Search Results (683)

Systemic disruptions in technologically complex projects often spread beyond the function in which they are first detected. This article follows that movement in two contrasting cases: research and development (R&D) of a software-hardware measuring complex and production of optical data-transmission equipment. The study uses a comparative two-case design and examines corporate records for 2020–2025 through factor screening, scenario analysis, and Monte Carlo simulation. The propagation paths differ. In R&D, restricted access to critical components leads to supplier search, technical qualification, documentation changes, and repeated testing; in production, the main consequences arise through component completeness, delivery delays, production scheduling, and shipment timing. Quantitative assessment estimates downside net present value (NPV) exposure rather than resilience itself. A compound sensitivity analysis indicates that simultaneous deterioration in component availability and delivery conditions may materially increase financial exposure. The process-role model connects the disruption signal to the affected process, the role responsible for interpreting its consequences, and the level at which corrective action can be authorized. Organizational resilience is treated as the management context for this interface, not as a directly measured outcome. Full article

Tribology, the science of friction, wear, and lubrication, governs the reliability of nearly every mechanical system. Tribological contacts account for approximately 23% of global energy consumption, including 20% used to overcome friction and 3% associated with remanufacturing worn components. Nevertheless, the field has remained constrained by low experimental throughput, operator variability, and a scarcity of standardized, reusable datasets. This review surveys two converging trends positioned to address these limitations: the development of robotic and automated platforms for tribological experimentation, and the growing application of artificial intelligence to tribological analysis. High-throughput tribometer architectures, robotic specimen preparation, and multi-modal in situ sensing are examined as components of an emerging automated tribometry infrastructure. Supervised learning, physics-informed neural networks, and Bayesian optimization are reviewed as AI methods organized by the data regime in which they operate. The convergence of these trends in closed-loop autonomous tribological experimentation is assessed, including system architecture, optimization target specification, current partial implementations, and tribology-specific integration barriers that distinguish this domain from adjacent self-driving laboratory applications. Application domains spanning industrial machinery, biomedical implants, and aerospace and automotive drivetrains are discussed. Key challenges including dataset standardization, model transferability, and hardware-software integration complexity are identified. Prospects for fully autonomous tribological discovery pipelines are outlined, with emphasis on open-access data infrastructure and physics-constrained learning as the enabling conditions for the field. Full article

Reinforcement learning (RL) is increasingly used for real-time control of complex dynamical systems, but its practical performance must be evaluated under hardware constraints that are often simplified in simulation. This paper presents a comprehensive review of published RL-based control studies using the Quanser Aero, Aero 2, 3-DOF Helicopter, and Autonomous Vehicles Research Studio (AVRS) platforms. The reviewed studies are compared according to the RL algorithm, control objective, hardware configuration, implementation environment, and reported experimental performance. The synthesis shows that Aero and Aero 2 are used primarily for stabilisation, trajectory tracking, and energy-aware control, whereas the 3-DOF Helicopter provides a more demanding benchmark for adaptive and Actor–Critic methods under nonlinear and coupled dynamics. AVRS offers significant potential for vision-based and multi-agent RL; however, the available experimental literature remains limited. Across the reviewed comparisons, policy-gradient and Actor–Critic methods, including PPO and SAC, generally demonstrate greater adaptability and smoother continuous-control behaviour, while conventional controllers frequently retain advantages in steady-state accuracy, computational predictability, and safety verification. Nevertheless, no RL algorithm can be identified as universally superior because the published studies employ heterogeneous reward functions, reference trajectories, sampling rates, performance measures, and hardware configurations. Recurring limitations include sample inefficiency, simulation-to-hardware discrepancies, computational latency, safety constraints, and incomplete reporting of experimental protocols. The review therefore identifies standardised evaluation procedures, reproducible reporting, safety-aware RL, interoperable software interfaces, and higher-fidelity digital twins as priorities for future research. No new experimental data are generated; the contribution is a comparative synthesis of experimental evidence reported in the literature. Full article

