Abstract: Sample-wise detection of P-, R-, and T-peaks in electrocardiograms (ECGs) is challenging because each peak type is sparsely represented (≈1:500 samples in a typical 10-s, 500-Hz ECG at 60 bpm), such that even a small number of false-positives (FPs) can markedly degrade positive predictive value (PPV) and limit the practicality of classifier-only approaches. This study proposes a lightweight ECG peak detection framework that combines binary classifiers with physiological temporal constraints (PTC) to address extreme sample-level class imbalance. Local morphological features are first evaluated using lightweight machine-learning models, among which XGBoost (XGB) exhibited the most stable score-ranking performance. Rather than directly thresholding classifier outputs, prediction scores are interpreted within the framework, which encodes physiological timing relationships. R-peaks are detected using score ranking combined with a refractory-period constraint, and the detected R-peaks serve as temporal landmarks for subsequent P- and T-peak detection within physiologically plausible time windows reflecting the P–QRS–T sequence. Quantitative evaluation was conducted using the Lobachevsky University Electrocardiography Database, hereafter referred to as LUDB. With a temporal tolerance of ±20 ms, the XGB-based system achieved an F1-score of 0.87 for R-peak detection (sensitivity 0.96, PPV 0.79), corresponding to approximately 9–10 true R-peaks with only 2–3 FP samples per 10-s segment. For P- and T-peaks, F1-scores of 0.70 and 0.69 were obtained, respectively. Additional evaluation on arrhythmic LUDB records demonstrated robust R-peak detection across rhythm types. In AF-related rhythms, where organized P waves are physiologically absent, the framework appropriately suppressed P-peak detections, with false-positive rates remaining below 0.31%. Qualitative application to ECG recordings from the PTB-XL database further demonstrated physiologically consistent behavior. These results indicate that reliable and interpretable ECG peak detection under extreme class imbalance can be achieved by integrating lightweight classifiers within the proposed framework, without reliance on complex deep learning architectures.

Abstract: The paper presents an analysis of the risk of failure of port structures in a modern seaport due to vessel impacts. The analysis addresses potential damage related to port maneuvers of self-maneuvering vessels and possible risk reduction options that can be applied to enhance port resilience. The proposed system model—including ship, port infrastructure and environment—enabled the observation of both implemented and anticipated future risk reduction measures. The analysis was carried out using the ferry terminal in the large Polish Port of Gdynia as a case study. A Bayesian influence diagram—including decisions related to the implementation of risk reduction options—was used to determine the total risk associated with ro-pax ferry port calls. Sustainable risk management led to the implementation of a cloud-based monitoring system and, subsequently, to the design of a new terminal in line with the green port concept. A comparative risk assessment for the two locations demonstrated improved safety and reduced environmental pollution in the new Public Ferry Terminal, primarily due to reduced spatial risk and the implementation of cold-ironing technology in the new terminal. The potential future implementation of an automated mooring system was also discussed.

Abstract: This study presents a large-scale empirical comparison of operational efficiency metrics derived from the IMO Data Collection System (DCS) and the EU Monitoring, Reporting and Verification (MRV) framework. Using a matched dataset of 15,755 dual-reported vessels and over 50,000 ship-year observations from 2019 to 2024, paired non-parametric tests, effect size estimation, and agreement diagnostics were applied to assess consistency across monitoring systems. Results indicate that although statistically significant differences are detected (p < 0.001), practical differences are negligible (Cohen’s d < 0.025), with MRV-based values averaging approximately 1.4% lower Annual Efficiency Ratio (AER) and fuel intensity than DCS values. Distributional analysis confirms substantial overlap between datasets, and temporal trends show progressive convergence following the implementation of the Carbon Intensity Indicator (CII) regulation. However, pronounced vessel-type heterogeneity is observed. Flexible cargo vessels exhibit consistent efficiency improvements in EU-related voyages, whereas container ships show minimal variation and LNG carriers demonstrate indicator-dependent patterns. Overall, the findings indicate that DCS and MRV provide broadly comparable representations of operational efficiency, with observed differences primarily reflecting vessel-type-specific operational characteristics rather than structural inconsistencies in reporting systems. The study contributes a scalable statistical validation framework for cross-regulatory monitoring assessment.

