Engineering

Abstract: The stability of financial institutions is crucial; however, current regulations evaluate credit, liquidity, and market risks separately, which hampers a consistent assessment of an entity's true loss-absorbing capacity. To address this, our study introduces the Risk Capacity Index (ICR) as a comprehensive indicator of financial sustainability for organizations in Colombia's solidarity sector. The approach adjusts a macrofinancial risk capacity model to fit the institutional setting, defining the ICR as the ratio of technical equity to total risk exposure, including expected credit losses, market Value-at-Risk, and the liquidity gap. This index was empirically tested with monthly data from 2025 from a closed savings and credit cooperative, using sensitivity tests and stress scenarios aligned with Basel III standards. Results show that liquidity risk is the main driver of capacity depletion, responsible for most of the index's fluctuations and causing non-linear deterioration during adverse conditions, while market risk effects are minor. Significant funding pressures sharply reduce the ICR below viability levels, leading to structural issues on the balance sheet. The ICR provides a new, integrated early-warning tool that complements traditional solvency measurements. The study highlights that managing liquidity and liabilities proactively, rather than just increasing capital, is key to preserving financial stability in cooperative models.

Abstract: Additive manufacturing of PEEK/regolith composites offers a promising route for lunar in-space manufacturing by reducing dependence on Earth-supplied materials. However, the processability of these composites and the elastic response of printed components are strongly influenced by regolith loading and manufacturing-induced defects. This study develops a hierarchical finite element framework to quantify the stiffness of additively manufactured PEEK containing 10-50 wt% LMS-1D lunar regolith simulant and to distinguish intrinsic composition effects from defect-driven stiffness losses. The approach combines composition-based estimation of regolith properties with microstructure-informed simulations of PEEK/regolith composites. Under defect-free assumptions, the predicted modulus increases monotonically from 1.27 GPa at 10 wt% to 1.97 GPa at 50 wt%, showing good agreement with experimental trends up to 40 wt%, where deviations remain within 3.5-10.1%. At 50 wt%, however, the experimental modulus decreases to 1.27 GPa, while the defect-free model predicts 1.97 GPa. Microscopy-informed single-layer analyses indicate that tall crack-like interfacial voids, polymer-starved welds, and interconnected weak seams significantly reduce load transfer and shift the mechanical response toward an interface-controlled regime. These results show that regolith additions can enhance stiffness only until defect connectivity becomes dominant. The findings provide insight into the process-structure-property relationships governing ceramic particle-reinforced high-performance thermoplastics in additive manufacturing.

Abstract: Classical thin-walled beam theory, following Vlasov, decomposes the longitudinal normal stress field into axial, bending and a single sectorial warping contribution governed by the bimoment B and the warping constant I_w. This work explores a natural generalization of that framework: a modal decomposition of warping in which the classical Vlasov sectorial mode is treated as the first member of a broader, orthogonal family of warping modes. Each additional mode φᵢ(y, z) is defined through orthogonality conditions with respect to area, bending and the sectorial coordinate, carries its own generalized bimoment qᵢ(x) and generalized warping inertia Iᵢ, and induces, through longitudinal equilibrium, a self-equilibrated shear flow governed by a Neumann-type Poisson problem on the cross-section. The resulting strain energy decomposes additively over modes, preserving full compatibility with energy methods. The physical origin of the higher-order modes is discussed, with eigenfunctions of a Laplacian cross-sectional operator, finite element cross-sectional eigenanalysis, and variational energy minimization proposed as candidate generating mechanisms. The framework is formulated independently of any thin-wall or mid-line kinematic hypothesis, making it directly applicable to thick-walled and solid cross-sections as well as to classical thin-walled members. Under this view, Vlasov warping theory emerges as the fundamental mode of a richer modal warping basis, analogous to the role of the fundamental mode in structural vibration theory.

Abstract: With the rapid growth of electric vehicles (EVs), high charging demand has in-creased uneven station utilization, and pressure on distribution networks. Therefore, this paper proposes a collaborative method for capacity planning and charging scheduling of multiple charging stations (CSs) considering time-of-use (TOU) pricing and energy storage system (ESS). A total system cost model is established, including transformer and ESS costs. Secondly, a deep neural network-guided im-proved sparrow search algorithm (DNN-ISSA) is proposed to optimize the number of chargers and parking spaces by predicting the initial capacity center. Furthermore, a charging scheduling algorithm is proposed to optimize user charging time by introducing a TOU price response function to modify charging probabilities. A case study of 36 CSs in Jinan shows that the proposed method reduces average charging time by 15.7, 15.4, and 15.2 minutes for 1,000, 5,000, and 10,000 demand points, while lowering the total system cost from 73.92 to 70.36 million yuan. The convergence value of DNN-ISSA reduces by 15.05%, 21.67%, and 11.61% compared with the sparrow golden-sine optimization algorithm (SSA-GSA), particle swarm optimization algorithm (PSO), and sparrow-particle swarm optimization algorithm (SSA-PSO), respectively. The proposed method enhances energy utilization, mitigates peak loads, and supports low-carbon EV charging operation.