The increasing demand for sustainable outdoor lighting has accelerated the development of solar-powered lighting systems. However, conventional solar lamps typically employ fixed illumination levels and simple day–night switching mechanisms, resulting in inefficient battery utilization and limited adaptability to changing environmental conditions. This study proposes a TRIZ-guided intelligent solar-powered lighting system that integrates photovoltaic energy harvesting, adaptive pulse-width modulation (PWM)-based illumination control, ultrasonic sensing, wireless communication, and embedded control into a unified standalone platform. The TRIZ contradiction matrix was employed during the conceptual design stage to systematically resolve key engineering contradictions involving illumination performance, energy efficiency, hardware complexity, battery lifetime, and user convenience. The proposed prototype was developed using an AT89S51 microcontroller to coordinate battery charging protection, environmental sensing, adaptive brightness regulation, and manual wireless operation. Experimental validation demonstrated stable photovoltaic charging with a regulated battery charging voltage of 14.4 V, reliable execution of embedded control functions, seamless transition between manual and autonomous operating modes, and adaptive LED brightness regulation according to real-time environmental conditions. The integrated PWM control strategy reduced unnecessary energy consumption by dynamically adjusting illumination intensity based on object detection rather than maintaining constant full-power operation. The experimental results further verified the feasibility of combining software-driven adaptive control with renewable energy harvesting to achieve intelligent energy management without increasing hardware complexity. Overall, the proposed system demonstrates that the integration of TRIZ-based systematic innovation with embedded intelligent control provides a practical, energy-efficient, and cost-effective solution for autonomous outdoor lighting. The proposed architecture offers valuable engineering insights for future smart lighting applications in off-grid infrastructure, sustainable communities, and smart city environments. Full article

Generative AI’s application in hands-on engineering systems has gained substantial attention, particularly in complex hardware–software configurations that involve physical operational risks. Traditional configuration tools often lack real-time contextual awareness, limiting the diagnostic capabilities of novice users. To overcome such limitations, we developed an AI-driven system configuration and diagnostics engine by integrating a deterministic rule-validation layer with a probabilistic generative large language model, balancing factual reliability with adaptive conversational debugging. Using task–technology fit (TTF) theory, the engine’s functional alignment was evaluated through an experimental study involving 100 participants. Structural equation modeling results showed that TTF significantly enhances user learning satisfaction and overall diagnostic learning outcomes. Accounting for 86.3% of the variance in diagnostic performance (R² = 0.863), the developed engine presents high effectiveness in supporting real-time, interactive engineering diagnostics. Full article

Accurate indoor localization is a fundamental enabling technology for modern smart environments and Location-Based Services (LBS). Among the various indoor positioning technologies, Bluetooth Low Energy (BLE) has emerged as a popular solution due to its low cost and low power consumption. However, BLE signals suffer from severe environmental noise, while continuously executing complex positioning models can quickly deplete the battery resources of mobile devices. To address both problems jointly, this paper proposes a robust and sustainable machine learning framework for indoor localization on mobile devices. The framework first applies a discrete-time Kalman Filter that suppresses noise induced by walls and moving human bodies, and then benchmarks 12 machine learning models (including a Neural Network baseline) on a real-world dataset of 15,000 samples from 10 smartphones under a $5\times 3$ repeated cross-validation (CV) protocol. To identify the best model for mobile deployment, we introduce the Green Efficiency Index (GEI), which balances positioning accuracy against software-estimated energy consumption in Joules. Results show that the evaluated Multi-Layer Perceptron (MLP) baseline struggles with the noisy BLE data, producing a Mean Absolute Error (MAE) of 1.81 m, whereas the evaluated tree-based models map indoor spatial patterns far more accurately. K-Nearest Neighbors (KNN) achieves the lowest MAE at 1.175 m but requires substantial memory, making it unsuitable for sustainable mobile deployment. Extra Trees therefore emerges as the optimal solution, achieving an MAE of 1.347 m with low energy consumption and a compact memory footprint. The framework also normalizes hardware differences across all 10 tested smartphones, providing an accurate, energy-efficient, and generalizable solution for indoor localization. Full article