Abstract: Chronic kidney disease (CKD) is a progressively worsening condition that erodes renal function over time, reduces quality of life, and can ultimately culminate in kidney failure with far-reaching systemic complications. In addition to reduced filtration, worsening kidney function disrupts mineral homeostasis and leads to CKD–mineral and bone disorder (CKD-MBD). Dysregulated calcium handling and maladaptive endocrine responses contribute to bone pathology and increase cardiovascular calcification risk; therefore, serial calcium monitoring remains clinically relevant for longitudinal CKD management. Conventional calcium measurements are typically obtained with centralized analyzers or laboratory assays (e.g., colorimetry and electrode/optical readouts). Despite high accuracy, the required instrumentation, controlled operating conditions, and pretreatment steps complicate rapid point-of-care deployment, especially when only microliter-scale biofluids are available. Accordingly, this study develops a finger-actuated microfluidic colorimetric platform capable of determining calcium ion concentrations in human biofluids, such as whole blood, serum, and urine. The platform integrates a three-dimensional PMMA/paper microchip with a compact reader that maintains stable temperature control while enabling CMOS-based optical detection. With just 6 μL of sample, a brief finger press propels the biofluid across an internal filtration layer, generating serum or cleaned urine that subsequently reacts with a pre-deposited murexide reagent. Under optimized conditions (1.6% reagent, 50°C, 3 min), the signal follows a strong logarithmic relationship with calcium concentration (Y = 47.273 ln X + 28.890; R² = 0.9905), supporting quantification over 1–40 mg/dL and a detection limit of 0.2 mg/dL. Across 80 clinical CKD specimens spanning serum, whole blood, and urine, results aligned closely with the NM-BAPTA reference assay, with R² values exceeding 0.97.

Abstract: Vibrations of thin sheet-metal parts during robotic manipulation on a production line create a number of serious challenges for production process planning. Modeling the behavior of an elastic plate or shell as a function of the robot manipulator trajectory is typically performed using the finite element method (FEM) and requires significant computational effort. The time factor remains a key limitation for integrating operations involving flexible parts into the virtual commissioning process. In this work, a methodology is proposed that enables accurate real-time reproduction of the behavior of an elastic part during linear robotic manipulation. The approach is based on modeling the response of an elastic part to a prescribed base excitation using the FEM and on the development of a reduced model compliant with the FMI/FMU standard. This reduced model computes, in real time, the convolution of the precomputed base response with the acceleration profile corresponding to the robot TCP trajectory. This makes it possible to determine the total cycle duration, which consists of the part transfer time and the time required for vibration decay at the end of the trajectory down to an acceptable threshold, as well as to perform collision checking while accounting for the deformation of the flexible part. As a result, operations involving elastic parts can be integrated into the virtual commissioning process.

Abstract: A quartz crystal microbalance-based biosensor for the specific detection of the first transgenic common bean (Phaseolus vulgaris L.) cultivar (BRS FC401 RMD) with resistance to bean golden mosaic virus (BGMV) was developed. The immobilization chemistry relies on the strong bond between the thiolated probe and the gold electrode surface. The probe sequence is internal to a region of the BGMV rep gene that was introduced into the common bean genome. The sensor's analytical performance was determined using synthetic oligonucleotides. Real samples of transgenic and wild-type bean seeds were also tested. Sample pretreatment consisted only of enzymatic fragmentation, followed by a thermal denaturation step combined with blocking oligonucleotides. Different biosensor regeneration approaches were studied. Immobilization showed good reproducibility (CV% of 5.8%). The biosensor proved specific for both synthetic oligonucleotides and non-amplified genomic DNA. A linear detection range of 0–1.4 ng/µL was observed, with a detection limit of 0.18 ng/µL. Three sequential detections were performed without loss of surface activity. The results demonstrate the biosensor's potential for direct, real-time, label-free detection of DNA samples for field screening of transgenic common bean cultivars.