Abstract: Light detection and ranging (LiDAR) sensors are widely known for their applications in 1 robotics, autonomous cars, remote sensing, and object tracking. These compact sensors are favored by 2 automation technologists for their accurate and long-range object detection capabilities. As a result, 3 a multitude of LiDAR sensors with diverse specifications have been introduced in the commercial 4 market. However, existing 3D LiDAR sensors lack the ability to customize their specifications, such 5 as measuring range, Field of View (FoV), angular resolution, number of scan points, and number of 6 scan layers, according to specific applications. This limitation poses several challenges for engineers, 7 including processing excessive data, requiring high computation power, demanding more storage 8 space for post-processing, and leading to false feature detection. To address these issues, this paper 9 presents a novel reconfigurable 3D LiDAR technology called 3D Customizable LiDAR (3D CS LiDAR), 10 which offers customization of sensor specifications based on the application requirements. The paper 11 discusses the mechanical and electrical systems of the developed LiDAR sensor and evaluates its 12 object detection performance by comparing it with a commercially available sensor. The results 13 demonstrate that the proposed reconfigurable 3D LiDAR system outperforms the commercial sensor 14 in all considered scenarios, indicating its potential for various perceptional applications. This research 15 contributes to the advancement of LiDAR technology by introducing a customizable approach, 16 addressing the limitations of existing sensors. The findings showcase the initial progress made 17 towards developing a comprehensive reconfigurable 3D LiDAR system, which holds promise for 18 diverse practical applications.

Abstract: A graph attention network-enhanced multi-agent proximal policy optimization (GAT-MAPPO) framework is proposed for cooperative guidance in counter-attack/defense scenarios. A dynamic heterogeneous interaction graph is formulated over interceptors and targets at every decision epoch. Through a multi-head graph attention encoder, relational features capturing both inter-interceptor cooperation and target threat dynamics are adaptively aggregated. These graph-enriched observations are processed by a Centralized-Training, Decentralized-Execution (CTDE) MAPPO architecture, guided by a hierarchical reward function that mandates miss distance minimization, simultaneity of arrival consensus, multi-directional encirclement, and smooth control effort. Furthermore, the integration of a three-stage curriculum learning strategy allows for robust cooperative policy derivation across transitions from rectilinear to highly adaptive evasion patterns, eliminating the need for explicit rule engineering. Extensive Monte Carlo simulations confirm GAT-MAPPO’s superior performance: achieving >95% interception success rate in 4-vs-4 scenarios and reducing mean simultaneity error by 41.4% compared to the MAPPO baseline. Comprehensive ablation studies validate the critical roles played by graph attention encoding, reward hierarchy design, and progressive curriculum staging.

Abstract: With increasingly strict air quality standards and growing concerns about air pollution, fine and ultrafine particulate matter remains a major challenge for conventional air cleaning technologies. Due to their small size, these particles are difficult to remove using traditional filtration and separation methods. Acoustic agglomeration can be used as a pre-treatment technology to increase particle size in a high-intensity acoustic field and improve the efficiency of particle removal. This study investigates acoustic-induced agglomeration of solid aerosol particles in a dynamic airflow system. The effects of acoustic frequency were evaluated at 3, 5.5, 7.5, and 15 kHz under a sound pressure level of 135 dB and at two airflow velocities: 0.75 m/s and 1.5 m/s. These velocities corresponded to different particle residence times in the acoustic field. Arizona test dust was used as the test aerosol, and particle number concentration and particle size distribution were measured before and after the acoustic field. The results showed that acoustic agglomeration of fine and ultrafine particles was strongly affected by both acoustic frequency and particle residence time. The highest agglomeration efficiency, reaching up to 42%, was obtained at 3 kHz, 135 dB, and longer particle residence time. These findings indicate that acoustic agglomeration can promote particle size redistribution in moving airflow and may be used as a pre-treatment method for improving particulate matter removal in air quality control systems.