The practicality of post-quantum lattice-based schemes is crucial for their real-world applications. Integrating these schemes requires efficient implementations through hardware, software, and algorithmic optimisation to achieve the necessary speed and resource capability. This paper aims to improve arithmetic operations in lattice-based cryptography by accelerating the Number Theoretic Transform (NTT/INTT). It optimises the transform’s main bottleneck, the twiddle-factor modular multiplication within the butterfly unit, by replacing it with constant modular multiplication derived from Barrett reduction. We introduce two constant multipliers: the Constant Barrett and a proposed Truncated-Modulus-Size Constant Barrett (TMSCB) variant, which pre-computes each twiddle constant together with its reciprocal, eliminating Barrett’s data dependency and enabling area–time trade-offs. A comprehensive evaluation of the proposed constant modular multiplication is conducted against the classical Barrett multiplication, incorporating analytical complexity analysis and experimental quantitative analysis using FPGA hardware design. The proposed optimisation technique is deployed in the hardware design of the NTT/INTT for the ML-DSA and Falcon parameter sets with optimal use of the DSP cores. Performance comparisons with state-of-the-art implementations of ML-DSA NTT show 46.73% and 29.82% execution time improvements and 17% and 35.4% area resource reductions using single- and dual-butterfly units, respectively. On the Falcon, our design achieves an execution time improvement of at least 27.6%, with area savings across several butterfly configurations. These results validate the effectiveness of the Constant Barrett optimisation technique for accelerating the NTT/INTT, paving the way for more efficient implementations in other applications. Full article

Purpose: While magnetic resonance-guided radiation therapy (MRgRT) may provide significant clinical advantages, delivery times tend to be longer than other radiation therapy (RT) modalities and published data on these times are limited. We evaluated longitudinal treatment times across our 8-year institutional MRgRT experience. Methods/Materials: A retrospective analysis of patients treated at our institution on a 0.35-Tesla MR-Linac between April 2018 and April 2026 was performed. All fractions were delivered with continuous intrafraction cine-MRI, soft tissue tracking, automatic beam gating, and online adaptive radiation therapy (oART) when indicated. The primary objective was to characterize changes in total in-room time (TIRT), treatment delivery time (TDT), and total adaptive time (TAT). For analysis of efficiency gains, our overall experience was separated into early (2018–2022) and late (2022–2026) periods. Results: A total of 1026 patients, 1203 treatment courses, and 7665 fractions were included. The median age was 69 years (range: 19–94) and the most commonly treated sites by treatment course were pancreas (n = 430; 35.7%), thorax (n = 203; 16.9%), abdominopelvic lymph nodes (n = 181; 15.0%), liver (n = 138; 11.5%), and adrenal gland (n = 83; 6.9%). The median prescription dose was 50 Gy (range: 16.0–76.0) in a median of five fractions (range: 1–36). Breath-hold and stereotactic body radiation therapy (SBRT) were used in 87.9% and 88.5% of treatment courses, respectively. From the early (2018–2022) to late (2022–2026) study period, the proportion of fractions utilizing SBRT (49.9% vs. 81.1%; p < 0.001), oART (32.9% vs. 80.2%; p < 0.001), respiratory gating (75.9% vs. 88.8%; p < 0.001), and elective nodal coverage (61.8% vs. 78.2%; p < 0.001) increased, as did the proportion of pancreatic treatments (20.4% vs. 38.4%; p < 0.001) and the utilization of single-fraction courses (3.4% vs. 10.3%; p < 0.001). Despite this increasing complexity between study periods, median oART TIRT decreased from 67.0 to 48.0 min (28.4% reduction; p < 0.001), driven primarily by a reduction in TAT from 20.0 to 7.0 min, p < 0.001. Pancreatic oART fractions showed the greatest improvement (median TIRT, 70.0 to 47.0 min; p < 0.001). In 2022–2026, 78.1% of oART fractions were completed within 60 min vs. 35.5% in 2018–2022 (p < 0.001). Conclusions: Ablative MRgRT, with or without oART, can often be delivered in 60 min or less, including for mobile and anatomically unfavorable tumors. Future software and hardware advances are expected to further improve MRgRT treatment efficiency. Full article

