Sensors

Autonomous vehicles (AVs) depend on perception, localization, and mapping to interpret their surroundings and navigate safely. This paper reviews existing methodologies and best practices in these domains, focusing on object detection, object tracking, localization techniques, and environmental mapping strategies. In the perception module, we analyze state-of-the-art object detection frameworks, such as You Only Look Once version 8 (YOLOv8), and object tracking algorithms like ByteTrack and BoT-SORT (Boosted SORT). We assess their real-time performance, robustness to occlusions, and suitability for complex urban environments. We examine different approaches for localization, including Light Detection and Ranging (LiDAR)-based localization, camera-based localization, and sensor fusion techniques. These methods enhance positional accuracy, particularly in scenarios where Global Positioning System (GPS) signals are unreliable or unavailable. The mapping section explores Simultaneous Localization and Mapping (SLAM) techniques and high-definition (HD) maps, discussing their role in creating detailed, real-time environmental representations that enable autonomous navigation. Additionally, we present insights from our testing, evaluating the effectiveness of different perception, localization, and mapping methods in real-world conditions. By summarizing key advancements, challenges, and practical considerations, this paper provides a reference for researchers and developers working on autonomous vehicle perception, localization, and mapping.

In human–robot collaborative environments, the inherent complexity of shared operational spaces imposes dual requirements on process safety and task execution efficiency. To address the limitations of conventional approaches that decouple planning and control modules, we propose a hierarchical planning–control framework. The proposed framework explicitly incorporates path tracking constraints during path generation while simultaneously considering path characteristics in the control process. The framework comprises two principal components: (1) an enhanced Dynamic Window Approach (DWA) for the local path planning module, introducing adaptive sub-goal selection method and improved path evaluation functions; and (2) a modified Model Predictive Control (MPC) for the path tracking module, with a curvature-based reference state online changing strategy. Comprehensive simulation and real-world experiments demonstrate the framework’s operational advantages over conventional methods.

Chromatographic fingerprint technology has become the standard of quality control for traditional herbal medicines (THMs). But several issues are associated with the wavelength selection of the representative fingerprint, such as contradictory evaluation results at different wavelengths and the accurate quantification of each composition at one wavelength. These problems can be addressed by projection profiling. Projection profiling is a collection of all sample compositions at the maximum absorption wavelength after baseline correction. In this paper, eleven baseline correction algorithms are optimized by using the effective information factor (EI) as an indicator. The influence of different integration methods and wavelengths on analytical method validation and similarity analysis results are discussed in detail to clarify the advantages of the projection profiling. A total of 33 batches of Compound Licorice Tablets (CLTs) were used to show the influence of different wavelengths in a similarity evaluation. The results show that projection profiling is a better choice than any chromatogram at a certain wavelength, because projection profiling is more informative, accurate, and stable.

Compressive spectral imaging (CSI) methods aim to reconstruct a three-dimensional hyperspectral image (HSI) from a single or a few two-dimensional compressive measurements. Conventional CSIs use separate optical elements to independently modulate the light field in the spatial and spectral domains, thus increasing the system complexity. In addition, real applications of CSIs require advanced reconstruction algorithms. This paper proposes a low-cost color-coded compressive snapshot spectral imaging method to reduce the system complexity and improve the HSI reconstruction performance. The combination of a color-coded aperture and an RGB detector is exploited to achieve higher degrees of freedom in the spatio-spectral modulations, which also renders a low-cost miniaturization scheme to implement the system. In addition, a deep learning method named Focus-based Mask-guided Spectral-wise Transformer (F-MST) network is developed to further improve the reconstruction efficiency and accuracy of HSIs. The simulations and real experiments demonstrate that the proposed F-MST algorithm achieves superior image quality over commonly used iterative reconstruction algorithms and deep learning algorithms.

Unmanned aerial vehicles (UAVs) equipped with communication devices are easy to deploy and can serve as temporary mobile communication stations, covering dynamic ground users and establishing communication links. The proper planning of UAV paths can improve user coverage while maintaining communication quality. The traditional particle swarm optimization (PSO) algorithm exhibits strong global search capabilities; however, it suffers from challenges such as uneven search coverage owing to the random distribution of the initial particle positions. Additionally, improper parameter selection often leads to premature convergence to local optima. This study proposes a method for multi-UAV coverage of dynamic users by combining Tent chaotic mapping with PSO (Tent–PSO). By using Tent chaotic mapping to adjust a UAV’s speed and position, the method ensures that particles are evenly distributed in the search space. Simulation results demonstrate that Tent–PSO improves the coverage of the global optimal solution by 10% and increases throughput by at least 37%, confirming its effectiveness in covering dynamic user scenarios.