Abstract: Active suspension can deliver better ride comfort and stability than a passive layout system, but only when the controller is properly tuned. This work proposes a simple tuning method: retaining Proportional-Integral-Derivative (PID) structure and employing two off-the-shelf optimizers (Optuna and Particle Swarm) to select the gains. The full-vehicle 7-Degrees of freedom (DOF) benchmark of Kumar et al. was used as a virtual test bench to compare four controllers: passive, Kumar et al. (2020) PID. A manually tuned baseline, and the two optimizer-tuned versions. The cost function combines RMS body acceleration, pitch and roll angular rates, peak actuator force and jerk across three simulation scenarios (bump, speed-breaker, cornering), with soft penalties for actuator saturation, suspension travel, and tire lift-off. The Optuna gains cut the global cost by 19% relative to the manual baseline and by 8.5% relative the Kumar et al. (2020) PID, primarily by reducing the peak actuator force from 2.7 kN to 1.78 kN. A 100-vehicle Monte-Carlo study (±20% on sprung mass, ±15% on stiffness and damping) confirms that the performance advantage is robust to variations in the nominal parameters.

Abstract: Olive is a major agricultural crop extensively cultivated throughout the Mediterranean region. However, olive trees are vulnerable to several diseases that can negatively affect productivity and yield. One of the most widespread foliar diseases is olive leaf peacock spot, caused by the fungus Cycloconium oleaginum. Early detection of this disease is essential for preventing leaf drop, limiting disease spread, maintaining tree health, and reducing treatment costs before the infection reaches an advanced stage. In this study, a multimodal hybrid deep learning framework is developed to detect peacock spot disease in olive leaves and assess disease severity based on visual and numerical features. The proposed framework integrates olive leaf images with soil conditions, environmental conditions, and vegetation and stress indices to provide a more comprehensive disease analysis than image-only approaches. A ResNet50-based convolutional neural network is used to extract visual features from leaf images, while a multilayer perceptron processes the numerical sensor-based and index-based data. These features are then fused within a unified learning framework to classify disease stages and estimate leaf damage severity, including lesion coverage and yellowing percentage. The performance of the proposed model was evaluated using standard performance metrics suitable for both classification and regression tasks. For classification, the model was evaluated on 494 testing samples and achieved an overall accuracy of 97.77 %, with a macro F1-score of 0.9809 and a weighted F1-score of 0.9776. In addition, the model achieved low regression errors, with mean absolute errors of 1.16 % for lesion coverage and 1.42 % for yellowing estimation. These results demonstrate the effectiveness of the proposed multimodal framework for accurate peacock spot detection and severity assessment, supporting its potential use in smart agricultural monitoring and disease management.

Abstract: High-speed electric rotating machines enhance power density and eliminate gearboxes in waste heat recovery microturbines, but conventional designs face high manufacturing costs and complex cooling requirements. This study presents the development, experimental validation, and comparative analysis of high-speed configurations. Initially, a lower-speed induction machine prototype operating at 13,000 rpm was built using standardized components to experimentally validate numerical loss models. Experimental testing of the initial prototype confirmed a total loss of 7.89 kW, closely matching the simulated 7.75 kW. Leveraging these findings, two next-generation topologies of decreased size, an induction machine and a surface permanent magnet machine, were designed and evaluated using finite element method and conjugate heat transfer simulations under sinusoidal and pulse-width modulation excitations. At a 14,000 rpm operational point, the surface permanent magnet prototype outperformed the induction machine configuration, ensuring the lower temperatures of the permanent magnet machine and achieving 63.2 kW of mechanical power and 96.21% efficiency compared to the induction machine's 52.4 kW and 94.64%.This paper builds upon the microturbine generator project introduced at the ICPGEEC 2025 conference, by presenting lighter, higher-speed machine designs.

