Search Results (2,275)

High-precision semantic segmentation of remote sensing imagery is crucial in geospatial analysis. It plays an immeasurable role in fields such as urban governance, environmental monitoring, and natural resource management. However, when confronted with complex objects (such as winding roads and dispersed buildings), existing semantic segmentation methods still suffer from inadequate target recognition capabilities and multi-scale representation issues. This paper proposes a neural network model, LMVMamba (LoRA Multi-scale Vision Mamba), for semantic segmentation of remote sensing images. This model integrates the advantages of convolutional neural networks (CNNs), Transformers, and state-space models (Mamba) with a multi-scale feature fusion strategy. It simultaneously captures global contextual information and fine-grained local features. Specifically, in the encoder stage, the ResT Transformer serves as the backbone network, employing a LoRA fine-tuning strategy to effectively enhance model accuracy by training only the introduced low-rank matrix pairs. The extracted features are then passed to the decoder, where a U-shaped Mamba decoder is designed. In this stage, a Multi-Scale Post-processing Block (MPB) is introduced, consisting of depthwise separable convolutions and residual concatenation. This block effectively extracts multi-scale features and enhances local detail extraction after the VSS block. Additionally, a Local Enhancement and Fusion Attention Module (LAS) is added at the end of each decoder block. LAS integrates the SimAM attention mechanism, further enhancing the model’s multi-scale feature fusion capability and local detail segmentation capability. Through extensive comparative experiments, it was found that LMVMamba achieves superior performance on the OpenEarthMap dataset (mIoU 52.3%, OA 69.8%, mF1: 68.0%) and LoveDA (mIoU 67.9%, OA 80.3%, mF1: 80.5%) datasets. Ablation experiments validated the effectiveness of each module. The final results indicate that this model is highly suitable for high-precision land-cover classification tasks in remote sensing imagery. LMVMamba provides an effective solution for precise semantic segmentation of high-resolution remote sensing imagery. Full article

The subject of this paper is an analysis of the use of wall heating and cooling modules with heat pipes for efficient space heating and cooling. The modules under consideration constitute a structural element installed in the room’s partition structure and consist of heat pipes embedded in a several-centimeter layer of concrete. Water-based central heating and chilled water systems were used as the heat and cooling source. The heat pipes are filled with a thermodynamic medium that changes state in repeated gas–liquid cycles. The advantage of this solution is the use of heat pipes as a heating and cooling element built into the wall, instead of a traditional water system. This solution offers many operational benefits, such as reduced costs for pumping the heat medium. This paper presents an analysis of the potential of using heat pipe modules for heating and cooling in real-world buildings in Poland. Taking into account the structural characteristics of the rooms under consideration (i.e., internal wall area, window area), an analysis was conducted to determine the potential use of the modules for space heating and cooling. The analysis was based on rooms where, according to the authors, the largest possible use of internal and external wall surfaces is possible, such as hotels and schools. Based on the simulations and calculations, it can be concluded that the modules can be effectively used in Poland as a real heating and cooling element: standalone, covering the entire heating and cooling demand of a room, e.g., a hotel room, or as a component working with an additional system, e.g., air cooling and heating in school buildings. The changes in outdoor air temperature, during the year analyzed in the article, were in the range of −24/+32 °C. Full article

Managing resource use in cloud and edge environments is crucial for optimizing performance and efficiency. Traditionally, this process is performed with detailed knowledge of the available infrastructure while being application-specific. However, it is common that users cannot accurately specify their applications’ low-level requirements, and they tend to overestimate them—a problem further intensified by their lack of detailed knowledge on the infrastructure’s characteristics. In this context, resource orchestration mechanisms perform allocations based on the provided worst-case assumptions, with a direct impact on the performance of the whole infrastructure. In this work, we propose a resource orchestration mechanism based on intents, in which users provide their high-level workload requirements by specifying their intended preferences for how the workload should be managed, such as prioritizing high capacity, low cost, or other criteria. Building on this, the proposed mechanism dynamically assigns resources to applications through a Reinforcement Learning method leveraging the feedback from the users and infrastructure providers’ monitoring system. We formulate the respective problem as a discrete-time, finite horizon Markov decision process. Initially, we solve the problem using a tabular Q-learning method. However, due to the large state space inherent in real-world scenarios, we also employ Deep Reinforcement Learning, utilizing a neural network for the Q-value approximation. The presented mechanism is capable of continuously adapting the manner in which resources are allocated based on feedback from users and infrastructure providers. A series of simulation experiments were conducted to demonstrate the applicability of the proposed methodologies in intent-based resource allocation, examining various aspects and characteristics and performing comparative analysis. Full article