Citation: Mahmud, M.F.; Ahmed, M.R.; Potluri, P.; Fernando, A. Understanding the Design and Sensory Behaviour of Graphene-Impregnated Textile-Based Piezoresistive Pressure Sensors. Abstract: Graphene-based textile pressure sensors are emerging as promising candidates for wearable sensing applications due to their high sensitivity, mechanical flexibility, and low energy consumption. This study investigates the design, fabrication, and electrome-chanical behaviour of graphene-coated nonwoven textile-based piezoresistive pressure sensors, focusing on the impact of different electrode materials and fabrication techniques. Three distinct sensor fabrication methods-drop casting, electrospinning, and electro-spraying-were employed to impregnate graphene onto nonwoven textile substrates, with silver-coated textile electrodes integrated to enhance conductivity. The fabricated sensors were characterised for their morphology (SEM), chemical composition (FTIR), and elec-tromechanical response under cyclic compressive loading. The results indicate that the drop-cast sensors exhibited the lowest initial resistance (~0.15 kΩ) and highest sensitivity (10.5 kPa −1) due to their higher graphene content and superior electrical connectivity. Electro-spun and electro-sprayed sensors demonstrated increased porosity and greater resistance fluctuations, highlighting the role of fabrication methods in sensor performance. Additionally, the silver-coated knitted electrodes provided the most stable electrical response , while spun-bonded and powder-bonded nonwoven electrodes exhibited higher hysteresis and resistance drift. These findings offer valuable insights into the optimisation of graphene-based textile pressure sensors for wearable health monitoring and smart textile applications, paving the way for scalable, low-power sensing solutions.

Travel time information has become an essential component of everyday commuting. Without such information, schedule delays would increase, leading to inevitable losses in traveler utility. In Korea, dedicated short-range communication transponders that identify vehicles have been installed along signalized arterials to collect travel time data. By matching vehicle identifications at consecutive points, travel times can be measured. However, for travel time information to be effective, two types of data processing techniques are required: outlier filtering and travel time prediction. This study proposes algorithms to address both challenges. An outlier filtering algorithm based on the median-based confidence interval was developed, taking into account the travel time characteristics on suburban arterials with frequent entry and exit points. Additionally, a travel time prediction algorithm that integrates Long Short-Term Memory (LSTM) and Convolutional Neural Networks (CNNs), referred to as LSTM-CNN, was developed to capture both long-term trends and local patterns in travel time data. The implementation of these algorithms resulted in a 2.2% reduction in error rates under congested conditions compared to current practices. At the 4 km study site, the annual benefits from this error reduction could amount to USD 135,200.

As a typical two-dimensional material, graphene and its derivatives exhibit many excellent properties, such as large specific surface area, electrical properties, and stability. Along with its derivatives, particularly graphene oxide (GO) and reduced graphene oxide (rGO), graphene materials have been studied in various fields due to the presence of aromatic ring, free π-π electron and reactive functional groups. This review focuses firstly on the synthesis methods of graphene and its derivatives along with their properties, followed by a discussion of the applications of their served as functional units in electrochemical sensing. Finally, this review describes the challenges, strategies, and outlooks on future developments.

Wearable inertial measurement units (IMUs) have been widely used in human movement analysis outside the laboratory. However, the IMU-based orientation estimation remains challenging, particularly in scenarios involving relatively fast movements. Increased sampling rate has the potential to improve accuracy, but it also increases power consumption and computational complexity. The relationship between sampling frequencies and accuracies remains unclear. We thus investigated the specific influence of IMU sampling frequency on orientation estimation across a spectrum of movement speeds and recommended sufficient sampling rates. Seventeen healthy subjects wore IMUs on their thigh, shank, and foot and performed walking (1.2 m/s) and running (2.2 m/s) trials on a treadmill, and a motion testbed with an IMU was used to mimic high-frequency cyclic human movements up to 3.0 Hz. Four widely used IMU sensor fusion algorithms computed orientations at 10, 25, 50, 100, 200, 400, 800, and 1600 Hz and were compared with marker-based optical motion capture (OMC) orientations to determine accuracy. Results suggest that the sufficient IMU sampling rate for walking is 100 Hz, running is 200 Hz, and high-speed cyclic movements is 400 Hz. The accelerometer sampling rate is less important than the gyroscope sampling rate. Further, accelerometer sampling rates exceeding 100 Hz even resulted in decreased accuracy because excessive orientation updates using distorted accelerations and angular velocity introduced more error than merely using angular velocity. These findings could serve as a foundation to inform wearable IMU development or selection across a spectrum of human gait movement speeds.