Abstract: Physics-based constitutive modelling remains a cornerstone for predicting ductile damage and fracture in metallic materials, particularly where microstructural mechanisms govern macroscopic response. Over the past two decades, a wide range of crystal plasticity, porous plasticity, and void-based fracture models have been proposed to capture deformation localisation, void growth, and coalescence under complex loading paths. However, these developments are often presented in isolation, obscuring their shared physical assumptions and limiting their transferability across material systems and length scales.This article provides a microstructure-sensitive perspective on constitutive modelling of ductile damage and fracture, with particular emphasis on crystal plasticity-based frameworks, void growth and coalescence mechanisms, and interface-driven fracture. Rather than attempting an exhaustive review, this review highlights unifying concepts, modelling trade-offs, and recurring challenges related to parameter identifiability, scale bridging, and predictive robustness. It further clarifies how physics-based constitutive descriptions can be systematically integrated into modern fatigue and fracture assessments and situate these developments relative to emerging data-assisted and machine-learning-enhanced modelling strategies.By reframing established constitutive models within a coherent physical narrative, this perspective aims to support more transparent model selection, improve interpretability, and guide future developments in multiscale damage and fracture modelling of metallic materials.

Abstract: This paper presents a reliability-oriented analytical framework for the quantitative assessment of fragment-resistant multilayer protective equipment subjected to impulsive fragment loading. The study is motivated by the stochastic nature of fragment generation and impact conditions in industrial and occupational accident scenarios, where deterministic penetration criteria are insufficient to describe protective performance. Fragment interactions are modelled as stochastic spatial events, with impact locations and kinematic characteristics treated as random variables and mapped onto a predefined protected region. System failure is formulated using an energy-based limit-state criterion defined by comparison between the absorbed energy demand induced by fragment impact and a critical admissible energy threshold. The fragment–PPE interaction is described using a reduced-order dynamic formulation with concentrated parameters, capturing the dominant normal deformation response under short-duration impulsive loading. Closed-form analytical expressions are derived that relate fragment mass and velocity to impact impulse and absorbed energy. The resulting formulation establishes a direct link between impulse-driven dynamic response, progressive multilayer engagement, and failure probability under single and repeated impact events. Application of the proposed framework to a representative multilayer protective configuration demonstrates physically consistent reliability trends and confirms its computational efficiency. The framework provides a practical tool for reliability-informed assessment and preliminary design of fragment-resistant multilayer protective equipment.

Abstract: The digital preservation of batik, a world intangible cultural heritage, is hindered by the difficulty in performing accurate semantic segmentation on its complex patterns with limited annotated samples. To address this few-shot learning challenge, we constructed a few-shot batik pattern dataset, and proposed a novel network architecture centered on attention weighting and hierarchical decoding. Our method leverages a pre-trained ResNet101 backbone for transfer learning to establish a strong feature foundation. It incorporates a dual-attention module that combines spatial and channel attention to dynamically highlight semantically rich regions and intricate texture boundaries specific to batik. For multi-scale context aggregation, a lightweight module utilizing parallel dilated convolutions is introduced to efficiently capture features from varying receptive fields. Finally, a hierarchical decoder progressively integrates these enhanced, multi-scale features with high-resolution shallow features to reconstruct precise segmentation maps. Comprehensive evaluations on a dedicated batik dataset show that our model achieves state-of-the-art performance, with a mean Intersection over Union (mIoU) of 79.22% and a Pixel Accuracy (PA) of 92.47%. It notably improves over the strong DeepLabV3+ baseline by 3.3% in mIoU and 0.95% in PA, demonstrating its effectiveness for the task of batik pattern segmentation under data-scarce conditions.