Traffic crashes remain a critical public safety issue and are among the leading causes of mortality worldwide. Understanding, analyzing, and forecasting crash trends are essential for implementing effective countermeasures and reducing injury severity. In response to the growing number of crashes and their associated economic and social costs, this study presents a geospatial analytical framework for prioritizing and classifying roadway segments based on crash trends. The framework focuses on a major freeway corridor in the United States, covering a four-year period across 20 counties. This methodology employs spatiotemporal analysis, which integrates both spatial (geographic) and temporal (time-based) dimensions to better understand how crash patterns evolve over time and space. A central component of the analysis is Space–Time Cube (STC) modeling, a three-dimensional GIS-based visualization, and an analytical approach that organizes data into spatial locations (x and y) across a sequence of temporal bins (z-axis) to reveal patterns that may not be evident in a two-dimensional analysis. Additionally, emerging pattern analysis, specifically Emerging Hotspot Analysis (EHA), is used to identify statistically significant trends in crash frequency over time. The results indicate a significant spatial clustering of crashes, with high-risk segments predominantly located in densely populated urban areas with high traffic volumes. Crash hotspots were classified into five distinct categories: persistent, intensifying, new, sporadic, and diminishing, enabling transportation agencies to tailor interventions based on temporal dynamics. The proposed geospatial framework enhances decision making for roadway safety improvements and can be adapted for use in other regional corridors to support infrastructure investment and advance public safety. Full article

This paper presents an analytical investigation of the free vibration behavior of fiber metal laminate (FML) beams with three types of boundary conditions, considering the temperature-dependent properties and the interfacial slip. In the proposed model, the non-uniform temperature field is derived based on one-dimensional heat conduction theory using a transfer formulation. Subsequently, based on the two-dimensional elasticity theory, the governing equations are established. Compared with shear deformation theories, the present solution does not rely on a shear deformation assumption, enabling more accurate capture of interlaminar shear effects and higher-order vibration modes. The relationship of stresses and displacements is determined by the differential quadrature method, the state-space method and the transfer matrix method. Since the corresponding matrix is singular due to the absence of external loads, the natural frequencies are determined using the bisection method. The comparison study indicates that the present solutions are consistent with experimental results, and the errors of finite element simulation and the solution based on the first-order shear deformation theory reach 3.81% and 3.96%, respectively. At last, the effects of temperature, the effects of temperature degree, interface bonding and boundary conditions on the vibration performance of the FML beams are investigated in detail. The research results provide support for the design and analysis of FML beams under high-temperature and vibration environments in practical engineering. Full article

Modern societies are increasingly experiencing limited contact with nature, a phenomenon referred to as the “nature deficit.” The article presents a systematic review of the literature on this issue, with particular emphasis on the role of forests in mitigating its effects. The analysis, based on the Scopus and Web of Science databases, synthesizes the current state of knowledge on the consequences of nature deficit for physical, mental, and social health, while also highlighting the potential of forests as spaces supporting human well-being. The review process followed a systematic methodology, using precisely defined keyword combinations and multi-stage screening. From an initial pool of 88 publications, a critical selection process led to 11 articles that met the inclusion criteria and were analyzed in depth. The findings show that regular contact with nature reduces stress, anxiety, and ADHD symptoms, supports cognitive development, and im-proves concentration, creativity, and social skills. At the same time, there is a lack of consistent tools for clearly diagnosing nature deficit, and existing studies face significant methodological limitations (small samples, subjective measurements, lack of laboratory control). The article also identifies research gaps, particularly in the context of sustainable forest management, cultural differences, and the long-term health effects of exposure to nature. Full article

The rapid advancement of kinetic pursuit technologies has significantly increased the difficulty of evasion for high-speed UAVs (HSUAVs), particularly in scenarios where two collaboratively operating pursuers approach from the same direction with optimized initial space intervals. This paper begins by deriving an optimal initial space interval to enhance cooperative pursuit effectiveness and introduces an evasion difficulty classification framework, thereby providing a structured approach for evaluating and optimizing evasion strategies. Based on this, an intelligent maneuver evasion strategy using semantic classification progressive learning with twin delayed deep deterministic policy gradient (SCPL-TD3) is proposed to address the challenging scenarios identified through the analysis. Training efficiency is enhanced by the proposed SCPL-TD3 algorithm through the employment of progressive learning to dynamically adjust training complexity and the integration of semantic classification to guide the learning process via meaningful state-action pattern recognition. Built upon the twin delayed deep deterministic policy gradient framework, the algorithm further enhances both stability and efficiency in complex environments. A specially designed reward function is incorporated to balance evasion performance with mission constraints, ensuring the fulfillment of HSUAV’s operational objectives. Simulation results demonstrate that the proposed approach significantly improves training stability and evasion effectiveness, achieving a 97.04% success rate and a 7.10–14.85% improvement in decision-making speed. Full article