Abstract: Surface electromyography (sEMG) is the dominant control modality for myoelectric hand prostheses, yet a persistent gap remains between high offline classification accuracy and real-time, on-device control: models are often large, evaluated under leaky window-level splits, and reported with a single inflated metric that hides instantaneous behaviour. Our contribution is primarily methodological: a strictly leak-safe within-user extrapolation protocol (calibrating on a user's earlier repetitions and testing on a held-out later repetition, split by contiguous recording segment so no overlapping window crosses the split), paired with honest dual reporting of per-window (instantaneous) and per-execution balanced accuracy alongside the rest false-activation rate. We instantiate it with a compact (28,199-parameter, 27.5 kB int8) decoder for a functionally-motivated set of six hand grips plus rest: a depthwise-separable CNN, leak-safe per-user calibration, and a causal sequence-reasoning decoder in which a population-learned, zero-parameter transition grammar reasons over per-window evidence. We evaluate on NinaPro DB2. With five calibration repetitions the system reaches 94.0% per-execution balanced accuracy at 2.2% false-activation with the deployed operating-point gate (92.7% from the gate-free decoder stream alone), and 75.2% per-window accuracy, with a microcontroller-compatible 28k-parameter model. We report a rigorous ablation: a classical time-domain baseline is competitive on the saturating per-execution metric, while the learned representation wins on the discriminating per-window metric, and the causal sequence decoder gives a small but statistically significant per-window gain over a matched-window majority vote (Wilcoxon p = 0.001) at zero added parameters. Under cross-subject 40-fold leave-one-subject-out evaluation the same pipeline reaches 50.6% per-execution balanced accuracy with no labelled calibration (unsupervised per-channel normalisation only), rising to 89.1% with 20% labelled calibration, where the sequence layer contributes a larger gain than within-user. We position the work honestly against recent ultra-low-power and foundation-model EMG decoders and report compute cost rather than claiming on-device measurements. Code and configurations are available at https://github.com/seanb9/emg-leaksafe-benchmark.

Abstract: Recent advances in miniaturized neural-interfacing probes have led to broadening of the knowledge on brain functions via recording of local field potentials and provide a great framework for interacting directly with neurons by different means of brain stimulation, typically by delivering electrical stimuli. Although fabrication of bidirectional implantable neural probes based on MEMS materials and methods is already a mature technology, the fact that electrical brain stimulation relies solely on a good electrical contact between the conductive electrode and the living tissue limits the lifespan of such implants, mainly due to foreign body inflammatory response and scar-tissue growth on the electrode-tissue interface. As a solution, we propose a neural probe for state-of-the-art biopotential recording, with a monolithic integrated miniaturized piezoelectric resonator for neural modulation which is fully encapsulated in biocompatible material.Theoretical analysis on piezoelectric resonating membranes is presented and its limitations are assessed and compared to finite element analysis (FEA), which is then used to tune the resonant frequency of the vibrating membrane to the desired value for neural stimulation. As well the acoustic output and penetration depth of the generated signal are analyzedFinally, the MEMS fabrication methodology for the proposed neural probe is presented.

Abstract: This paper analyzes the use of Omani limestone, collected from the mountain ranges of the Sultanate of Oman, a major global exporter of limestone, as a natural reinforcement to improve the mechanical properties of recycled aluminum alloys obtained from automotive engine cylinder scraps. The recycled aluminum was melted and mixed with limestone powder at volume fractions of 2.5%, 5%, and 7.5% using the stir-casting technique. Mechanical test results showed progressive, statistically significant improvements (one-way ANOVA, p < 0.05) with increasing reinforcement content. At the optimal 5.0 vol.% fraction, tensile strength increased by 16.9%, surface hardness improved by 19.7%, and impact resistance increased by 28.6% relative to the unreinforced alloy. Scanning Electron Microscopy (SEM) confirmed uniform particle distribution and microstructural densification at 5.0 vol.%, but severe porosity, particle agglomeration, and microcrack initiation at 7.5 vol.%. X-ray diffraction (XRD) analysis confirmed the stable presence of CaCO₃ compounds and positive interaction with the aluminum matrix. The 5.0 vol.% fraction is identified as optimal, delivering the best balance of mechanical enhancement and microstructural soundness. These results suggest promising prospects for Oman's limestone waste in advanced engineering applications with significant environmental and economic benefits.

Abstract: Visual simultaneous localization and mapping in indoor dynamic environments remains challenging because moving objects introduce unreliable correspondences, whilst removing dynamic feature points often leaves insufficient static features in low-texture regions. This paper proposes a robust visual SLAM framework that combines semantic-geometric feature filtering with texture-aware feature compensation to improve pose estimation under dynamic interference. The framework first identifies potentially dynamic regions through pixel-level semantic segmentation and removes features associated with highly dynamic objects. To reduce over-filtering and address semi-static objects, depth variation and multi-view geometric consistency are further used to distinguish static and moving feature points across consecutive frames. After dynamic filtering, a learned local feature extractor is introduced to improve descriptor discriminability and feature density in reliable static regions. Two additional modules, semantic confidence weighting and static region feature compensation, adaptively adjust feature extraction thresholds so that low-texture but geometrically useful areas can contribute more stable correspondences. The proposed system is implemented within a visual SLAM pipeline and evaluated on public dynamic RGB-D benchmarks, including TUM and Bonn sequences. Experimental results indicate that the method improves localization robustness in high-dynamic scenarios and reduces trajectory error compared with conventional ORB-based SLAM and several dynamic SLAM baselines. The study demonstrates the potential of combining semantic priors, geometric verification and adaptive feature enhancement for dynamic indoor localization.