Abstract: The mandates of the Circular Economy (CE) and Industry 5.0 necessitate unprecedented life cycle traceability and management for complex products, such as electronics and lithium-ion batteries. The Digital Product Passport (DPP) has emerged as the centralized data framework for this transition. However, effective DPP implementation requires the robust integration of technologies to overcome the technical challenges of scalability, data integrity, and legacy system management. This work shifts from a high-level conceptual overview to an in-depth technical analysis, detailing the integration architecture of the DPP with the Internet of Things (IoT), Digital Twins (DTs), and Artificial Intelligence (AI). Specifically, we focus on how IoT, via embedded sensors and NFC/RFID tags, acts as the dynamic data carrier that feeds DT. For the battery sector, we propose a technical framework that utilizes the DPP to continuously track State of Health (SoH) and State of Charge (SoC), throughout the entire life cycle. AI/Machine Learning is then integrated within the DT to enable accurate predictions of degradation and failure, directly addressing reliability concerns and optimizing the second life and recycling phases. Our focus is on tackling critical bottlenecks, such as interoperability between IT systems and data storage scalability (e.g., via robust Cloud or Blockchain solutions), ensuring the DPP is established not just as a static data repository, but as a dynamic and predictive tool for product reliability management and regulatory compliance.

Abstract: Air Quality (AQ) plays a critical role in public health and urban sustainability, but drawing insights from Air Quality data remains challenging due to fragmented sources, inconsistent formats and varying measurement standards and devices. This paper explores the architecture and standardization of Air Quality datasets from major global monitoring systems, specifically the U.S. EPA’s Air Quality System (AQS) and European Environment Agency (EEA) networks, emphasizing discrepancies in pollutant units, reporting frequencies and metadata quality. The report outlines key pollutants due to road transport emissions and how they are measured using a range of technologies, from fixed regulatory stations to low-cost and satellite-based sensors. The inconsistency in schema design and the lack of interoperability across datasets hinder the scalability of machine learning (ML) pipelines, which rely on clean and harmonized inputs. To address this, an application named “Data Manager Tool” is introduced that ingests, transforms and standardizes heterogeneous AQ data into a centralized “PostgreSQL” database using a star schema. This allows more efficient querying, integration and modeling. The report discusses practical applications of this system, and how it paves the way for scalable ML-based analysis of pollution trends. Future efforts will focus on professional ML approaches, integration of mobile sensor data, and extending the framework to support predictive models and optimization using meteorological and transport datasets.

Abstract: This study aims to conduct and experimental modal analysis of a very lightweight composite structure representative of UAV application and to evaluate the suitability of different testing approaches for reliable identification of its dynamics characteristics. The investigated structure is a winglet made of carbon fibre reinforced polymer (CFRP) with a lightweight foam core. The experimental campaign was based on impact hammer excitation combined with triaxial accelerometer measurements. Modal test were performed under three different boundary conditions: free-free suspension using elastic cords, free-free approximation using compliant foam support, and fixed conditions reflecting to the operational mounting of the winglet. The objective of the study is not only to identify natural frequencies and mode shapes, but primarily to assess the influence of support conditions, excitation quality and measurement-induced mass loading on the reliability of the extracted modal parameters. By comparing the obtained frequency response functions and modal characteristics across different test configurations, the work seeks to identify the most appropriate experimental approach for modal analysis of ultra-lightweight composite UAV structures, providing practical guidance for future vibration investigations.

Abstract:

A solar-air source absorption heat pump (SAAHP), which mainly consists of a solar collector, a fan coil, and an absorption heat pump equipped with a gas-fired combustor, has been proposed for water heating. This system runs in SD (Solar-energy driving) or GD (Gas-combustion-heat driving) mode, designed to utilize renewable energies as much as possible. The models for each component were built and the corresponding heat and mass balance equations were established. The performance of SAAHP based on LiBr/H₂O and LiNO₃/H₂O working fluids was simulated and compared with an air source absorption heat pump (AAHP) based on LiBr/H₂O. Results indicated that SAAHP based on LiNO₃/H₂O has a higher solar energy utilization rate than that based on LiBr/H₂O due to its lower solar collector inlet temperature in SD mode. In comparison to AAHP based on LiBr/H₂O, SAAHP based on both of LiBr/H₂O and LiNO₃/H₂O achieved a higher primary energy COP throughout a year. Relative to a gas-fired hot water boiler, SAAHP based on LiNO₃/H₂O and LiBr/H₂O achieved yearly primary energy saving rates of 46.2% and 40.0%, respectively, whereas AAHP just achieved 12.2%. SAAHP based on LiNO₃/H₂O shows significant energy saving potential in the building energy consumption.

