Search Results (28,902)

Robust condition monitoring of wind turbine blades is essential for reducing downtime and maintenance costs, particularly under variable operating conditions. While recent studies suggest that combining trend monitoring (TM) with change point detection (CPD) can improve diagnostic performance, it remains unclear whether such integration is beneficial for all fault types. This study experimentally evaluates the integration of TM and CPD using vibration data from a laboratory-scale wind turbine for two representative blade faults: leading-edge erosion and twist misalignment. For the erosion case, discrete wavelet transform (DWT) energy features exhibit a clear and persistent increase in mid-frequency content, with energy deviations of approximately 34–45% relative to the healthy state. However, Bayesian Online Change Point Detection (BOCPD) does not reveal distinct change points, indicating that CPD provides limited additional value for gradual, steady-state degradation. In contrast, for twist misalignment, the short-time Fast Fourier Transform (FFT) features reveal dynamic spectral redistribution, and CPD applied to spectral centroid trends produces a sharp, localized detection signature. These results demonstrate that integrating TM with CPD significantly enhances fault detectability for dynamic, instability-driven faults, while TM alone is sufficient for smooth, steady-state degradation. This study provides an evidence-based guideline for selectively integrating CPD into wind turbine blade condition monitoring systems based on fault physics. Full article

Object detection in Unmanned Aerial Vehicle (UAV) imagery has gained significant traction in applications such as railway inspection and waste management. While emerging end-to-end detectors like DEIM show promise, they often struggle with weak feature responses and spatial misalignment in aerial scenarios. To address these issues, this paper proposes SCA-DEIM, a context-aware real-time detection framework. Specifically, we introduce the Adaptive Spatial and Channel Synergistic Attention (ASCSA) module, which refines existing attention paradigms by transitioning from a static gating mechanism to an active signal amplifier. Unlike traditional designs that impose rigid bounds on feature responses, this improved architecture enhances feature extraction by dynamically boosting the saliency of faint small-target signals amidst complex backgrounds. Furthermore, drawing inspiration from infrared small object detection, we propose the Cross-Stage Partial Shifted Pinwheel Mixed Convolution (CSP-SPMConv). By synergizing asymmetric padding with a spatial shift mechanism, this module effectively aligns receptive fields and enforces cross-channel interaction, thereby resolving feature misalignment and scale fusion issues. Comprehensive experiments on the VisDrone2019 dataset demonstrate that, compared with the baseline model, SCA-DEIM achieves improvements of 1.8% in Average Precision (AP), 2.3% in AP for small objects ($A{P}_s$), and 2.0% in AP for large objects ($A{P}_l$), while maintaining a competitive inference speed. Notably, visualization results under different illumination conditions demonstrate the strong robustness of the model. In addition, further validation on both the UAVVaste and UAVDT datasets confirms that the proposed method effectively enhances the detection performance for small objects. Full article

Reliable recognition of people on water from UAV imagery remains a challenging task due to strong glare, wave-induced distortions, partial submersion, and small visual scale of targets. This study proposes a hybrid method for human detection and position recognition in aquatic environments by integrating the YOLO12 object detector with optical-flow-based motion analysis, Kalman tracking, and BlazePose skeletal estimation. A combined training dataset was formed using four complementary sources, enabling the detector to generalize across heterogeneous maritime and flood-like scenes. YOLO12 demonstrated superior performance compared to earlier You Only Look Once (YOLO) generations, achieving the highest accuracy (mAP@0.5 = 0.95) and the lowest error rates on the test set. The hybrid configuration further improved recognition robustness by reducing false positives and partial detections in conditions of intense reflections and dynamic water motion. Real-time experiments on a Raspberry Pi 5 platform confirmed that the full system operates at 21 FPS, supporting onboard deployment for UAV-based search-and-rescue missions. The presented method improves localization reliability, enhances interpretation of human posture and motion, and facilitates prioritization of rescue actions. These findings highlight the practical applicability of YOLO12-based hybrid pipelines for real-time survivor detection in flood response and maritime safety workflows. Full article