Being a cornerstone of Industry 4.0, Asset Administration Shell (AAS) enables seamless integration and interaction among the physical and digital worlds. There are multiple different tools and technologies available for implementing AAS. The purpose of this study is to support the tool and technology selection decision of AAS modelers and implementers. For that purpose, we conducted a literature survey and identified four active tools, and in the study, we included all of them: AASX server, Eclipse BaSyx, FA3ST service, and NOVAAS. Using a comprehensive criteria list, we conducted a thorough comparison of the selected technologies. The comparison was made in two steps: first for initial learning exercises and second for a real case study where digital twins belong to real assets in a facility belonging to the petrochemical industry. Among the evaluated tools, Eclipse BaSyx demonstrated superior performance compared to the other three tools investigated in this study. Future research will focus on incorporating machine learning (ML) and deep learning (DL) models associated with the assets, leveraging datasets generated by the sensors installed on the system.

The focus on Speech Emotion Recognition has dramatically increased in recent years, driven by the need for automatic speech-recognition-based systems and intelligent assistants to enhance user experience by incorporating emotional content. While deep learning techniques have significantly advanced SER systems, their robustness concerning speaker gender and out-of-distribution data has not been thoroughly examined. Furthermore, standards for SER remain rooted in landmark papers from the 2000s, even though modern deep learning architectures can achieve comparable or superior results to the state of the art of that era. In this research, we address these challenges by creating a new super corpus from existing databases, providing a larger pool of samples. We benchmark this dataset using various deep learning architectures, setting a new baseline for the task. Additionally, our experiments reveal that models trained on this super corpus demonstrate superior generalization and accuracy and exhibit lower gender bias compared to models trained on individual databases. We further show that traditional preprocessing techniques, such as denoising and normalization, are insufficient to address inherent biases in the data. However, our data augmentation approach effectively shifts these biases, improving model fairness across gender groups and emotions and, in some cases, fully debiasing the models.

Yield maps and in-season forecasts help optimize agricultural practices. The traditional approaches to predicting yield during the growing season often rely on ground-based observations, which are time-consuming and labor-intensive. Remote sensing offers a promising alternative by providing frequent and spatially extensive information on crop development. In this study, we evaluated the feasibility of high-resolution satellite imagery for the early yield prediction of an under-investigated crop, Japanese squash (Cucurbita maxima), in a small farm in Hollister, California, over the growing seasons of 2022 and 2023 using vegetation indices, including the Normalized Difference Vegetation Index (NDVI) and the Soil-Adjusted Vegetation Index (SAVI). We identified the optimal time for yield prediction and compared the performances across satellite platforms (Sentinel-2: 10 m; PlanetScope: 3 m; SkySat: 0.5 m). Pearson’s correlation coefficient (r) was employed to determine the dependencies between the yield and vegetation indices measured at various stages throughout the squash growing season. The results showed that SkySat-derived vegetation indices outperformed those of Sentinel-2 and PlanetScope in explaining the squash yields (R2 = 0.75–0.76; RMSE = 0.8–1.9 tons/ha). Remote sensing showed very strong correlations with yield as early as 29 days after planting in 2022 and 37 and 76 days in 2023 for the NDVI and the SAVI, respectively. These early dates corresponded with the vegetative stages when the crop canopy became denser before fruit development. These findings highlight the utility of high-resolution imagery for in-season yield estimation and within-field variability detection. Detecting yield variability early enables timely management interventions to optimize crop productivity and resource efficiency, a critical advantage for small-scale farms, where marginal yield changes impact economic outcomes.

