Search Results (6,734)

Distracted and fatigue-related driving behaviors are major causes of road traffic accidents, creating an urgent need for reliable driver monitoring systems. Vision-based detection methods have garnered widespread attention due to their low cost of deployment and practical applicability. However, existing lightweight models often suffer from limited global contextual perception and insufficient preservation of fine details. Motivated by these challenges, this study introduces an improved distracted and fatigue-related driving behavior detection model, YOLOv12-LAD, built on the YOLOv12 architecture. The proposed framework integrates a Large Separable Kernel Attention module (LSKA) to enhance global contextual perception, an Adaptive Downsampling module (ADown) to mitigate information loss during feature compression, and a Dynamic Sampling module (DySample) to enable content-adaptive feature reconstruction and improve multi-scale behavior representation. Experimental results show that YOLOv12-LAD achieved 97.5% precision, 96.3% recall, and 98.4% mAP@50 with only 2.5 million parameters, 6.2 GFLOPs, and an inference speed of 249 FPS. Ablation studies, comparisons with representative models, cross-dataset evaluation, and real-vehicle tests further verify the effectiveness and robustness of the proposed method. The proposed method demonstrates strong performance while maintaining computational efficiency, making it suitable for real-time vision-based driver monitoring applications. Full article

The increasing integration of intermittent renewable energy sources (RESs) into islanded Hybrid Power Systems (HPSs) is a critical step towards global energy sustainability; however, it poses significant challenges to frequency stability owing to low system inertia and stochastic power fluctuations. To address these challenges and enable higher penetration of green energy, this study proposes a novel and robust Load Frequency Control (LFC) strategy based on the Crested Porcupine Optimizer (CPO). A customized Mode-Dependent Adaptive Balloon (MDAB) controller is developed, wherein the virtual control gain is dynamically tuned based on the real-time operating modes and disturbance severity. Furthermore, to optimize communication resources and mitigate actuator wear in networked microgrids, an intelligent event-triggered (ET) mechanism is seamlessly integrated into the adaptive logic. The proposed control framework is rigorously validated through comprehensive nonlinear simulations and comparative analyses with state-of-the-art metaheuristic algorithms (GTO, GWO, JAYA, and GO). The evaluation encompasses step load disturbances, severe parametric uncertainties (+25%), realistic 24-h diurnal cycles with solar cloud shading and wind turbulence, and extended practical constraints, including Battery Energy Storage System (BESS) integration and Internet of Things (IoT) communication delays. The results demonstrate the superiority of the CPO-tuned framework, which achieved the fastest transient recovery (settling time of 3.4367 s) and the lowest absolute Integral Absolute Error (IAE). Additionally, the proposed ET-based strategy not only reduced the communication burden but also improved the overall control performance by 37% in terms of IAE compared with continuous approaches. By inherently filtering measurement noise, mitigating control signal chattering, and maintaining resilience under nonideal latency, the proposed architecture offers a highly robust and resource-efficient solution that directly guarantees the operational sustainability and reliability of modern smart microgrids. Full article

Cropland protection enforcement is central to food security and sustainable land management, yet small-scale encroachments within Permanent Basic Farmland (PBF) boundaries frequently evade conventional field surveys and reactive inspection regimes. Existing remote sensing approaches rely mainly on comprehensive land-cover classification or bi-temporal change detection, which often generate alerts beyond the regulatory scope and require annotation efforts that limit county-scale deployment. To address this gap, this study reframes PBF monitoring as a boundary-constrained anomaly screening task, defined as the detection of surface conditions that deviate from expected cultivation norms within legally defined parcels. To operationalise this task, we adapt a DeepLabv3+-based segmentation pipeline by incorporating an auxiliary edge branch and a composite loss to improve sensitivity to minority-class anomalies and preserve fragmented parcel boundaries. The model is trained on the LoveDA dataset and evaluated in Mancheng District, Hebei Province, China, without site-specific fine-tuning. Multi-temporal imagery from 2021 to 2023 is further used as a post hoc consistency check to distinguish persistent anomalies from transient surface conditions, rather than to model temporal dynamics explicitly. Cross-regional zero-shot evaluation further examines model robustness under heterogeneous environmental conditions. Benchmarked against five comparison architectures, the adapted pipeline achieves a Recall of 61.25%, representing a 10.24 percentage-point improvement over DeepLabv3+ and expanding the set of candidate encroachments for field verification. This result should be interpreted in terms of screening sensitivity rather than overall segmentation optimisation. The outputs are intended as preliminary screening leads that support, rather than replace, expert review. The principal contribution of this study therefore lies in reframing PBF monitoring as an operational anomaly-screening task aligned with enforcement needs, rather than in proposing a fundamentally new segmentation architecture. Full article

