Search Results (356)

This article presents the design, development, and evaluation of an AI-based assistant tailored to support users in the application of the Specific Operations Risk Assessment (SORA) methodology for unmanned aircraft systems. Built on a customized language model, the assistant was trained using system-level instructions with the goal of translating complex regulatory concepts into clear and actionable guidance. The approach combines structured definitions, contextualized examples, constrained response behavior, and references to authoritative sources such as JARUS and EASA. Rather than substituting expert or regulatory roles, the assistant provides process-oriented support, helping users understand and complete the various stages of risk assessment. The model’s effectiveness is illustrated through practical interaction scenarios, demonstrating its value across educational, operational, and advisory use cases, and its potential to contribute to the digital transformation of safety and compliance processes in the drone ecosystem. Full article

Although Unmanned Aircraft Systems (UASs) offer valuable services, they also introduce certain risks—particularly to individuals on the ground—referred to as third-party risk (TPR). In general, ground-level TPR tends to rise alongside the density of people who might use these services, leading current regulations to heavily restrict UAS operations in populated regions. These operational constraints hinder the ability to gather safety insights through the conventional method of learning from real-world incidents. To address this, a promising alternative is to use dynamic simulations that model UAS collisions with humans, providing critical data to inform safer UAS design. In the automotive industry, the modelling and simulation of car crashes has been well developed. For small UAS, this dynamical modelling and simulation approach has focused on the effect of the varying weight and kinetic energy of the UAS, as well as the geometry and location of the impact on a human body. The objective of this research is to quantify the effects of UAS material and shape on-ground TPR through dynamical modelling and simulation. To accomplish this objective, five camera–drone types are selected that have similar weights, although they differ in terms of airframe structure and materials. For each of these camera–drones, a dynamical model is developed to simulate impact, with a biomechanical human body model validated for impact. The injury levels and probability of fatality (PoF) results, obtained through conducting simulations with these integrated dynamical models, are significantly different for the camera–drone types. For the uncontrolled vertical impact of a 1.2 kg UAS at 18 m/s on a model of a human head, differences in UAS designs even yield an order in magnitude difference in PoF values. Moreover, the highest PoF value is a factor of 2 lower than the parametric PoF models used in standing regulation. In the same scenario for UAS types with a weight of 0.4 kg, differences in UAS designs even considered yield an order when regarding the magnitude difference in PoF values. These findings confirm that the material and shape design of a UAS plays an important role in reducing ground TPR, and that these effects can be addressed by using dynamical modelling and simulation during UAS design. Full article

A primary objective of this research entails the development of an innovative algorithm capable of tracking a drone in real-time. This objective serves as a fundamental requirement across various applications, including collision avoidance, formation flying, and the interception of moving targets. Nonetheless, regardless of the efficacy of any detection algorithm, achieving 100% performance remains unattainable. Deep neural networks (DNNs) were employed to enhance this performance. To facilitate real-time operation, the DNN must be executed within a Deep Learning Processing Unit (DPU), Neural Processing Unit (NPU), Tensor Processing Unit (TPU), or Graphics Processing Unit (GPU) system on board the UAV. Given the constraints of these processing units, it may be necessary to quantify the DNN or utilize a less complex variant, resulting in an additional reduction in performance. However, precise target detection at each control step is imperative for effective flight path control. By integrating multiple algorithms, the developed system can effectively track UAVs with improved detection performance. Furthermore, this paper aims to establish a versatile Unmanned Aerial System (UAS) development platform constructed using open-source components and possessing the capability to adapt and evolve seamlessly throughout the development and post-production phases. Full article

Ransomware, a significant cybersecurity threat, encrypts files and causes substantial damage, making early detection crucial yet challenging. This paper introduces a novel multi-phase framework for early ransomware detection, designed to enhance accuracy and minimize false positives. The framework addresses the limitations of existing methods by integrating operational data with situational and threat intelligence, enabling it to dynamically adapt to the evolving ransomware landscape. Key innovations include (1) data augmentation using a Bi-Gradual Minimax Generative Adversarial Network (BGM-GAN) to generate synthetic ransomware attack patterns, addressing data insufficiency; (2) Incremental Mutual Information Selection (IMIS) for dynamically selecting relevant features, adapting to evolving ransomware behaviors and reducing computational overhead; and (3) a Deep Belief Network (DBN) detection architecture, trained on the augmented data and optimized with Uncertainty-Aware Dynamic Early Stopping (UA-DES) to prevent overfitting. The model demonstrates a 4% improvement in detection accuracy (from 90% to 94%) through synthetic data generation and reduces false positives from 15.4% to 14%. The IMIS technique further increases accuracy to 96% while reducing false positives. The UA-DES optimization boosts accuracy to 98.6% and lowers false positives to 10%. Overall, this framework effectively addresses the challenges posed by evolving ransomware, significantly enhancing detection accuracy and reliability. Full article