Background: Sensors for continuous glucose monitoring (CGM) are now commonly used by people with type 1 and type 2 diabetes. However, the response of these devices to potentially interfering nutritional, pharmaceutical, or endogenous substances is barely explored. We previously developed an in vitro test method for continuous and dynamic CGM interference testing and herein explore the sensitivity of the Abbott Libre2 (L2) and Dexcom G6 (G6) sensors to a panel of 68 individual substances. Methods: In each interference experiment, L2 and G6 sensors were exposed in triplicate to substance gradients from zero to supraphysiological concentrations at a stable glucose concentration of 200 mg/dL. YSI Stat 2300 Plus was used as the glucose reference method. Interference was presumed if the CGM sensors showed a mean bias of at least ±10% from baseline with a tested substance at any given substance concentration. Results: Both L2 and G6 sensors showed interference with the following substances: dithiothreitol (maximal bias from baseline: L2/G6: +46%/−18%), galactose (>+100%/+17%), mannose (>+100%/+20%), and N-acetyl-cysteine (+11%/+18%). The following substances were found to interfere with L2 sensors only: ascorbic acid (+48%), ibuprofen (+14%), icodextrin (+10%), methyldopa (+16%), red wine (+12%), and xylose (>+100%). On the other hand, the following substances were found to interfere with G6 sensors only: acetaminophen (>+100%), ethyl alcohol (+12%), gentisic acid (+18%), hydroxyurea (>+100%), l-cysteine (−25%), l-Dopa (+11%), and uric acid (+33%). Additionally, G6 sensors could subsequently not be calibrated for use after exposure to dithiothreitol, gentisic acid, l-cysteine, and mesalazine (sensor fouling). Conclusions: Our standardized dynamic interference testing protocol identified several nutritional, pharmaceutical and endogenous substances that substantially influenced L2 and G6 sensor signals. Clinical trials are now necessary to investigate whether our findings are of relevance during routine care.

Stroke is the fifth leading cause of death in Taiwan. In the process of stroke treatment, rehabilitation for gait recovery is one of the most critical aspects of treatment. The Gait Assessment and Intervention Tool (G.A.I.T.) is currently used in clinical practice to assess the gait recovery level; however, G.A.I.T. heavily depends on physician training and clinical judgment. With the advancement of technology, today’s small, lightweight inertial measurement unit (IMU) wearable sensors are rapidly revolutionizing gait assessment and may be incorporated into routine clinical practice. In this paper, we developed a gait data acquisition and analysis system based on IMU wearable devices, proposed a simple yet accurate calibration process to reduce the IMU drifting errors, designed a machine learning algorithm to obtain real-time coordinates from IMU data, computed gait parameters, and derived a formula for G.A.I.T. scores with significant correlation with the physician’s observational scores.

With the rapid advancement of intelligent driving technology, multi-source information fusion has become a vital topic in the field of environmental perception. To address the fusion deviation resulting from changes in sensor performance due to environmental variations, this paper proposes a multi-source information fusion algorithm based on the improved Sage-Husa adaptive extended Kalman filtering (SHAEKF) algorithm. First, a multi-source information fusion system is constructed based on the vehicle kinematic model and the sensor measurement model. Then, the Sage-Husa adaptive fading extended Kalman filtering (SHAFEKF) algorithm is constructed by introducing a fading factor into the SHAEKF algorithm to enhance the influence of newly incoming data. Finally, the experimental results indicate that the positional average errors of the algorithm in the two scenarios are 0.137 and 0.071. When compared to the SHAEKF algorithm, the positional average errors have been reduced by 2.8% and 13.4%, while the mean squared errors have decreased by 64% and 72%. This demonstrates that the SHAFEKF algorithm offers high accuracy and low fluctuation, enhancing its adaptability in multi-source information fusion systems.

The main goal of this study was to determine whether the type of spike can influence the final sprint result by comparing step by step the kinematics of four 50-m sprints. Twelve well-trained junior sprinters (ages 17-19) from the Polish National Team (ranging from 100 to 400 m) participated in the study, with personal bests in the 100-m sprint of 10.70 ± 0.19 s. The OptoJump Next-Microgate sensor measurement system (Optojump, Bolzano, Italy) was used to measure the essential kinematic sprinting variables. Following the sprint distance, photocells were placed on the track at the start, at 10 m, at 20 m, at 30 m, and at the finish (50 m). Fifty-meter sprints were completed alternately, two with classic and two with the carbon-plated spikes. For every sprinter, the order in which the spikes were chosen was randomized. To better understand the problem of variability in kinematic parameters, in addition to the actual statistics, the profile analysis process was applied. The analysis of the four 50 m sprints did not show significant differences between the kinematic parameters considering runs in both the classic Nike and carbon-plated Nike ZoomX Flymax spikes. It may be suggested that spikes' sole bending stiffness may not affect short-distance (up to 50-60 m) sprinting performance. From a practical point of view, training focused on maximum speed development can be carried out with both classic and carbon-plated spikes. Finally, our experiment can guide the preparation of a research methodology that assesses the effect of carbon-plated spikes on prolonged sprinting, e.g., 200-400 m.