In the practice of historic building conservation and restoration, the authentic restoration of damaged components often faces challenges due to the lack of definitive design evidence. To address this issue, this paper proposes a restoration derivation method that integrates digital survey technologies, such as UAV oblique photogrammetry and 3D laser scanning, with the analysis of historical mathematical composition rules. Taking five Ming Dynasty pavilion-style stone pagodas in Fuzhou as subjects, this study first employed digital surveying and cross-verification with ancient texts to reveal their shared, precise proportional system: the eave–column ratio of the Ruiyun Pagoda approaches √2 (≈1.414), while the other four pagodas approach the golden ratio of 1.618. Furthermore, the pagoda silhouettes are governed by a √2 hierarchical system and a √3/2 visual correction mechanism. Based on these mathematical rules, a triple logical chain of “historical evidence verification–functional constraints–traditional adaptation” was constructed and applied to the quantitative restoration design of the damaged finial of the Luoxing Pagoda. This process ultimately derived the relationship between its total height and the first-story width as (L/2 + √2/2), with the finial height being 1/7 of the pagoda body’s total height. This case study validates the effectiveness of the proposed method in transforming profound historical wisdom into clear engineering parameters, offering a replicable and verifiable technical pathway for the digital conservation and scientific restoration of similar architectural heritage. Full article

The mathematical characterization of non-stationary signals remains a significant challenge, particularly when impulsive components are obscured by high-dimensional noise and structural coupling. This paper proposes an application-driven mathematical methodology for a learnable discrete wavelet transform (LDWT) that combines classical multi-resolution analysis with task-optimized data-driven adaptivity. Rather than introducing entirely new foundational theory, our approach strategically relaxes constraints from orthogonal wavelet theory within the non-perfect reconstruction filter bank framework, enabling controlled spectral decomposition optimized for supervised fault diagnosis. We introduce a specialized regularization term based on the half-band property to ensure spectral complementarity and minimize cross-band correlation, while a Jacobian-based stabilization approach is formulated to ensure the convergence of filter coefficients during optimization. The proposed algorithmic architecture, LDBRFnet, features a dual-branch encoder system designed to capture the mathematical synergy between sub-band-level global statistics and time-domain local morphology. This dual-view representation effectively mitigates feature leakage and overconfidence in classification. Theoretical analysis and numerical experiments demonstrate that the learned filters satisfy the frequency-shift property and maintain robust spectral partitioning even under low signal-to-noise ratios. Validation on complex vibration datasets confirms that the framework achieves superior diagnostic accuracy (over 95.5%) and computational efficiency, reducing model parameters by 96.7% compared to state-of-the-art baselines. This work provides a generalizable mathematical approach for adaptive signal decomposition and robust pattern recognition in interdisciplinary applications. Full article

Decentralised manufacturing is expanding as digitally controlled fabrication tools become accessible to SMEs, independent operators, and community workshops outside traditional factory settings, but the resulting heterogeneous, autonomously operated network introduces systemic uncertainty that no central authority governs. This paper proposes a systems-theoretic framework in which Free and Open-Source Software (FOSS) governance acts as the structural interoperability layer of a distributed cyber–physical manufacturing system (CPS), and node-local digital twins—each hosting a machine learning (ML) disturbance estimator—provide local adaptive compensation without centralised data aggregation. A defining property of the architecture is automatic improvement propagation: learned corrections distribute via federated learning to structurally similar nodes without operator intervention, and the open, observable FOSS ecosystem enables advances in one fabrication modality to transfer to others through shared interface standards. The framework is applied analytically to three disturbance classes: regulatory restriction, technical process variability, and supply chain disruption. Across cases, the analysis shows how open modular interfaces and local adaptation preserve functional continuity under perturbations that would more strongly affect centralised architectures. The contribution is a unified mathematical basis for robustness analysis in decentralised manufacturing CPS and a foundation for future simulation and empirical validation. Full article