This paper addresses the modeling challenge of significant asymmetry between acceleration and deceleration processes in car-following behavior by proposing an Asymmetric Acceleration and Deceleration Car Following (AAD-CF) model. The model characterizes driving decisions using both desired speed and desired spacing, and incorporates an asymmetric correction factor to capture differences in acceleration and deceleration behavior. Based on real vehicle trajectory data from the I-80 dataset, the model was compared at the microscopic level against classical models such as Gipps in terms of trajectory fitting error. The results show that the AAD-CF model consistently achieves lower trajectory fitting errors across different simulation time-steps, with error reduction exceeding 10%. At the macroscopic traffic flow level, the model successfully reproduced three-phase traffic flow states—free flow, synchronized flow, and wide moving jams. By implementing both startup and emergency braking scenarios, it was further revealed that braking waves propagate approximately 40% faster than startup waves, demonstrating asymmetric wave propagation. This study provides quantitative evidence for understanding the intrinsic relationship between microscopic driving behavior and macroscopic traffic phenomena, and the proposed model can support traffic simulation systems and theoretical analysis. Full article

Background: There is an emerging need for new scalable behavioral assays, i.e., assays that are feasible to administer from the comfort of the person’s home, with ease and at higher frequency than clinical visits or visits to laboratory settings can afford us today. This need poses several challenges which we address in this work along with scalable solutions for behavioral data acquisition and analyses aimed at diversifying various populations under study here and to encourage citizen-driven participatory models of research and clinical practices. Methods: Our methods are centered on the biophysical fluctuations unique to the person and on the characterization of behavioral states using standardized biorhythmic time series data (from kinematic, electrocardiographic, voice, and video-based tools) in naturalistic settings, outside a laboratory environment. The methods are illustrated with three representative studies (58 participants, 8–70 years old, 34 males, 24 females). Data is presented across the nervous systems under a proposed functional taxonomy that permits data organization according to nervous systems’ maturation and decline levels. These methods can be applied to various research programs ranging from clinical trials at home, to remote pedagogical settings. They are aimed at creating new standardized biometric scales to screen and diagnose neurological disorders across the human lifespan. Results: Using this remote data collection system under our new unifying statistical platform for individualized behavioral analysis, we characterize the digital ranges of biophysical signals of neurotypical participants and report departure from normative ranges in neurodevelopmental and neurodegenerative disorders. Each study provides parameter spaces with self-emerging clusters whereby data points corresponding to a cluster are probability distribution parameters automatically classifying participants into different continuous Gamma probability distribution families. Non-parametric analysis reveals significant differences in distributions’ shape and scale (p < 0.01). Data reduction is realizable from full probability distribution families to a single parameter, the Gamma scale, amenable to represent each participant within each subclass, and each cluster of similar participants within each cohort. We report on data integration from stochastic analyses that serve to differentiate participants and propose new ways to highly scale our research, education, and clinical practices. Conclusions: This work highlights important methodological and analytical techniques for developing personalized and scalable biometrics across various populations outside a laboratory setting. Full article

This research aims to actively suppress vibrations at the end-effector of a flexible manipulator. When configured in a locked state, the system behaves as a two-link manipulator subjected to disturbances on the first link. To analyze its behavior, Finite Element Analysis (FEA) is employed to extract the natural frequencies (eigenvalues) and corresponding mode shapes (eigenvectors) of a two-link, two-joint flexible manipulator (2L2JM). The obtained eigenvectors are transformed into uncoupled state-space equations using balanced realization and the Match-DC-Gain model reduction algorithm. An H-infinity controller is then designed and applied to both the full-order and reduced-order models of the manipulator. The objective of this study is to validate an analytical framework through FEA, demonstrating its applicability to complex manipulators with multiple joints and flexible links. Given that the full state-space representation typically results in high-dimensional matrices, model reduction enables effective vibration control with a minimal number of states. The derivation of the 2L2JM state space, its model reduction, and a subsequent control strategy have not been previously addressed in this manner. Simulation results showcasing vibration suppression of a cantilever beam are presented and benchmarked against two alternative modeling approaches. Full article

Classical spectral analysis characterizes linear systems effectively but often fails to reveal the nonlinear temporal structure of chaotic dynamics. We introduce the ordinal spectrum, a frequency-domain characterization derived from the ordinal-pattern representation of a time series. Applied to both synthetic and real-world datasets—including periodic, stochastic, and chaotic signals from physical, biological, and astronomical sources—the ordinal spectrum identifies the temporal scales implied in a possible chaotic behavior. By providing an interpretable, data-driven view of symbolic dynamics in the frequency domain, this approach complements state–space reconstructions and enhances the detection of nonlinear temporal organization that classical spectra may obscure. Its ability to distinguish between qualitatively different dynamics make it a useful tool for exploring complex time series across diverse scientific domains. Full article