Nighttime light remote sensing imagery is limited by its single band and low spatial resolution, hindering its ability to accurately capture ground information. To address this, a dual-sampling adjustment method is proposed to enhance nighttime light remote sensing imagery by fusing daytime optical images with nighttime light remote sensing imagery, generating high-quality color nighttime light remote sensing imagery. The results are as follows: (1) Compared to traditional nighttime light remote sensing imagery, the spatial resolution of the fusion images is improved from 500 m to 15 m while better retaining the ground features of daytime optical images and the distribution of nighttime light. (2) Quality evaluations confirm that color nighttime light remote sensing imagery enhanced by dual-sampling adjustment can effectively balance optical fidelity and spatial texture features. (3) In Beijing’s central business district, color nighttime light brightness exhibits the strongest correlation with business, especially in Dongcheng District, with r = 0.7221, providing a visual tool for assessing urban economic vitality at night. This study overcomes the limitations of fusing day–night remote sensing imagery, expanding the application field of color nighttime light remote sensing imagery and providing critical decision support for refined urban management.

Atmospheric refraction is a significant challenge to accurate distance and angle measurements in open-air environments, often limiting the precision of measurements obtained using electro-optic geodetic instruments despite their nominal accuracies. This study introduces a novel model, 3D-RM, designed to mitigate atmospheric effects on both distance and vertical angle measurements. The 3D-RM integrates in situ meteorological data from a network of automatic data-loggers, terrain information from a digital terrain model (DTM), and sensible heat flux from the fifth generation of European Centre for Medium-Range Weather Forecast reanalysis (ERA5), which is used in the application of the Turbulence Transfer Model (TTM) for estimating vertical refractivity gradients at various height levels. The model was tested with total station observations to 10 target points during two field campaigns. The results show that applying the model for distance correction leads to improvements in terms of closeness to reference values when compared to the standard method, which relies only on meteorological data collected at the station. Furthermore, the model has been additionally tested by removing the station meteorological data (3D-RM2). The results demonstrate that accurate corrections can be obtained even without the need of meteorological sensors specifically installed at the station point, which makes it more flexible. The 3D-RM is a cost-effective and relatively easy-to-implement solution, offering a promising alternative to existing methodologies, such as measuring meteorological values at both station and target points or the development of new instruments that can compensate the refractivity (such as a multiple-color electronic distance meter).

Personal protective equipment (PPE) is crucial for infection prevention and is effective only when worn correctly and consistently. Health organizations often use education or inspections to mitigate non-compliance, but these are costly and have limited success. This study developed a novel on-device, AI-based computer vision system to monitor healthcare worker PPE adherence in real time. Using a custom-built image dataset of 7142 images of 11 participants wearing various combinations of PPE (mask, gloves, gown), we trained a series of binary classifiers for each PPE item. By utilizing a lightweight MobileNetV3 model, we optimized the system for edge computing on a Raspberry Pi 5 single-board computer, enabling rapid image processing without the need for external servers. Our models achieved high accuracy in identifying individual PPE items (93–97%), with an overall accuracy of 85.58 ± 0.82% when all items were correctly classified. Real-time evaluation with 11 unseen medical staff in a cardiac intensive care unit demonstrated the practical viability of our system, maintaining a high per-item accuracy of 87–89%. This study highlights the potential for AI-driven solutions to significantly improve PPE compliance in healthcare settings, offering a cost-effective, efficient, and reliable tool for enhancing patient safety and mitigating infection risks.