Financial time series in turbulent markets exhibit complex long-memory dynamics and multifractal features that traditional deep learning models fail to capture due to inherent exponential forgetting mechanisms. To address this, we propose Frac-Attn-GL, a novel Fractional-order Spatiotemporal Attention-based GRU-LSTM framework. Grounded in the Fractal Market Hypothesis, the model embeds Grünwald–Letnikov fractional-order operators into a dual-channel architecture (FracLSTM and FracGRU) to characterize long-range memory with rigorous power-law decay priors. Furthermore, an extreme-aware asymmetric loss function is designed to drive a dynamic spatiotemporal routing mechanism, enabling adaptive shifts between long-term macro trends and short-term micro shocks. Empirical tests on major U.S. stock indices reveal three significant findings. First, the Frac-Attn-GL framework substantially reduces prediction errors, achieving up to a 93.1% RMSE reduction on the highly volatile NASDAQ index compared to standard baselines. Second, the adaptively learned fractional-order parameters exhibit a consistent quantitative alignment with the market’s empirical multifractal singularity spectrum, supporting the physical interpretability of the model’s endogenous memory mechanism. Finally, hybrid residual multifractal diagnostics indicate that the framework effectively captures deep long-range correlations, reducing the Hurst exponent of the prediction residuals from ~0.83 to approximately 0.50, a level consistent with the absence of significant long-range dependence. Full article

This article presents the design and development of an interactive architecture oriented toward the management of traditional vallenato, a musical genre recognized as an Intangible Cultural Heritage of Humanity by UNESCO. Architecture combines the principles of contextual awareness and the use of massive open online courses (MOOCs) to face the current challenges of preservation, dissemination, and teaching of this cultural expression, threatened by commercialization and the loss of its traditional roots. Through a modular structure, adaptive technological tools are integrated to capture, process, and use contextual information, personalizing learning experiences and strengthening the link between communities and their cultural heritage. The proposal consists of several functional layers, including context management, user profiles, educational resources, and a persistence unit, each designed to ensure the interoperability and sustainability of cultural data. In addition, the capacity of architecture to be used in other cultural contexts is highlighted, expanding its impact on different artistic manifestations and heritages worldwide. This article includes a comparative analysis with other existing models, highlighting the advantages of this solution in terms of customization and adaptability. Finally, opportunities for improvement and expansion are explored, as well as the pending challenges in the implementation of this technological tool in educational and cultural environments. Full article

Deepfake content forgery compromises the integrity of digital media and the protection of personal identity, making its detection essential for preserving trust and enabling effective forensic analysis. Most deepfake detection approaches focus on global classification with a binary decision, which is inadequate for precise localization of manipulated regions. This limitation becomes particularly evident under image processing distortions. This paper proposes a dual-decoder architecture for the detection and segmentation of original and deepfake facial manipulations. Unlike conventional single-decoder segmentation models, the proposed approach introduces two decoding branches that learn complementary feature representations of authentic and forgery facial textures. In addition, attention mechanism modules are incorporated to refine encoder features based on decoder context, introducing adaptive feature selection during reconstruction. This architectural design reduces feature interference during reconstruction and enhances the localization of subtle inconsistencies introduced by deepfake manipulations. This approach generates complementary masks for real and forged regions, providing more precise boundary delineation. Experimental results highlight the robustness of the proposed method under image processing distortions, achieving intersection over union (IoU) scores of 0.9387 for real faces and 0.9254 for deepfake segmentation. These results underscore the effectiveness of the dual-decoder architecture in accurately detecting and localizing deepfake facial manipulations. Full article

