction. Furthermore, the integration of AI-based BESSs with photovoltaic systems, wind farms, microgrids, and smart grids is critically analyzed. The review highlights the advantages of AI techniques in improving system efficiency, reliability, adaptability, and decision-making capabilities under uncertain operating conditions. Current challenges, including data quality, model interpretability, computational requirements, cybersecurity concerns, and real-time implementation issues, are also discussed. Finally, emerging research directions such as digital twins, explainable artificial intelligence, federated learning, and edge intelligence are explored to provide insights into the future development of intelligent battery storage systems. This review aims to serve as a valuable reference for researchers, engineers, and practitioners working at the intersection of artificial intelligence, battery technologies, and renewable energy systems. Full article

Artificial intelligence can extend the functional capabilities of embedded sensors by extracting application-level information from physical measurements. This study investigates whether footstep-induced floor vibrations acquired with a single triaxial MEMS accelerometer contain sufficient information to characterize pedestrian path orientation and travel sense. A custom sensing platform based on an ADXL355 accelerometer and an ESP32 microcontroller was developed to acquire the structural vibration response at 4 kSPS. Lightweight temporal features were processed using a Random Forest classifier. The primary assessment used leakage-aware event-level cross-validation, with complete footsteps as the data-partitioning units. Under this more conservative protocol, discrimination of the complete A–K movement-label set was poor, whereas a compact 12-dimensional descriptor representation achieved 73.18% accuracy, 71.52% balanced accuracy, and 70.98% macro-F1 for X/Y path-orientation classification. Reliable positive/negative travel-sense discrimination could not be demonstrated from isolated footsteps. For historical comparison, the original sample-level procedure yielded 97.09% accuracy, but this value is retained only as a within-sequence reference because densely sampled observations contain strongly overlapping information. The findings provide proof-of-concept evidence that a single floor-mounted MEMS accelerometer can capture coarse pedestrian path-orientation information without cameras or spatially distributed vibration-sensor networks. Feature extraction and classifier inference were performed offline; broader validation across participants, sessions, floor structures, and realistic disturbances is required before deployment as an embedded edge AI sensing node. Full article

The present systematic review set out to reassess whether claims of energy-efficient deep learning in transportation infrastructure are supported by direct sustainability evidence. Deep learning models run in transportation systems on edge devices with a limited energy budget, and the literature labels them “green” or “energy-efficient”; the share of studies that support the label with measurement had not been quantified. Following PRISMA 2020, the Scopus, IEEE Xplore, and Web of Science databases were searched for the period from January 2020 to June 2026. Included were 721 studies applying Green AI techniques: pruning, quantization, knowledge distillation, lightweight architectures, TinyML, and dedicated accelerators. The review covers the transport domains of roads and ADAS, railway, connected and autonomous vehicles, and sensor networks. Studies were classified by the strongest efficiency evidence they report: a direct sustainability metric (energy, power, power efficiency, battery life, CO₂) or computational proxies. A direct metric is reported by 58 studies (8.0%); the share is a lower-bound estimate. Full-text verification of a stratified random sample of 34 Tier 2 studies found one study with a direct metric not stated in its abstract (2.9%); the sample-adjusted estimate of the share is 10.7% (95% confidence interval 8.5 to 21.8%). The evidence levels differ: 48 studies (6.7% of the corpus) report power or energy measured on the target hardware, two derive battery life from a measured energy budget, seven report modelled or simulated values, and one a macro-level CO₂ estimate. Railway contributes three studies. The largest measured reduction in energy per inference is 1961.8-fold (0.005 J on an FPGA against 9.77 J on a 95 W CPU); the largest modelled factor in the corpus is approximately 2400-fold (a memristor accelerator against an embedded GPU). Measured and modelled values are distinguished throughout the text. The studies that measure show that rigorous reporting is feasible; from the evidence presented follows the recommendation that an efficiency claim in transportation AI be supported by a direct metric measured or explicitly modelled on a named target platform. Full article