As the integration of renewable energy sources into microgrid operations deepens, their inherent uncertainty poses significant challenges for dispatch scheduling. This paper proposes a hybrid game-theoretic optimization strategy to address the uncertainty of renewable energy in microgrid scheduling. An energy trading framework is developed, involving integrated energy microgrids (IEMS), shared energy storage operators (ESOS), and user aggregators (UAS). A mixed game model combining master–slave and cooperative game theory is constructed in which the ESO acts as the leader by setting electricity prices to maximize its own profit, while guiding the IEMs and UAs—as followers—to optimize their respective operations. Cooperative decisions within the IEM coalition are coordinated using Nash bargaining theory. To enhance the generality of the user aggregator model, both electric vehicle (EV) users and demand response (DR) users are considered. Additionally, the model incorporates renewable energy output uncertainty through distributionally robust chance constraints (DRCCs). The resulting two-level optimization problem is solved using Karush–Kuhn–Tucker (KKT) conditions and the Alternating Direction Method of Multipliers (ADMM). Simulation results verify the effectiveness and robustness of the proposed model in enhancing operational efficiency under conditions of uncertainty. Full article

Supercritical carbon dioxide (s-CO₂) Brayton cycles have emerged as a promising technology for high-efficiency power generation, owing to their compact architecture and favorable thermophysical properties. However, their performance degrades significantly under cold-climate conditions—such as those encountered in Greenland, Russia, Canada, Scandinavia, and Alaska—due to the proximity to the fluid’s critical point. This study investigates the behavior of the recompression Brayton cycle (RBC) under subzero ambient temperatures through the incorporation of low-critical-temperature additives to create CO₂-based binary mixtures. The working fluids examined include methane (CH₄), tetrafluoromethane (CF₄), nitrogen trifluoride (NF₃), and krypton (Kr). Simulation results show that CH₄- and CF₄-rich mixtures can achieve thermal efficiency improvements of up to 10 percentage points over pure CO₂. NF₃-containing blends yield solid performance in moderately cold environments, while Kr-based mixtures provide modest but consistent efficiency gains. At low compressor inlet temperatures, the high-temperature recuperator (HTR) becomes the dominant performance-limiting component. Optimal distribution of recuperator conductance (UA) favors increased HTR sizing when mixtures are employed, ensuring effective heat recovery across larger temperature differentials. The study concludes with a comparative exergy analysis between pure CO₂ and mixture-based cycles in RBC architecture. The findings highlight the potential of custom-tailored working fluids to enhance thermodynamic performance and operational stability of s-CO₂ power systems under cold-climate conditions. Full article

This study presents a cooperative unmanned aerial system (UAS) designed to enable precise autonomous landings in unstructured environments using low-cost onboard vision technology. This approach involves a carrier UAV with a stabilized RGB camera and a neural inference system, as well as a lightweight tailsitter payload UAV with an embedded grayscale vision module. The system relies on visually recognizable landing markers and does not require additional sensors. Field trials comprising full deployments achieved an 80% success rate in autonomous landings, with vertical touchdown occurring within a 1.5 m radius of the target. These results confirm that vision-based marker detection using compact neural models can effectively support autonomous UAV operations in constrained conditions. This architecture offers a scalable alternative to the high complexity of SLAM or terrain-mapping systems. Full article

We aimed to simulate tau abnormalities—specifically hyperphosphorylation and aggregation—that are hallmarks of tauopathies, including Alzheimer’s disease, to evaluate tau-targeting therapies. To model pathological p-tau accumulation at early disease stages, we exposed mouse cortical cultures to redox-active iron from hemin (Hm), a breakdown product of hemoglobin, or challenged them with the excitatory neurotransmitter glutamate. Using the AT8 phospho-specific antibody, we demonstrate that a subtoxic concentration of Hm (3 µM) promotes pathological p-tau accumulation in a subpopulation of cultured cortical neurons and their proximal neurites. Uric acid (UA; 0.1–200 µM), the metabolic end-product of purines in humans, prevented p-tau build-up. Neither xanthine, the immediate precursor of UA, nor allantoin, its oxidized product, reproduced this effect. Live cell imaging studies revealed that UA operates by repressing iron-driven lipid peroxidation. DOT (3 µM), a brain-permeant tetracycline (TC) without antibiotic activity, mimicked UA’s anti-tau and antioxidant effects. Interestingly, both UA and DOT remained effective in preventing p-tau accumulation induced by glutamate (10 µM). To simulate tau aggregation at more advanced disease stages, we conducted a Thioflavin-T aggregation assay. Our findings revealed that UA and DOT prevented tau aggregation seeded by heparin. However, only DOT remained effective when heparin-assembled tau fibrils were used as the seeding material. In summary, our results indicate that UA-elevating agents may hold therapeutic utility for tauopathies. The non-purine compound DOT could serve as an effective alternative to UA-related therapies. Full article