The increasing complexity of next-generation wireless systems challenges the scalability and generalization capabilities of traditional model-driven physical layer (PHY) design, which relies on analytically derived channel models and optimization frameworks. This paper presents a comprehensive survey and critical review of deep learning (DL) architectures enabling the transition toward AI-native PHY design. A unified optimization perspective is developed in which all PHY tasks—including channel estimation, channel state information (CSI) feedback, massive MIMO processing, signal detection, channel coding, beamforming, resource allocation, and semantic-aware transmission—are formulated under a common empirical risk minimization (ERM) framework. Neural architectures such as autoencoders, convolutional and recurrent networks, transformers, and reinforcement learning models are examined through their underlying optimization formulations, loss functions, training methodologies, and representation learning mechanisms. The review compares model-driven and AI-native approaches in terms of performance metrics, computational complexity, robustness, generalization capability, and practical deployment constraints, including hardware limitations, energy efficiency, and real-time feasibility. The analysis highlights the conditions under which AI-native architectures provide adaptability and performance improvements while identifying trade-offs in complexity, latency, and interpretability. The study concludes by outlining prioritized research directions toward fully adaptive and self-optimizing wireless communication systems. Full article

The Industrial Internet of Things (IIoT) has become a cornerstone of modern industrial automation, enabling real-time monitoring, intelligent decision-making, and large-scale connectivity across cyber–physical systems. However, the growing scale, heterogeneity, and dynamic behavior of IIoT environments significantly expand the attack surface and challenge the effectiveness of conventional security mechanisms. In this paper, we propose DeepSense, a hybrid and adaptive anomaly and intrusion detection framework specifically designed for resource-constrained and heterogeneous IIoT deployments. DeepSense integrates three complementary components: DataSense, a realistic data pipeline and experimental testbed supporting synchronized sensor and network data processing; RuleSense, a lightweight rule-based detection layer that provides fast, deterministic, and interpretable anomaly screening at the edge; and NeuroSense, a learning-driven detection module comprising an adaptive ensemble of 22 machine learning and deep learning models spanning classical, neural, hybrid, and Transformer-based architectures. NeuroSense operates as a second detection stage that validates suspicious events flagged by RuleSense and enables both coarse-grained and fine-grained attack classification. To support rigorous and practical assessment, this work further introduces a comprehensive performance evaluation framework that extends beyond accuracy-centric metrics by jointly considering detection quality, latency, resource efficiency, and detection coverage, alongside an optimization-based process for selecting Pareto-optimal model ensembles under realistic IIoT constraints. Extensive experiments across diverse detection scenarios demonstrate that DeepSense exhibits strong generalization, lower false positive rates, and robust performance under evolving attack behaviors. The proposed framework provides a scalable and efficient IIoT security solution that meets the operational requirements of Industry 4.0 and the resilience-oriented objectives of Industry 5.0. Full article

This work explores the application of Spiking Neural Networks (SNNs) and Continual Learning (CL) methodologies to the problem of steering angle regression, using autonomous driving simulation as the experimental context, with a focus on energy efficiency and alignment with sustainable computing objectives. The primary goal was to design and implement CL techniques in SNNs to assess the model’s ability to maintain accuracy in explored environments while reducing CO${}_2$ emissions through the optimized use of a subset of the data. This study emerges in response to the increasing energy demand of deep learning models, which poses a challenge to sustainability. SNNs, inspired by the efficiency of biological neural systems, offer significant advantages in terms of computational and energy consumption, making them a promising alternative. CL techniques, such as Elastic Weight Consolidation and replay memory, are integrated to mitigate catastrophic forgetting in sequential learning tasks. The methodology includes adapting the PilotNet architecture for SNNs, preprocessing datasets generated in the Udacity driving simulator, and evaluating models in incremental learning scenarios. The experiments compare the performance of SNNs with CL against baseline models without CL, using mean squared error (MSE), computational efficiency, and equivalent CO${}_2$ emissions as evaluation metrics. The results demonstrate that replay memory enables the retention of prior knowledge with a limited increase in energy consumption. This work concludes that SNNs with CL are a viable alternative for sustainable AI applications. Future research directions include a focus primarily on hardware-specific implementations and real-world testing. Full article