The You Only Look Once (YOLO) has been widely adopted in aquaculture monitoring and management due to its real-time performance and deployment flexibility. The recently introduced YOLO26 architecture incorporates Non-Maximum Suppression (NMS)-free end-to-end inference and is optimized for deployment on resource-constrained CPU-based devices, making it particularly relevant for edge deployment in commercial aquaculture applications. Nevertheless, its performance, operational efficiency, and deployment suitability compared with previous YOLO generations remain largely unvalidated in aquaculture-specific scenarios. This study benchmarks YOLO26 against three Ultralytics predecessors (YOLOv5u, YOLOv8, and YOLO11) across nano, small, and medium model scales for the detection of fish mortality, a critical indicator of fish population health and welfare, in recirculating aquaculture systems (RAS). Twelve model variants were evaluated for detection accuracy, training efficiency across seven dataset sizes, and inference performance on both high-performance NVIDIA A100 GPUs and the resource-constrained, CPU-only Raspberry Pi 5 edge device. All models achieved comparable performance on the full dataset, with mAP50 varying by only 1.25 percentage points across three independent training runs, indicating minimal influence of architectural generation on final mortality detection accuracy when sufficient training data are available. However, notable differences emerged in data efficiency and deployment performance. YOLOv8 demonstrated the strongest training efficiency, achieving 90% mAP50 with only 400 training images, whereas YOLO26 nano and small variants required 1000 images to reach comparable accuracy. In contrast, YOLO26 exhibited advantages during edge deployment, with YOLO26n achieving the highest inference speed on the Raspberry Pi 5 at 7.84 ± 0.13 FPS across three benchmark sessions, while YOLOv5mu outperformed all contemporary medium-scale architectures on CPU-based hardware. These results demonstrate that architectural novelty alone is an insufficient criterion for model selection. The findings support a deployment-oriented framework in which training data availability, target hardware, and inference requirements collectively inform model selection for aquaculture applications. Full article

Physics-informed anomaly detection for industrial control systems (ICSs) is usually validated on synthetic data. How much of that performance survives on real hardware-in-the-loop (HIL) data is an open question. We re-implement PhySec-Edge, a hybrid BiLSTM/β-VAE/GAT/Kalman pipeline, as a full architecture and evaluate it on HAI 23.05. The gap is consistent across five seeds. AI-only F1 falls from 0.774 (synthetic) to 0.525 (HAI); the full hybrid falls from 0.798 to 0.512. The components diverge. PVE residual sharing transfers but weakens on real data (ΔF1 = +0.010 ± 0.008 across five seeds against +0.017 synthetic; ΔMCC = +0.013, positive in every seed). The PVE NIS decision score fails outright (raw FPR = 0.482; a correction brings it to 0.013, but detection does not recover). Under the retrospective fixed-FPR comparison, the preferred configuration is GNN+BiLSTM (F1 = 0.530 ± 0.002), outperforming full-EADE fusion (0.512 ± 0.003). From this split we derive the information-vs-decision pathway distinction, a way to reason about component-level generalization in hybrid architectures. We then set out what it means for physics-informed ICS detector design and edge deployment. Full article

As smart sensing applications grow rapidly, the IoT edge architectures need to support low latency, make decisions with little memory, power, and communication resources while preserving the privacy of the data. But traditional cloud-based IoT solutions come with transmission delay, increased energy consumption, and privacy issues because of the constant transfer of raw data. In this paper, we propose a privacy-preserving TinyML-driven IoT edge architecture that enables real-time smart sensing and autonomous AI-based control. The proposed framework includes on-device sensor pre-processing, lightweight TinyML inference, adaptive model selection, encrypted feature-level communication, trust-aware decision validation, and local control execution. Raw data streams from the sensors are also processed locally, and only compact encrypted features or a summary of the decisions are sent if necessary to minimize privacy exposure. The experimental evaluation reveals that the proposed architecture has an accuracy of 97.4%, an F1 score of 96.9%, and a secure-event detection rate of 98.1% and reduces the inference latency by 14.8%, the energy consumption by 4.1 mJ per inference and the amount of data transmitted by 74.5% compared to conventional edge-cloud processing. Results show that the proposed architecture is a scalable, privacy-aware, and energy-efficient solution for real-time autonomous IoT control of smart environments with limited resources. Full article