This study shows a comprehensive simulation to assess and enhance the throughput capacity of unmanned air system vertiports, one of the most essential elements of urban air mobility ecosystems. The framework integrates dynamic grid-based spatial management, probabilistic mission duration algorithms, and EASA-compliant operational protocols to address the infrastructural and logistical demands of high-density UAS operations. It was focused on two use cases—high-frequency food delivery utilizing small UASs and extended-range package logistics with larger UASs—and the model incorporates adaptive vertiport zoning strategies, segregating operations into dedicated sectors for battery charging, swapping, and cargo handling to enable parallel processing and mitigate congestion. The simulation evaluates critical variables such as vertiport dimensions, UAS fleet composition, and mission duration ranges while emphasizing scalability, safety, and compliance with evolving regulatory standards. By examining the interplay between infrastructure design, operational workflows, and resource allocation, the research provides a versatile tool for urban planners and policymakers to optimize vertiport layouts and traffic management protocols. Its modular architecture supports future extensions. This work underscores the necessity of adaptive, data-driven planning to harmonize vertiport functionality with the dynamic demands of urban air mobility, ensuring interoperability, safety, and long-term scalability. Full article

SmallUnmanned Aerial Systems (sUAS) have seen rapid adoption thanks to advances in endurance, communications, autonomy, and manufacturing costs, yet most testing remains focused on GPS-supported, above-ground operations. This study introduces new test methodologies and presents comprehensive experimental evaluations of collision tolerance, navigation, and trajectory following for commercial sUAS platforms in GPS-denied indoor environments. We also propose numerical and categorical metrics—based on established vehicle collision protocols such as the Modified Acceleration Severity Index (MASI) and Maximum Delta V (MDV)—to quantify collision resilience; for example, the tested platforms achieved an average MASI of 0.1 g, while demonstrating clear separation between the highest- and lowest-performing systems. The experimental results revealed that performance varied significantly with mission complexity, obstacle proximity, and trajectory requirements, identifying platforms best suited for subterranean or crowded indoor applications. By aggregating these metrics, users can select the optimal drone for their specific mission requirements in challenging enclosed spaces. Full article

Accurate vehicle type recognition in low-light environments remains a critical challenge for intelligent transportation systems (ITSs). To address the performance degradation caused by insufficient lighting, complex backgrounds, and light interference, this paper proposes a Twin-Stream Feature Fusion Graph Neural Network (TFF-Net) model. The model employs multi-scale convolutional operations combined with an Efficient Channel Attention (ECA) module to extract discriminative local features, while independent convolutional layers capture hierarchical global representations. These features are mapped as nodes to construct fully connected graph structures. Hybrid graph neural networks (GNNs) process the graph structures and model spatial dependencies and semantic associations. TFF-Net enhances the representation of features by fusing local details and global context information from the output of GNNs. To further improve its robustness, we propose an Adaptive Weighted Fusion-Bagging (AWF-Bagging) algorithm, which dynamically assigns weights to base classifiers based on their F1 scores. TFF-Net also includes dynamic feature weighting and label smoothing techniques for solving the category imbalance problem. Finally, the proposed TFF-Net is integrated into YOLOv11n (a lightweight real-time object detector) with an improved adaptive loss function. For experimental validation in low-light scenarios, we constructed the low-light vehicle dataset VDD-Light based on the public dataset UA-DETRAC. Experimental results demonstrate that our model achieves 2.6% and 2.2% improvements in mAP50 and mAP50-95 metrics over the baseline model. Compared to mainstream models and methods, the proposed model shows excellent performance and practical deployment potential. Full article

In this paper, we propose and prove an innovative radar antenna concept suitable for collision avoidance (CA) systems installed onboard small, unmanned aircraft (UA). The proposed architecture provides 360° monopulse coverage around the host platform, enabling the detection and accurate position estimation of airborne, non-cooperative intruders using lightweight, low-profile antennas. These antennas can be manufactured using low-cost 3D printing techniques and are easily integrated into the UA airframe without compromising airworthiness. We present a Detect and Avoid (DAA) concept of operations (ConOps) aligned with the SESAR U-space ConOps, Edition 4. In this ConOps, the Remain Well Clear (RWC) and CA functions are treated separately: RWC is the responsibility of ground-based U-space services, while CA is implemented as an airborne safety net using onboard equipment. Based on this framework, we derive operation-centric design requirements and propose an antenna architecture based on a fixed circular array of sector waveguides. This solution overcomes key limitations of existing radar antennas for UAS CA systems by providing a wider field of view, higher power handling, and reduced mechanical complexity and cost. We prove the proposed concept through a combination of simulations and measurements conducted in an anechoic chamber using a 24 GHz prototype. Full article