Wearable electrocardiographic technologies have become increasingly important for continuous cardiac monitoring; however, most existing portable systems are limited by the number of recorded leads or provide only basic signal acquisition without advanced automated analysis. This study presents a third-generation wearable twelve-lead electrocardiographic platform developed for multilead ECG acquisition and automated spatial ECG analysis. Compared with the previous generation, the hardware modification primarily consists of architectural consolidation: functions previously distributed across an STM32 microcontroller and separate wireless communication modules are integrated into a single ESP32-S3-based architecture, while the ECG acquisition principle, ten-electrode configuration, and sampling rate remain unchanged. The main methodological contribution of the present work is the software pipeline for lead-specific ST₈₀ measurement and analysis of ST-segment deviations across anatomically contiguous leads. The system uses an ADS1298 analog front-end for synchronized multichannel ECG acquisition. The host software performs digital preprocessing, R-peak detection, ECG feature extraction, twelve-lead reconstruction, lead-specific ST₈₀ measurement, contiguous-lead analysis, and generation of a preliminary computer-assisted ECG report. The developed platform was clinically evaluated using sequential recordings acquired with the proposed system and a reference clinical electrocardiograph. Quantitative comparison of automated PR, QRS, QT, and QTc measurements in 30 paired recordings demonstrated positive correlations with the reference BTL Flexi 12 ECG (r = 0.756–0.820, all p < 0.001), with mean absolute errors ranging from 2.53 ms for QRS duration to 10.40 ms for the QT interval. The system successfully recorded diagnostically interpretable twelve-lead ECGs in all participants and produced stable signal quality suitable for clinical assessment. The software automatically identified ECG waves and intervals, reconstructed twelve-lead recordings, evaluated ST-segment deviations across individual leads, and localized ischemia-related changes according to standard anatomical lead groups. Integration of signal acquisition, processing, visualization, and automated interpretation into a single telemedicine-oriented platform reduced hardware complexity while maintaining reliable multichannel ECG monitoring. The proposed wearable platform demonstrates the feasibility of combining compact embedded hardware with automated multilead ECG analysis for remote cardiovascular monitoring. The presented architecture provides a practical foundation for telemedicine applications and may support earlier recognition of clinically significant electrocardiographic abnormalities during ambulatory monitoring. Full article

Software-defined radio (SDR) has become an attractive platform for radar research due to its flexibility, compact size, and ability to reconfigure radar methods for different situations. However, achieving high-resolution time-domain radar with SDR remains challenging due to demanding requirements for data throughput, sampling rate, and hardware configuration. Frequency-modulated radar implementations face similar limitations, often suffering from slow data acquisition and increased system complexity including exact waveform linearity, phase coherence, frequency-stepping control, and computationally intensive signal processing. In this work, a high-resolution impulse radar is implemented on a USRP X310 equipped with a UBX-160 daughterboard. The system fully exploits the hardware’s wide instantaneous bandwidth, enabling high range resolution in real time. Detailed hardware requirements and software configurations are presented step-by-step to ensure reliable operation at high data rates. To maximize performance and flexibility, the radar system is implemented using the low-level USRP Hardware Driver C++ API rather than graphical programming frameworks. This enables improved memory handling, reduced processing overhead, and precise timing control. The resulting modular design permits adaptability for different radar modalities with minimal code modification. The proposed platform provides a practical and reproducible framework for developing wideband, real-time SDR-based radar systems. Full article