Abstract: The transition from internal combustion engine vehicles toward hybrid electric vehicles and battery electric vehicles has transformed not only automotive engineering but also the sensory and emotional experience of driving. While previous studies have examined environmental and performance differences between propulsion systems, limited researches investigated their direct impact on human neurophysiology and emotional perception during real-world driving. This study investigates the neurophysiological and autonomic responses of drivers exposed to three vehicle categories: electric vehicles, internal combustion engine vehicles, and hybrid electric vehicles. A standardized driving sessions were performed in urban driving environment, while electroencephalography, heart rate, heart rate variability, galvanic skin response and peripheral blood oxygen saturation were continuously recorded. Measurements were collected during three phases: a five-minute pre-driving baseline, twenty minutes of active driving, and a five-minute post-driving recovery period. Electroencephalography power in the theta (4–8 Hz), alpha (8–12 Hz), low beta (12–20 Hz), and high beta (20–30 Hz) frequency bands were analyzed as indicators of cognitive workload, cortical relaxation, and attentional engagement. Cardiovascular and electrodermal signals were interpreted as markers of autonomic arousal, sympathetic activation, and stress regulation. Peripheral oxygen saturation was included as complementary index of cardiorespiratory and metabolic demand across vehicle conditions.

Abstract: Hydrogen blending in natural gas pipelines is a promising decarbonization pathway. This study investigates a coaxial-swirl static mixer for hydrogen-natural gas mixing at ratios of 5% to 30% H₂. The mixer features nine ring-shaped cavities with 120° helical torsion to enhance turbulent mixing. A calibrated 2D axisymmetric computational model was developed and validated against experimental data. Results show that the configuration achieves 95% mixing uniformity within 8.2D to 9.0D across all blending ratios, meeting industry targets with minimal pressure penalty (<0.04% of operating pressure). Validation shows good agreement with literature, with mixing intensity profiles matching within 5%. This work supports the integration of hydrogen into existing infrastructure for near-term decarbonization.

Abstract: Personal voice activity detection (PVAD) identifies whether each detected speech frame originates from a designated target speaker. Modern PVAD systems are typically trained offline and then deployed with frozen model parameters and a fixed, pre-enrolled speaker embedding, leaving them unable to adapt to distribution shifts at inference time such as unseen acoustic environments, changing speaking styles, or mismatches between enrollment and test conditions. Test-time training (TTT) and test-time adaptation have shown promise in language, vision, and several speech tasks, yet their behavior on PVAD has not been studied. In this work, we present an empirical study of two complementary test-time adaptation mechanisms built on top of the recently proposed FDE-Mamba backbone. The first is a VAD-gated TTT adapter, which instantiates the TTT-Linear formulation within the personalization pathway and augments it with a VAD-probability gate and exponential moving-average stabilization, adapting an internal weight matrix on the speaker-conditioned feature stream of each test utterance. The second is TEA (Test-time Embedding Adaptation), a scheme that keeps all model parameters frozen and instead adapts the target speaker d-vector itself via self-supervised objectives at inference time, directly targeting enrollment–test mismatch. We evaluate both mechanisms on the LibriSpeech PVAD benchmark across two backbones (LSTM-based FDE-RNN and Mamba-based FDE-Mamba), reporting category-wise average precision, mean average precision (mAP), accuracy, recall, precision, and real-time factor. Our results show that test-time adaptation yields consistent but modest gains over the FDE-Mamba baseline (e.g., mAP from 0.9605 to 0.9641 and precision from 0.881 to 0.899), while slightly reducing recall and increasing inference cost when TEA is enabled. Through ablation studies, we quantify the independent and combined contribution of each component, characterize the recall–precision trade-off introduced by adaptation, and isolate the effect of post-hoc Gaussian smoothing so that its benefit is not conflated with that of the adaptation mechanisms. These findings, together with a discussion of their limitations and cost–benefit profile, provide a measured baseline and design insights for future work on adaptive PVAD, particularly under stronger acoustic and enrollment mismatches than those captured by the LibriSpeech protocol.