Forested wetlands in temperate mountain ecosystems play a critical role in carbon sequestration and biodiversity maintenance, yet their accurate delineation remains challenging due to spectral similarity with forests and anthropogenic interference. Here, we present an optimized two-stage Random Forest framework integrating 2019–2022 growing season datasets from Sentinel-1 C-SAR, ALOS-2 L-PALSAR, Sentinel-2 MSI, and Landsat-8 TIRS with environmental covariates. The methodology first applied NDBI thresholding (NDBI > 0.12) to exclude 94% of urban/agricultural areas through spectral masking, then implemented an optimized Random Forest classifier (ntree = 1200, mtry = 28) with 10-fold cross-validation, leveraging 42 features including L-band HV backscatter (feature importance = 47), Sentinel-2 SWIR (Band12; importance = 57), and land surface temperature gradients. This study pioneers a 10 m resolution forest swamp map in the Changbai Mountain wetlands, achieving 87.18% overall accuracy (Kappa = 0.84) with strong predictive performance (AUC = 0.89). Forest swamps showed robust classification metrics (PA = 80.37%, UA = 86.87%), driven by L-band SAR’s superior discriminative power (p < 0.05). Quantitative assessment demonstrated that L-band SAR increased classification accuracy in canopy penetration scenarios by 4.2% compared to optical-only approaches, while thermal-IR features reduced confusion with forests. Forested swamps occupied 229.95 km² (9% of protected areas), predominantly in transitional ecotones (720–850 m elevation) between herbaceous wetlands and forest. This study establishes that multi-sensor fusion enables operational wetland monitoring in topographically complex regions, providing a transferable framework for temperate mountain ecosystems. The dataset advances precision conservation strategies for these climate-sensitive habitats, supporting sustainable development goals targets for wetland protection through enhanced machine learning interpretability and anthropogenic interference mitigation. Full article

This study develops and implements a scalable system architecture for dynamic data acquisition and knowledge modeling in industrial contexts. The objective is to efficiently process large datasets to support decision-making and process optimization within Industry 4.0. The architecture integrates modern technologies, such as the ibaPDA system for data acquisition, and employs communication standards like Modbus TCP and OPC UA to ensure broad compatibility with diverse equipment. In addition, it leverages native protocols offered by certain controllers, enabling direct data exchange without the need for conversion layers. A developed prototype demonstrates the practical applicability of the architecture, tested in a real industrial environment with a focus on processing speed, data integrity, and system reliability. The results indicate that the architecture not only meets the requirements for dynamic data acquisition but also enhances knowledge modeling. This leads to more efficient process control and opens new perspectives for managing and analyzing big data in production environments. The study emphasizes the importance of an integrated development approach and highlights the need for interdisciplinary collaboration to address operational challenges. Future extensions may include the implementation of Python interfaces and machine learning algorithms for data simulation, enabling more accurate predictive models. These findings provide valuable insights for industry, software development, data science, and academia, helping to tackle the challenges of Industry 4.0 and drive innovation forward. Full article

Background/Objectives: This study examines trends in appendectomy utilization in US emergency departments (EDs) from 2012 to 2021 using National Emergency Department Sample (NEDS) data. The objective is to explore appendectomy frequency, appendicitis management, disease progression, and resource distribution in EDs. A predictive model was developed to forecast trends from 2022 to 2032, aiming to improve patient outcomes and support operational planning in EDs. Methods: A cross-sectional analysis was conducted using NEDS data from 2012 to 2021. Appendectomy trends were assessed in four ways: first, comparing the total number of appendectomies with total ED visits to determine relative frequencies; second, comparing trends in Complicated Appendicitis (CA) and Uncomplicated Appendicitis (UA) patients; third, categorizing each appendicitis type based on clinical complications and comorbidities; and finally, using a linear regression model to predict trends through 2032. Results: During the study period, the overall appendectomy rate decreased, while the proportion of patients with Complicated Appendicitis rose. Appendectomies in patients without complications or comorbidities showed a decline, while those in patients with complications or comorbidities increased. Predictive modeling suggests that trends in all subgroups will continue to rise until 2032. Conclusions: This study highlights evolving appendicitis management trends in EDs. The results advocate for fast-track appendectomy pathways and better resource allocation to enhance efficiency, reduce complications, and improve patient care. These findings assist healthcare systems in preparing for ED throughput challenges and refining surgical management strategies. Full article