This study focuses on accelerating adaptive beamforming in the Multicarrier Wireless Internet Local Loop (McWiLL) professional radio communication system using “Korona” ring smart antennas. The work investigates algorithms for calculating complex weight coefficients in an eight-element uniform circular antenna array. The main objective is to reduce beam pattern adaptation time while maintaining interference suppression depth and robustness under multipath propagation. To achieve this, an ensemble machine learning approach based on the Random Forest algorithm is employed to approximate the optimal Minimum Variance Distortionless Response (MVDR) solution using elements of the sample covariance matrix of received signals. The training dataset is generated through McWiLL channel simulations considering mutual coupling between array elements, signal-to-noise ratio (SNR) variation, and different angles of arrival of the desired and interfering signals. The proposed method is evaluated against the classical MVDR algorithm in terms of radiation pattern null depth, robustness to phase distortions, and inference time on a Field-Programmable Gate Array (FPGA) hardware platform. Results demonstrate that the Random Forest-based approach achieves more than a fourfold reduction in computation time while forming radiation-pattern nulls of about 30–35 dB toward the interferers (versus 44–46 dB for the classical MVDR); the synthesized core uses no hardware multipliers (DSP48), and its functional equivalence to the software model is confirmed by bit-exact RTL co-simulation. The findings show promise for deployment in McWiLL base stations and other professional radio systems requiring fast, adaptive beamforming under dynamic channel conditions. Full article

The three-dimensional (3D) documentation of the Similaun mummy (Ötzi the Iceman) and the associated Copper Age equipment and clothing presents unique challenges due to diverse material properties and strict conservation constraints. This study presents a comprehensive digital preservation workflow, primarily utilizing Structure from Motion (SfM) close-range photogrammetry (a method that reconstructs precise 3D geometry from overlapping 2D digital photographs), integrated with Image-Based Modeling (IBM) and Neural Radiance Field (NeRF) algorithms (a machine learning approach that models a complex scene as a continuous volumetric function method). To overcome the non-Lambertian properties of the mummy’s protective ice layer and wet skin (surfaces that reflect light specularly rather than diffusely, creating glares that can disorient standard reconstruction algorithms), a specialized Polarized Light Photography (PLP) strategy was implemented using custom-built hardware. This integration required advanced anatomical segmentation to resolve postural discrepancies caused by taphonomic processes. The resulting web-based application provides the scientific community with a metrically accurate digital twin, featuring interactive tools for cross-sectioning and X-ray visualization. By adopting a Free/Libre and Open-Source Software (FLOSS) ecosystem, this project establishes a sustainable, modular framework for future forensic investigations and diachronic monitoring, ensuring the long-term digital life of one of the world’s most significant archaeological finds. Full article

Water scarcity in semi-arid regions makes efficient irrigation scheduling a priority, yet farm-level adoption of soil-moisture monitoring remains limited by the cost, low portability and installation complexity of commercial sensing systems. This study presents the development and field validation of WaziSense, a low-cost, solar-powered Internet-of-Things (IoT) smart tensiometer, developed within the OSIRRIS platform for soil-water monitoring and irrigation scheduling. The device couples a Watermark granular-matrix sensor and a DS18B20 temperature probe to an ATmega328P microcontroller (Arduino Pro-Mini, 3.3 V, 8 MHz) with long-range LoRa communication and a maximum-power-point-tracking (MPPT) solar-charging stage, logging soil matric potential and soil temperature every 15 min. An open-source edge/cloud stack (WaziGate, WaziApp) retrieves weather forecasts from an open API and runs an automated machine learning (AutoML) regression pipeline that forecasts soil-water dynamics and the time to a user-defined threshold, from which irrigation is scheduled and its applied volume verified by a flow meter. The system was deployed at three bioclimatic sites in Tunisia (durum wheat at Cherfech, citrus at Nabeul, apple at Sbeitla), with tensiometers installed at 20 and 40 cm depths, and validated against commercial 10HS capacitive probes coupled to a ZL6 data logger, with which the co-located readings were significantly correlated (r = 0.81). Calibrated readings showed a strong relationship between soil–water content and soil–water potential (R² = 0.99), and the edge forecasting model reproduced soil–water dynamics on unseen data (Sbeitla apple site, 5-day horizon) with R² = 0.73, RMSE = 0.35, MAE = 0.23 and MPE = 12.52%. With a material cost under about 90 EUR per node and fully open-source hardware and software, WaziSense is one to two orders of magnitude cheaper than commercial monitoring stations, offering an affordable, reproducible and scalable tool for data-driven irrigation in water-limited agriculture. Full article