Quantum communication is emerging as a foundation for next-generation networks, offering unprecedented capabilities in security, entanglement-based connectivity, and distributed computation. However, the classical Open Systems Interconnection (OSI) model, designed for deterministic, error-tolerant systems, is incompatible with quantum phenomena such as decoherence, probabilistic entanglement, and the no-cloning theorem. This paper surveys and redefines the OSI model for quantum networking in the context of 7G systems. We propose a Quantum-Converged OSI stack by extending the classical seven-layer model with two additional layers: (i) Layer 0, the Quantum Substrate, responsible for entanglement management, coherence preservation, and teleportation; and (ii) Layer 8, the Cognitive Intent Plane, which enables AI- and QML-driven orchestration. The survey synthesizes over 150 research works published between 2018 and 2025, classifying them by OSI layer, enabling technologies (e.g., Quantum Key Distribution, Quantum Error Correction, and Post-Quantum Cryptography), and application domains such as satellite quantum links, quantum IoT, and federated edge systems. We further provide a taxonomy of cross-layer enablers and discuss simulation tools, including NetSquid, QuNetSim, and QuISP. Finally, an evaluation framework with quantum-native metrics, such as entropy throughput, coherence latency, and entanglement fidelity, is introduced, along with open challenges for programmable stacks, digital twins, and AI-defined quantum agents. The specific and novel contribution of this work is a Quantum-Converged OSI stack that extends the classical seven-layer model with two additional layers: Layer 0, the Quantum Substrate, responsible for entanglement management, coherence preservation, and teleportation; and Layer 8, the Cognitive Intent Plane, which enables AI- and QML-driven orchestration. Unlike prior technology-centric surveys, the proposed framework classifies over 150 research works by OSI layer, maps enabling technologies (QKD, QEC, PQC) and application domains (satellite quantum links, quantum IoT, federated edge systems) to their functional layers, and introduces a quantum-native evaluation framework based on entropy throughput, coherence latency, and entanglement fidelity. This layer-resolved synthesis, together with the formal definition of cross-layer quantum-native metrics, constitutes the principal novelty distinguishing this survey from existing quantum-networking reviews. Full article

AI-assisted renewal of living vernacular settlements must preserve regional identity and historical layering while keeping professional judgment central; however, current facade-generation studies rarely connect field-derived architectural knowledge, controllable data construction, and subsequent design development within one verifiable workflow. Taking Yun’ao Town as a case study, this study develops a Grasshopper–Pix2Pix method for preliminary renewal of principal (front) facades—the primary public-facing elevations—in a hot-humid coastal setting exposed to disaster-prone weather, which is treated only as maintenance and renewal background rather than as a generative input, constraint, or performance metric. Field surveys documented 310 representative buildings and 1240 facade images. Typological and component analyses informed 160 parametric base models, which were expanded into 870 Facade Semantic Layout (FSL)–Building Elevation (BE) image pairs using four fixed parameters and eleven groups of linked parameters. The dataset was allocated at approximately 80%/10%/10% for training, validation, and testing at the base-model-group level to reduce information leakage, and Pix2Pix was trained for 1000 epochs. Across six representative cases, SSIM ranged from 0.33 to 0.65, with a descriptive mean of 0.54; these values do not represent aggregate validation- or test-set performance. Qualitative inspection identified distorted openings, incomplete balcony edges, interrupted roof lines, unclear railing relationships, ornamental discontinuity, and occasional component overlap. A mini-program and Huangang Road application examined how selected outputs could enter a human-supervised sequence of preliminary comparison, professional screening, manual correction, three-dimensional modeling, and technical drawing development. The contribution is a field-informed, human-supervised generative workflow rather than autonomous design; its outputs remain preliminary visual references requiring measured-survey, conservation, structural, environmental, material, and construction verification. Full article