In order to protect the high-sensitivity optical lens of the “magnetic field and velocity field imager” in extreme deep space environments, this paper proposes a new type of dual redundant planetary differential lens cover drive mechanism. In view of the critical vulnerability that traditional single-motor direct drive is prone to sudden mechanical jamming and catastrophic single-point failure (SPF) in severe tasks such as Jupiter exploration, this study constructs a “dual input single output (DISO)” rigid decoupling architecture from the perspective of physical topology. Through theoretical analysis and kinematic modeling, the adaptive decoupling mechanism of the two-degree-of-freedom (2-DOF) system under unilateral mechanical stalling is revealed. Dynamic analysis shows that in the nominal dual-motor synergy mode, the system shows a significant “kinematic load-sharing effect”, thus greatly reducing the sliding friction and gear wear rate. In addition, under the severe dynamic fault injection scenario (maximum gravity deviation and sudden jam superposition of a single motor), the cold standby motor is activated and the dynamic takeover is quickly performed. The high-fidelity transient simulation based on ADAMS verifies that although the fault will produce transient global torque spikes and pulsed internal gear contact forces at the moment, all extreme dynamic loads remain well within the structural safety margin. The output successfully achieved a smooth transition, which is characterized by a non-zero-crossing velocity recovery. This research provides an innovative theoretical basis and a practical engineering paradigm for the design of high-reliability fault-tolerant mechanisms in deep space exploration. Full article

This study proposes a robust data-driven control framework, LSTM-MPC, designed to enhance the velocity stabilization of Autonomous Underwater Vehicles (AUVs) operating under stochastic marine disturbances. Traditional control methods often struggle with the highly nonlinear and time-varying hydrodynamics of irregular waves. To address this, we employ a Long Short-Term Memory (LSTM) recurrent neural network to capture complex temporal dependencies and provide accurate multi-step-ahead velocity predictions. These predictions are integrated into a Model Predictive Control (MPC) scheme, which optimizes control actions while respecting actuator constraints. A key contribution is the integration of an error-triggered online learning mechanism. Utilizing run-time weight synchronization via MATLAB Coder, the framework dynamically adapts to plant mismatches and high-frequency MEMS noise without an explicit analytical model. The architecture was validated using experimental data from a Pixhawk/ArduSub baseline. Results demonstrate that, under these stochastic conditions, the data-driven approach significantly outperforms the standard PID-based baseline. While adaptive PID variants offer improvements, the suggested framework drastically reduces tracking errors in rotational axes while maintaining high precision in translational velocities. This research confirms that adaptive, data-driven strategies can effectively bridge the gap between simulation and real-world deployment, offering a scalable solution for robust AUV autonomy in unpredictable environments. Full article

Accurate electric vehicle (EV) charging load forecasting is essential for grid planning and resource allocation, yet existing approaches struggle with the inherent sparsity of charging data—a phenomenon characterized by excessive zeros representing periods of no charging activity. This paper addresses this challenge through a novel framework combining a Zero-Inflated Neural Network (ZINN) architecture with an Evolutionary Neural Architecture Search (ENAS) algorithm. ZINN explicitly decomposes the forecasting problem into binary classification (predicting charging occurrence) and regression (estimating energy magnitude conditioned on occurrence), enabling the model to learn distinct patterns for the absence and presence of charging events. Rather than relying on manually designed architectures, ENAS automatically discovers optimal encoder and decoder configurations from a comprehensive search space encompassing modern architectures (LSTM, GRU, Transformer, and iTransformer), layer configurations, activation functions, and hyperparameters. The evolutionary algorithm balances prediction accuracy with computational efficiency through multi-objective optimization. Extensive experiments on real-world EV charging data from 30 stations in Wuhan demonstrate that the ZINN+ENAS framework achieves the lowest prediction error compared to conventional baselines, with the discovered optimal configuration substantially outperforming hand-crafted designs. Comprehensive ablation studies reveal that the asymmetric dual-head architecture and adaptive regularization strategies are critical for handling data sparsity. These findings highlight the importance of explicit zero-inflation modeling and automated architecture discovery for specialized forecasting tasks, providing practitioners with an open-source framework for practical EV charging load prediction. Full article