This study investigates the dynamics of carbon flux at the water–air interface during the ecological restoration of eutrophic water bodies. A controlled simulation of the eutrophic aquatic environment was carried out. A series of experiments was established, centered on submerged aquatic plants as key agents for carbon sequestration and enhancement of carbon sink capacity, supplemented by biological manipulation techniques aimed at pollution reduction and algal control. Results show that restoration systems based on submerged plants significantly enhance carbon sequestration, whereas systems relying solely on filter-feeding fish tend to increase the carbon emission burden. The submerged plant-only treatment (HV) exhibited the highest carbon absorption capacity (−72.53 mg·m⁻²·h⁻¹), followed by submerged plant + fish + snail (HSXB) and submerged plant + fish (HSX) treatments. CH₄ emissions were initially higher in the combined biological treatments but were eventually surpassed by the control group as algal cell density increased. Carbon sink potential and CH₄ emissions were strongly correlated with algal cell density and chlorophyll a concentration. While combination treatments (HSX and HSXB) effectively suppressed algal proliferation, the submerged plant-only treatment demonstrated superior nutrient removal efficiency. The findings provide theoretical support for ecologically based management strategies that simultaneously address eutrophication control and carbon sequestration in freshwater ecosystems, contributing to both water quality improvement and climate change mitigation. Full article

Indoor air quality (IAQ) in high-rise residential buildings is an increasing concern, especially in hot and humid climates where prolonged indoor exposure elevates health risks. This study evaluates the performance of Fresh Air Handling Units (FAHUs) using two complementary approaches: (1) real-time sensor data to quantify IAQ conditions and (2) occupant survey responses to capture perceived comfort and pollution indicators. The results show that floor level did not predict satisfaction, even though AQI data revealed clear differences between flats, suggesting perceptions are driven more by sensory cues than by actual pollutant levels. Longer weekday exposure emerged as a stronger predictor of dissatisfaction. These gaps between perceived and measured IAQ highlight the need for improved ventilation scheduling and greater occupant awareness. FAHUs were found to be inefficient, consuming 21–26% of total building energy while lacking pollutant-specific monitoring capabilities. To address these issues, the study recommends the integration of IoT-enabled sensors for real-time pollutant detection, enhanced facade sealing to minimize external infiltration, and the upgrade of filtration systems with HEPA filters and UV purification. Additionally, AI-driven predictive maintenance and automated ventilation optimization through Building Management Systems (BMS) are suggested. These findings offer valuable insights for improving IAQ management in high-rise buildings, with future research focusing on AI-based predictive modeling for dynamic air quality control. Full article

Background/Objectives: HER2-low breast cancer has emerged as a clinically meaningful category that challenges the historical HER2-positive versus HER2-negative classification. Although not defined as a distinct biological subtype, HER2-low tumors exhibit unique clinicopathological features and differential sensitivity to novel antibody–drug conjugates. Accurate identification remains difficult due to limitations in immunohistochemistry performance, inter-observer variability, intratumoral heterogeneity, and dynamic shifts in HER2 expression over time. This review synthesizes current evidence on the biological and clinical characteristics of HER2-low breast cancer and evaluates emerging diagnostic innovations, with emphasis on liquid biopsy approaches and evolving technologies that may enhance diagnostic accuracy and monitoring. Methods: A narrative literature review was conducted, examining tissue-based HER2 testing, liquid biopsy modalities, including circulating tumor cells, circulating nucleic acids, extracellular vesicles, and soluble HER2 extracellular domains, and applications of artificial intelligence (AI) across histopathology and multimodal diagnostic systems. Results: Liquid biopsy technologies offer minimally invasive, real-time assessment of HER2 dynamics and may overcome fundamental limitations of tissue-based assays. However, these platforms require rigorous analytical validation and face regulatory and standardization challenges before widespread clinical adoption. Concurrently, AI-enhanced histopathology and multimodal diagnostic systems improve reproducibility, refine HER2 classification, and enable more accurate prediction of treatment response. Emerging biosensor- and AI-enabled monitoring frameworks further support continuous disease evaluation. Conclusions: HER2-low breast cancer sits at the intersection of evolving pathology and technological innovation. Integrating liquid biopsy platforms with AI-driven diagnostics has the potential to advance precision stratification and guide personalized therapeutic strategies for this expanding patient subgroup. Full article

This paper examines the impact of fintech adoption and financial inclusion on financial stability in selected countries. Using panel data from 30 countries spanning 2011–2024, the study employs an empirical strategy based on Two-Way Fixed Effects, a dynamic two-step System GMM estimator, and Panel Quantile Regression. This multi-method approach captures both average and distributional effects while addressing key econometric challenges, including endogeneity, heteroskedasticity, serial correlation, and cross-sectional dependence. The empirical findings differ across estimation techniques but reveal two consistent patterns: financial inclusion exerts a positive and significant effect on financial stability across all models, whereas the impact of fintech is model-dependent. While fintech appears insignificant under the Two-Way Fixed Effects and Driscoll–Kraay specifications, the System GMM and quantile regression analyses confirm that both fintech and financial inclusion significantly enhance financial stability. Overall, the results show that fintech can boost financial stability, but only when supported by broad financial inclusion and solid institutions. The findings highlight that policymakers must pair the growth of digital finance with clear regulatory standards and programs designed to deepen financial inclusion. Full article