The latest generation of 4D imaging radar demonstrates significant potential in autonomous driving environmental perception, leveraging its capability to provide target elevation data and dense point clouds. This paper introduces a complete method for constructing a multimodal 4D imaging radar dataset for three-dimensional traffic scenes. It illustrates the hardware and software configurations of the data-acquisition vehicle. Methods including multi-sensor coordination, parameter calibration, timestamp synchronization and spatial datum synchronization are proposed. And eight typical three-dimensional traffic scenarios are designed, such as rainy weather environments, dense heterogeneous targets, enclosed tunnels, high-speed cut-in of multiple vehicles, multi-layered stereoscopic structures and edge working condition reproduction. In addition, this paper puts forward a frame-by-frame processing method for high-resolution images and point cloud data collected by the high-definition camera-LiDAR-4D imaging radar collaborative system. A large model-based 3D annotation method for multiple types of targets is proposed, generating a spatio-temporal sequence-optimized four-dimensional annotation sequence, and finally constructs a complete and high-quality multimodal 4D imaging radar dataset for three-dimensional traffic scenes. The results show that the constructed dataset enables the synchronization of timestamps and spatial coordinate systems. The large model can achieve high-precision 3D annotation for the four predefined target types. The dataset contains 11,400 frames of data from high-definition cameras, LiDAR, and 4D imaging radar, with 131,642 labels. This study will provide reliable fundamental support for the training and verification of 4D imaging radar perception algorithms, vehicle decision-making and planning in complex scenarios, and multi-sensor fusion technologies. Full article

Ambient Assisted Living (AAL) environments aim to enable older people to remain active and lead an autonomous and independent life for as long as possible. Among the technologies that can be incorporated into these settings, socially assistive robots (SAR) seek to establish a more natural and intuitive means of interaction with people, whilst helping them to carry out everyday tasks. One of the main challenges facing the design of these robots is how to enable them to undertake more complex tasks. Recent advances in Large Language Models (LLMs) have opened new avenues for flexible robot deliberation, yet their integration into real-time robotic systems remains challenging due to latency constraints, reasoning reliability, and the complexity of coordinating multi-step tasks. This paper proposes a hierarchical multi-agent architecture for robot deliberation that addresses these challenges by combining LLM-based planning with structured execution mechanisms within the ROS 2 ecosystem. The proposed architecture employs a supervisor agent that decomposes high-level natural language instructions into prioritised subtasks, enabling a priority-driven execution model that dynamically adapts to task relevance, temporal constraints, and environmental feedback. Subtasks are delegated to a set of Single-Purpose Agents (SPAs), orchestrated via LangGraph state machines and coordinated through a priority-aware scheduling mechanism. A key design principle is the use of Behaviour Trees (BTs) as high-level callable tools through the Model Context Protocol (MCP), encapsulating closed-loop control strategies while enabling preemptive and priority-consistent execution. This reduces the number of LLM inference steps required per task and improves robustness under dynamic conditions. A further contribution concerns the deployment of fine-tuned, lightweight LLMs—on the order of 0.6 billion parameters—specifically adapted for both the supervisor and the individual SPA roles through parameter-efficient low-rank adaptation (LoRA). These models are trained on role-specific tool-calling datasets to specialise in constrained reasoning patterns and task-specific decision-making, enabling efficient, low-latency inference directly on edge hardware. The combination of fine-tuning and hierarchical priority control enhances both the determinism and responsiveness of the system while mitigating error propagation across agent interactions. The paper presents the full software architecture, a formal characterisation of the system as a priority-aware hierarchical policy over a graph of agent workflows, and an experimental evaluation in an Ambient Assisted Living scenario assessing task success rate, inference efficiency, responsiveness under competing priorities, and overall user experience. Because SPA execution is decoupled from the supervisor’s own reasoning loop, the architecture is designed to keep accepting, processing, and queuing new user queries while previously dispatched SPAs are still executing their tasks. Full article