Abstract: The main purpose of the paper is to show the possibility of assessing the dynamic properties of crankshaft in the early design phase of engine development, without performing dynamic forced response simulation. This is achieved by carrying out modal analysis of crankshaft under existing boundary conditions, namely by taking the radial stiffness in main bearings and the masses of moving conrods and pistons into account. The spectra of eigenfrequencies and corresponding mode shapes as a result of such supported modal analysis are compared to those of free modal analysis, emphasizing the influence of the boundary conditions. To easily identify the modes and to compare them with each other, kinetic energy-based method is used, alongside visualization and animation of mode shapes. The examples of crankshafts considered in the paper are taken from model catalog of virtual engines of six different sizes and configurations, being compared to that of in-line 4-cylinder engine as the reference case. All types of modal analysis are performed on structured FE models of crankshafts using software tool AVL EXCITE™ Shaft Modeler.

Abstract: Effective implementation of university extension projects requires a structured ap-proach with explicit objectives, dependencies, and resource constraints. Automated planning, particularly Hierarchical Task Network (HTN) planning, provides an opera-tional method to schedule complex activities by decomposing high-level goals into ex-ecutable tasks. This paper presents a tool that converts university project documents into BPMN 2.0 declarative processes and automatically produces HTN planning in-puts. Following the Design Science Research methodology, the architecture integrates (i) a pre-planner parser that extracts activities, roles, and precedence relations from BPMN models, (ii) generators that create domain and problem files, and (iii) the HTN planners SHOP 2 and PyHOP to synthesize executable plans. The system was validated with three categories of projects provided by two public universities in Colombia: Universidad Nacional de Colombia and Universidad de Caldas. The platform produces multiple alternative plans for each project and reports plan length and solution-search cost, enabling direct comparison across planners. Results show that the proposed workflow reduces manual scheduling effort, improves consistency of implementation roadmaps, and supports evidence-based selection of implementation strategies under different constraints. These capabilities help extension offices formalize knowledge, audit decisions, and reuse plans across initiatives. Traceability links BPMN elements to planning tasks.

Abstract: Multimodal Named Entity Recognition (MNER) leverages both textual and visual information to improve entity recognition, particularly in unstructured scenarios such as social media. While existing approaches predominantly rely on raster images (e.g., JPEG, PNG), Scalable Vector Graphics (SVG) offer unique advantages in resolution independence and structured semantic representation—an underexplored potential in multimodal learning. To fill this gap, we propose MNER-SVG, the first framework that incorporates SVG as a visual modality and enhances it with ChatGPT-generated auxiliary knowledge. Specifically, we introduce a Multimodal Similar Instance Perception Module that retrieves semantically relevant examples and prompts ChatGPT to generate contextual explanations. We further construct a Full Text Graph and a Multimodal Interaction Graph, which are processed via Graph Attention Networks (GATs) to achieve fine-grained cross-modal alignment and feature fusion. Finally, a Conditional Random Field (CRF) layer is employed for structured decoding. To support evaluation, we present SvgNER, the first MNER dataset annotated with SVG-specific visual content. Extensive experiments demonstrate that MNER-SVG achieves state-of-the-art performance with an F1 score of 82.23%, significantly outperforming both text-only and existing multimodal baselines. This work validates the feasibility and potential of integrating vector graphics and large language model–generated knowledge into multimodal NER, opening a new research direction for structured visual semantics in fine-grained multimodal understanding.