This review examines the impact of data analytics powered by the Internet of Things (IoT), edge computing, and artificial intelligence (AI) on improving energy efficiency in smart environments, with a focus on smart factories, smart cities, and smart territories. Advanced AI, machine learning (ML), and deep learning (DL) techniques enable real-time energy optimization and intelligent decision-making in complex, data-intensive systems. Integrating edge computing reduces latency and improves responsiveness in IoT and Industrial Internet of Things (IIoT) networks, enabling local energy management and reducing grid load. Federated learning further enhances data privacy and efficiency by enabling decentralized model training across distributed smart nodes without exposing sensitive information or personal data. Emerging 5G and 6G technologies provide the necessary bandwidth and speed for seamless data exchange and control across energy-intensive, connected infrastructures. Blockchain increases transparency, security, and trust in energy transactions and decentralized energy trading in smart grids. Together, these technologies support dynamic demand response mechanisms, predictive maintenance, and self-regulating systems, leading to significant improvements in energy sustainability. Case studies of smart cities and industrial ecosystems within Industry 4.0/5.0/6.0 demonstrate measurable reductions in energy consumption and carbon emissions through these synergistic approaches. Despite significant progress, challenges remain in interoperability, scalability, and regulatory frameworks. This review demonstrates that AI-based edge computing, supported by robust connectivity and secure IoT and IIoT architectures, has a transformative potential for creating energy-efficient and sustainable smart environments. Full article

This study investigates the role of family involvement and succession planning in shaping the long-term sustainability of family businesses in Indonesia, with a specific focus on the mediating effect of transgenerational entrepreneurial orientation (TEO). This research responds to calls for integrative models that move beyond examining these factors in isolation. Indonesia offers a unique context due to its dominance of family-controlled firms and informal succession traditions, which shape entrepreneurial value transmission across generations. A quantitative approach was employed using survey data from 210 respondents representing active family businesses in Indonesia. Partial least squares structural equation modeling (PLS-SEM) with SmartPLS 4.0 was used to test reliability, validity, and structural relationships. Additional analyses included HTMT for discriminant validity, CVPAT for predictive relevance, and importance–performance map analysis (IPMA) to identify managerial priorities. The results reveal that family involvement and succession planning both exert significant positive effects on long-term family business sustainability, with TEO playing a mediating role. Family involvement strongly enhances both sustainability and entrepreneurial orientation, while succession planning contributes more indirectly through the development of TEO. The IPMA indicates that family leadership in governance and openness to innovation are highly important but underperforming drivers, suggesting key areas for improvement. The model explains 51.9% of the variance in TEO and 48.6% in long-term sustainability, with significant mediation paths (β = 0.092–0.104, p < 0.05). The cross-sectional design limits causal inference, and the focus on Indonesian firms may constrain generalizability to other cultural contexts. Future research could adopt longitudinal and cross-country comparative designs while also examining the role of digital transformation and generational differences in sustaining family firms. The findings highlight the need for Indonesian family firms to professionalize succession planning while strengthening transgenerational entrepreneurial orientation. Practical steps include formal mentoring, clear successor criteria, and embedding innovation and proactiveness in family governance. This study extends the family business literature by conceptualizing TEO as a dynamic capability that bridges family involvement, succession planning, and sustainability. By integrating these perspectives, it offers a more comprehensive understanding of how family firms can achieve resilience and continuity across generations. Full article