The increasing integration of renewable energy sources (RES) and rising energy demand have created challenges in maintaining stability in interconnected power systems, particularly in terms of frequency, voltage, and tie-line power. While traditional load frequency control (LFC) and automatic voltage regulation (AVR) strategies have been widely studied, they often fail to address the complexities introduced by RES and nonlinear system dynamics such as boiler dynamics, governor deadband, and generation rate constraints. This study introduces the Arithmetic Optimization Algorithm (AOA)-optimized PI(1+DD) controller, chosen for its ability to effectively optimize control parameters in highly nonlinear and dynamic environments. AOA, a novel metaheuristic technique, was selected due to its robustness, efficiency in exploring large search spaces, and ability to converge to optimal solutions even in the presence of complex system dynamics. The proposed controller outperforms classical methods such as PI, PID, I–P, I–PD, and PI–PD in terms of key performance metrics, achieving a settling time of 7.5 s (compared to 10.5 s for PI), overshoot of 2.8% (compared to 5.2% for PI), rise time of 0.7 s (compared to 1.2 s for PI), and steady-state error of 0.05% (compared to 0.3% for PI). Additionally, sensitivity analysis confirms the robustness of the AOA-optimized controller under ±25% variations in turbine and speed control parameters, as well as in the presence of nonlinearities, demonstrating its potential as a reliable solution for improving grid performance in complex, nonlinear multi-area interconnected power systems. Full article

High-performance deep learning models require extensive computational resources and datasets, making their ownership protection a pressing concern. To address this challenge, we focus on advancing model security through robust watermarking mechanisms. In this work, we propose a novel deep neural network watermarking method that embeds ownership information directly within the image feature space. Unlike existing approaches that often suffer from low embedding success rates and significant performance degradation, our method leverages convolutional kernels with orthogonal preferences to extract multiperspective features, which are then linearly mapped at the output layer for watermark embedding. Furthermore, we introduce an orthogonal regularization constraint into the loss function to increase the watermark robustness. This constraint enforces orthogonality in both convolutional and fully connected layer weights, suppresses redundancy in hidden layer representations, and minimizes interference between the watermark and the model’s original feature space. Through these innovations, we significantly improve the embedding reliability and preserve model integrity. Experimental results obtained on ResNet-18 and ResNet-101 demonstrate a 100% watermark detection rate with less than 1% performance impact, underscoring the practical security value of our approach. Comparative analysis further validates that our method achieves superior harmlessness and effectiveness relative to state-of-the-art techniques. These contributions highlight the role of our work in strengthening intellectual property protection and the trustworthy deployment of deep learning models. Full article

Climate change and urbanization are intensifying the urban heat island effect and negatively impacting outdoor thermal comfort in cities. Innovative planning strategies are required to design more livable and resilient urban spaces. Building on a state of the art of current Urban Digital Twins (UDTs) for outdoor thermal comfort analysis, this paper presents the design and implementation of a functional UDT prototype. Developed for a pilot area in Huelva, Spain, the system integrates real-time environmental data, spatial modeling, and simulation tools within an open-source architecture. The literature reveals that while UDTs are increasingly used in urban management, their application to outdoor thermal comfort remains limited and technically challenging, especially in terms of real-time data, modeling accuracy, and user interaction. The case study demonstrates the feasibility of a modular, open-source UDT capable of simulating mean radiant temperature and outdoor thermal comfort indexes at high resolution and visualizing the results in a 3D interactive environment. UDTs have strong potential for supporting microclimate-sensitive planning and improving outdoor thermal comfort. However, important challenges remain, particularly in simulation efficiency, model detail, and stakeholder accessibility. The proposed prototype addresses several of these gaps and provides a basis for future improvements. Full article

Background: As diabetes mellitus continues to escalate into a global health crisis, particularly in China, the limitations of conventional pharmacotherapy underscore the need for complementary interventions. This study systematically reviews two decades of research progress on acupuncture for diabetes management. Methods: A total of 391 publications met the inclusion criteria from the Web of Science Core Collection (2004–2024) using the search terms “acupuncture” AND “diabetes”. These comprised 294 original studies and 97 reviews. CiteSpace 6.3.R1 was used to perform multidimensional analyses, including co-occurrence networks, centrality algorithms, and silhouette metrics across countries/regions, institutions, authors, journals, references, and keywords. Results: The analysis shows a significant increase in publications on acupuncture for diabetes management after 2013. China and the United States lead in research output, yet collaboration between the two countries remains limited. Most researchers currently work within isolated clusters, underscoring the need for greater exchanges and cooperation. Furthermore, this study identified three key research hotspots: insulin resistance, complications, and interdisciplinary research. Conclusions: This bibliometric analysis reveals dynamic growth patterns and paradigm shifts in acupuncture and diabetes research. The findings provide valuable implications for integrating acupuncture into diabetes treatment. Full article