Abstract: A laboratory–scale mechanical draft cooling tower equipped with eight sections of perforated inclined plates was designed to determine the effect of operating conditions on the volumetric mass transfer coefficient (kya) between water and air. A three–factor, three–level design of experiments (DOE) was implemented, considering liquid mass flow rate L (120, 240, and 360 kg/h), gas mass flow rate G (36, 57, and 75 kg/h), and top water temperature TL2 (50, 60, and 70◦C). A total of 54 runs were performed, and the global volumetric mass transfer coefficient was calculated by combining energy and mass balances with the Mickley method. The experimental data were fitted to a power–law correlation using multivariable regression. The ANOVA showed that TL2 is the dominant factor, followed by L, whereas the influence of G is comparatively small in the studied range. The selected correlation, based on the nominal gas flow rate, achieved R2=0.869 and a RMSE of 5930 kg/(m3h). The kya values were found in the range from 4600 to 62000 kg/(m3h). Vertical temperature profiles of water and air along the column revealed that, for high liquid flow rates, most of the cooling occurs in the lower stages, suggesting that the upper sections are underutilized.

Abstract: This paper addresses distributed formation control for multiple unmanned aerial vehicles (UAVs) operating in obstacle-dense environments under directed switching communication topologies. A leader–follower architecture is adopted, wherein the leader performs online trajectory replanning while followers rely on delayed and intermittently available neighbor information. To simultaneously tackle collision avoidance, formation feasibility under narrow passages, and communication intermittency, we propose an integrated deformable formation navigation framework. The framework couples Safe Flight Corridor (SFC)-constrained Bézier trajectory planning with a dynamic formation scaling mechanism, allowing the swarm to adaptively shrink or expand its geometric configuration when traversing constricted spaces, thereby ensuring all agents remain within certified collision-free corridors. A nonlinear distributed consensus-based estimator is designed to propagate leader reference states under directed switching graphs with bounded delays. Using a max-min contraction analytical approach, we establish guaranteed practical convergence for both leader tracking and inter-follower agreement without requiring persistent connectivity. Extensive simulations in complex cluttered environments demonstrate that the proposed approach enables flexible and real-time formation reshaping, enhancing navigational safety and robustness while maintaining cohesive swarm behavior under challenging communication and spatial constraints.

Abstract: The widescale deployment of radars, distributed across a platform and across multiple platforms for reliable 360° situational awareness (SA), introduces the challenge of radar interference. Interference can broadly be categorised as self-interference (between radars mounted on the same platform) and mutual interference (signals received from radars on other platforms). Both types of interference impede the reliability of SA delivered by such systems, particularly in dense environments where numerous radars operate simultaneously within the same frequency band. This work presents a comprehensive evaluation of a multi-modal beamforming approach that combines unfocused synthetic aperture radar with the traditional Multiple-Input, Multiple-Output beamformer to enhance radar resolution and suppress interference. Additionally, various aspects of sensor configurations defining hardware and software capabilities of state-of-the-art radars are discussed, and a systematic analysis of signal-to-interference-plus-noise ratio at each step of the processing is presented. Extensive simulations and experimental results in both automotive and maritime environments are shown to validate the effectiveness of the proposed approach.

Abstract: The design, fabrication, and characterization of a highly transparent and flexible monopole antenna optimized for the 3–6 GHz frequency band is presented in this paper. In traditional Transparent Conductive Oxide (TCO) designs, there is always a trade-off between the RF efficiency and transparency. Therefore, an Aerosol Jet® 5X system was used to directly print a silver nanoparticle mesh over a 50-µm polyimide (PI) substrate. With this fabrication method, a durable structure was yield which works well both electrically and mechanically with 85% transparency and a gain of −2.5 dBi. In order to demonstrate how the antenna is flexible and compatible with other devices, it was bent over a cylindrical body and was integrated with a commercial solar panel. The results show that impedance matching and radiation characteristics of the antenna remain stable under bending conditions, and no critical decrease was observed in solar energy harvesting. Consequently, this design presents a suitable solution for energy-autonomous IoT systems, smart windows, and CubeSat applications.