The dynamic characteristics of flotation foam, such as velocity and breakage rate, are critical factors that influence mineral separation efficiency. However, challenges inherent in foam images, including weak textures, severe deformations, and motion blur, present significant technical hurdles for dynamic monitoring. These issues lead to a fundamental conflict between the efficiency and accuracy of traditional feature matching algorithms. This paper introduces a novel progressive framework for dynamic feature matching in flotation foam images, termed “stable extraction, efficient coarse screening, and precise matching.” This framework first employs the Accelerated-KAZE (AKAZE) algorithm to extract robust, scale- and rotation-invariant feature points from a non-linear scale-space, effectively addressing the challenge of weak textures. Subsequently, it innovatively incorporates the Grid-based Motion Statistics (GMS) algorithm to perform efficient coarse screening based on motion consistency, rapidly filtering out a large number of obvious mismatches. Finally, the Progressive Sample and Consensus (PROSAC) algorithm is used for precise matching, eliminating the remaining subtle mismatches through progressive sampling and geometric constraints. This framework enables the precise analysis of dynamic foam characteristics, including displacement, velocity, and breakage rate (enhanced by a robust “foam lifetime” mechanism). Comparative experimental results demonstrate that, compared to ORB-GMS-RANSAC (with a Mean Absolute Error, MAE of 1.20 pixels and a Mean Relative Error, MRE of 9.10%) and ORB-RANSAC (MAE: 3.53 pixels, MRE: 27.36%), the proposed framework achieves significantly lower error rates (MAE: 0.23 pixels, MRE: 2.13%). It exhibits exceptional stability and accuracy, particularly in complex scenarios involving low texture and minor displacements. This research provides a high-precision, high-robustness technical solution for the dynamic monitoring and intelligent control of the flotation process. Full article

Optimising the learning rate is essential for efficient neural network training, but static methods can cause overshooting or undershooting, while adaptive techniques like ADAM often struggle to balance exploration and exploitation. We introduce the Polling Method, an ensemble-based optimisation approach that dynamically selects the most effective learning rate at each step, improving convergence and mitigating issues inherent in traditional optimisation strategies. By evaluating base models with varying learning rates at each epoch, the method adaptively balances exploration and exploitation without being constrained by predefined functions or gradient noise. This study details the theoretical foundation, implementation, and integration of the Polling Method with the ADAM optimiser, demonstrating its effectiveness in Artificial Neural Networks and Bayesian variational inference. The results demonstrate that Polling Method-ADAM reduces absolute error by 50% compared to ADAM alone, while also accelerating convergence. In Bayesian optimisation, it reduces the mean gradient shift from 0.85 to 0.7835 over 500 iterations, indicating improved stability in high-dimensional problems. By introducing adaptive learning rate selection within training, the Polling Method enhances optimisation efficiency while mitigating noise accumulation. This framework provides a computationally efficient, flexible alternative for deep learning applications, offering significant improvements over traditional optimisers and a potential breakthrough in neural network training strategies. Full article

In the context of energy structure transition, Enhanced Geothermal Systems (EGSs) represent a core technology for developing hot dry rock (HDR) resources. However, the ultra-long-term natural recovery patterns of reservoir temperature after heat extraction cessation remain unclear, hindering sustainable lifecycle assessment of the system. This study establishes a dual-well EGS numerical model based on the finite element method to simulate the impact mechanisms of flow rate, injection temperature, initial reservoir temperature, and well spacing on natural reservoir temperature compensation during a 1000-year shut-in period following 40 years of heat extraction. Results indicate that reservoir temperature fails to recover to its initial state after shut-in, with final recovery rates ranging from 60.63% to 89.51% of the initial temperature. Each parameter exerts nonlinear control over recovery: lower flow rates yield higher final recovery temperatures (87.62% at 20 kg/s versus 60.63% at 100 kg/s); increased injection temperature from 10 °C to 70 °C reduces the absolute recovery magnitude from 10.65 °C to 7.05 °C but raises the final recovery rate from 78.16% to 86.07%; higher initial reservoir temperatures from 100 °C to 260 °C significantly enhance absolute recovery temperatures from 79.48 °C to 199.58 °C; reduced well spacing from 500 m to 100 m improves final recovery rates from 72.77% to 89.51%. After shut-in in dual-well EGS, the vertical fracture configuration recovered to 78.16% of the initial temperature, the horizontal fracture to 74.39%, and the no-fracture configuration only to 67.87%. Due to optimal heat flow and thermal compensation efficiency, vertical fractures exhibit the best recovery performance, while the no-fracture configuration shows the worst. This study reveals the dynamic mechanism of heat recovery dominated by heat conduction in surrounding rocks, establishes a long-term temperature recovery evaluation framework for EGS, and provides novel scientific evidence and perspectives for the sustainable development and research of geothermal systems. Full article