Abstract: Next-generation wireless systems are becoming more complex, and the need for intelligent mobility management mechanisms that can ensure service continuity and efficiently utilize network resources has been growing. Frequent handovers, unequal distribution of traffic, variable network conditions, and multiple radio access technologies are some of the challenges that traditional mobility control strategies are likely to face in 5G and future 6G networks with dense deployments of small cells, heterogeneous architectures, and highly mobile users. These constraints frequently lead to sub-optimal user experience, higher overhead in signalling and inefficient use of resources. The introduction of new tools through the advancements of artificial intelligence (AI), specifically machine learning and deep learning techniques have created new opportunities for "predictive" and "adaptive" mobility optimization. The use of data-driven decision-making can enable AI-based solutions to predict user movements, fine-tune the execution of handover and dynamically allocate radio resources to enhance network performance. This survey This paper presents a comprehensive survey of AI-enabled handover strategies for mobility load management 5G, Beyond-5G (B5G), and upcoming 6G networks. This study provides a review of current studies, classifies the framework approaches of mobility management based on AI technologies, and identifies their architecture, learning and optimization goals. Moreover, the survey assesses the performance of intelligent handover schemes to solve the critical issues like load balancing, interference mitigation, connection reliability and quality of service maintenance. Key performance indicators related to mobility robustness, resource efficiency and service continuity are compared between the conventional mobility management methods and the AI based ones. Last but not least, the paper outlines unsolved problems, new trends and potential areas of research that will guide the evolution of autonomous mobility management solutions for the future of wireless communication networks.

Abstract: The cement industry is one of the most challenging sectors to decarbonize due to the coexistence of high-temperature thermal demand and process-related emissions from limestone calcination. This study presents an energy and emissions assessment of cement manufacturing based on representative mass and energy balances derived from literature benchmarks and industrial operating data. Typical cement production requires 2.8–3.6 GJ of thermal energy and 80–120 kWh of electricity per ton of final product, resulting in total emission in the range 500–850 kg CO₂/t cement, of which 55–65% originate from clinker calcination. Moving from this baseline, possible decarbonization pathways are evaluated, including energy efficiency improvements, clinker substitution through supplementary cementitious materials use of alternative fuels, electrification, hydrogen utilization and carbon capture technologies. The analysis shows that energy efficiency measures provide relatively limited reductions (10–30 kg CO₂/t cement), while alternative fuels and clinker substitution can achieve larger but still partial benefits. Hydrogen emerges as a promising option for decarbonizing the combustion-related share of emissions, with a potential reduction ranging from 50 to 250 kg CO₂/t cement, particularly when integrated with oxy-fuel combustion systems. Deep decarbonization ultimately requires carbon capture and storage (CCS), the only technology capable of addressing the substantial process emissions inherent to clinker production and use of hydrogen can be relevant too.

Abstract: Diabetes mellitus is a major global public health challenge, with prevalence rising due to aging populations, urbanization, sedentary lifestyles, and dietary changes. Effective glucose monitoring is essential for diagnosis, treatment, and long-term disease management, driving significant research into improved sensing technologies. Conventional invasive and minimally invasive glucose monitoring methods often cause discomfort and reduce patient compliance, motivating the development of non-invasive alternatives. Recent advances in photonics, biomedical engineering, nanotechnology, wearable devices, and artificial intelligence have accelerated the emergence of innovative glucose sensing approaches capable of improving comfort, safety, and monitoring frequency. This review presents a comprehensive overview of recent non-invasive glucose monitoring technologies reported in the literature, including optical, electromagnetic, nanotechnology-based, and physiological sensing methods evaluated through human studies, biological samples, or tissue-equivalent models. The underlying sensing principles, measurement sites, performance characteristics, and practical implementation challenges of these technologies are discussed. Particular attention is given to the integration of machine learning algorithms, which have demonstrated significant potential for enhancing glucose prediction accuracy and supporting real-time monitoring applications. The review also critically examines the advantages, limitations, clinical feasibility, and commercialization prospects of existing technologies, highlighting the key barriers that continue to impede widespread adoption. By consolidating recent developments across multiple scientific and engineering disciplines, this work provides researchers, clinicians, and technology developers with a concise assessment of the current state of non-invasive glucose sensing and identifies future research directions necessary for advancing reliable, accurate, and user-friendly next-generation diabetes management systems.