Winter jujube harvesting is traditionally labor-intensive, yet declining labor availability and rising costs necessitate robotic automation to maintain agricultural competitiveness. Path planning for robotic arms in orchards faces challenges due to the unstructured, dynamic environment containing densely packed fruits and branches. To overcome the limitations of existing robotic path planning methods, this research proposes BMGA-RRT Connect (BVH-based Multilevel-step Gradient-descent Adaptive RRT), a novel algorithm integrating adaptive multilevel step-sizing, hierarchical Bounding Volume Hierarchy (BVH)-based collision detection, and gradient-descent path smoothing. Initially, an adaptive step-size strategy dynamically adjusts node expansions, optimizing efficiency and avoiding collisions; subsequently, a hierarchical BVH improves collision-detection speed, significantly reducing computational time; finally, gradient-descent smoothing enhances trajectory continuity and path quality. Comprehensive 2D and 3D simulation experiments, dynamic obstacle validations, and real-world winter jujube harvesting trials were conducted to assess algorithm performance. Results showed that BMGA-RRT Connect significantly reduced average computation time to 2.23 s (2D) and 7.12 s (3D), outperforming traditional algorithms in path quality, stability, and robustness. Specifically, BMGA-RRT Connect achieved 100% path planning success and 90% execution success in robotic harvesting tests. These findings demonstrate that BMGA-RRT Connect provides an efficient, stable, and reliable solution for robotic harvesting in complex, unstructured agricultural settings, offering substantial promise for practical deployment in precision agriculture. Full article

The increasing frequency and severity of urban crowd disasters underscore a critical need for intelligent surveillance systems capable of real-time crowd anomaly detection and early warning. While deep learning models such as LSTMs, ConvLSTMs, and Transformers have been applied to video-based crowd anomaly detection, they often face limitations in long-term contextual reasoning, computational efficiency, and interpretability. To address these challenges, this paper proposes HiMeLSTM, a crowd anomaly detection framework built around a hippocampal-inspired memory-enhanced LSTM backbone that integrates Long Short-Term Memory (LSTM) networks with an Episodic Memory Unit (EMU). This hybrid design enables the model to effectively capture both short-term temporal dynamics and long-term contextual patterns essential for understanding complex crowd behavior. We evaluate HiMeLSTM on two publicly available crowd-anomaly benchmark datasets (UCF-Crime and ShanghaiTech Campus) and an in-house CrowdSurge-1K dataset, demonstrating that it consistently outperforms strong baseline architectures, including Vanilla LSTM, ConvLSTM, a lightweight spatial–temporal Transformer, and recent reconstruction-based models such as MemAE and ST-AE. Across these datasets, HiMeLSTM achieves up to 93.5% accuracy, 89.6% anomaly detection rate (ADR), and a 0.89 F1-score, while maintaining computational efficiency suitable for real-time deployment on GPU-equipped edge devices. Unlike many recent approaches that rely on multimodal sensors, optical-flow volumes, or detailed digital twins of the environment, HiMeLSTM operates solely on raw CCTV video streams combined with a simple manually defined zone layout. Furthermore, the hippocampal-inspired EMU provides an interpretable memory retrieval mechanism: by inspecting the retrieved episodes and their att ention weights, operators can understand which past crowd patterns contributed to a given decision. Overall, the proposed framework represents a significant step toward practical and reliable crowd monitoring systems for enhancing public safety in urban environments. Full article

Sustainable language education necessitates scalable, accessible learning environments that foster long-term learner autonomy and reduce educational inequality. While online courses have democratized access to language learning globally, persistent deficiencies in instructor-student interaction and learner engagement compromise their sustainability. The “face effect,” denoting the influence of instructor facial appearance on learning outcomes, remains underexplored as a resource-efficient mechanism for enhancing engagement in digital environments. Furthermore, effective measures linking psychological engagement to sustained learning experiences are notably absent. This study addresses three research questions within a sustainable education framework: (1) How does instructor identity, particularly facial appearance, affect second language learners’ outcomes and interactivity in scalable online environments? (2) How can digital human technology dynamically personalize instructor appearance to support diverse learner populations in resource-efficient ways? (3) How does instructor identity influence learners’ flow state, a critical indicator of intrinsic motivation and self-directed learning capacity? Two controlled experiments with Japanese language learners examined three instructor identity conditions: real teacher identity, learner self-identity, and idol-inspired identity. Results demonstrated that the self-identity condition significantly enhanced oral performance and flow state dimensions, particularly concentration and weakened self-awareness. These findings indicate that identity-adaptive digital human instructors cultivate intrinsic motivation and learner autonomy, which are essential competencies for lifelong learning. This research advances Sustainable Development Goal 4 (Quality Education) by demonstrating that adaptive educational technology can simultaneously improve learning outcomes and psychological engagement in scalable, cost-effective online environments. The personalization capabilities of digital human instructors provide a sustainable pathway to reduce educational disparities while maintaining high-quality, engaging instruction accessible to diverse global populations. Full article