This paper presents a ZnO-Pt/Ru sensor prepared by a two-step hydrothermal method with in situ-grown ZnO nanorods and doped with Pt and Ru elements by immersion sintering. Characterization results showed that Pt and Ru were successfully modified on the surface of ZnO nanorods. ZnO-Pt/Ru achieved a response of 25–50 ppm H2S at the optimum operating temperature of 198 °C. In addition, the lower limit of H2S detection of ZnO-Pt/Ru reached 50 ppb with a response of about 10%, indicating a wide concentration detection range. Due to the good catalytic properties of Pt, the recovery characteristics of ZnO at high concentrations of H2S were significantly improved. The response time of ZnO-Pt/Ru (30 s) was also significantly shorter than pristine ZnO (56 s), with excellent selectivity. As far as the gas-sensitive enhancement mechanism is concerned, at the macroscopic level, the ZnO surface was modified by Pt and Ru, and this special structure of ZnO-Pt/Ru significantly increased the specific surface area. At the microscopic level, the PN junction formed between Pt/Ru and ZnO provided abundant holes for electron migration.

The current methods for detecting coal gangue face several challenges, including low detection accuracy, a high probability of missed detections, and inadequate real-time performance. These issues stem from the complexities associated with diverse industrial environments and mining conditions, such as the mixing of coal gangue and insufficient illumination within coal mines. A detection model, referred to as EBD-YOLO, is proposed based on YOLOv11n. First, the C3k2-EMA module is integrated with the EMA attention mechanism within the C3k2 module of the backbone network, thereby enhancing the model’s feature extraction capabilities. Second, the introduction of the BiFPN module reduces computational complexity while enriching both semantic information and detail within the model. Finally, the incorporation of the DyHead detector head further enhances the model’s ability to express features in complex environments. The experimental results indicate that the precision (P) and recall (R) of the EBD-YOLO model are 88.7% and 83.9%, respectively, while the mean average precision (mAP@0.5) is 91.7%. These metrics represent increases of 3.4%, 3.7%, and 3.9% compared to those of the original model, respectively. Additionally, the frames per second (FPS) improved by 10.01%. Compared to the mainstream YOLO target detection algorithms, the EBD-YOLO detection model achieves the highest mAP@0.5 while maintaining superior detection speed. It exhibits a slight increase in computational load, despite an almost unchanged number of parameters, and demonstrates the best overall detection performance. The EBD-YOLO detection model effectively addresses the challenges of missed detections, false detections, and real-time detection in the complex environment of coal mines.

To push the boundaries of confocal microscopy beyond its current limitations by predicting sensor responses for complex surface geometries, we build digital twins using three rigorous models, the finite element method (FEM), Fourier modal method (FMM), and boundary element method (BEM) to model light–surface interactions. Fourier optics are then used to calculate the sensor signals at the back focal plane and at the detector. A 3D illumination model is applied to 2D periodic structures for FEM and FMM modelings and to 3D aperiodic structures for BEM modeling. The lateral and vertical scanning processes of the confocal microscope are achieved through focal-point shifts of the objective, using plane-wave illuminations with varying incident and azimuthal angles. This approach reduces the need for repeated, time-intensive rigorous simulations of the scattering process when a fine scanning is desired. Furthermore, we give an in-depth description of a novel confocal microscopy method using FMM. For rectangular grating surfaces, the three models yield identical, highly accurate results, as validated by measured results. Simulations of the instrument transfer function, tilted gratings, and gratings with edge rounding offer insights into some experimentally observed effects. This research therefore provides a promising approach for correcting systematic errors in confocal microscopy.

Railway tight and wide gauges are critical factors affecting the safety and reliability of railway systems. Undetected tight and wide gauges can lead to derailments, posing significant risks to operations and passenger safety. This study explores a novel approach to detecting railway tight and wide gauges by integrating accelerometer data, machine-learning techniques, and building information modeling (BIM). Accelerometers installed on axle boxes provide real-time dynamic data, capturing anomalies indicative of tight and wide gauges. These data are processed and analyzed using supervised machine-learning algorithms to classify and predict potential tight- and wide-gauge events. The integration with BIM offers a spatial and temporal framework, enhancing the visualization and contextualization of detected issues. BIM’s capabilities allow for the precise mapping of tight- and wide-gauge locations, streamlining maintenance workflows and resource allocation. Results demonstrate high accuracy in detecting and predicting tight and wide gauges, emphasizing the reliability of machine-learning models when coupled with accelerometer data. This research contributes to railway maintenance practices by providing an automated, data-driven methodology that enhances the proactive identification of tight and wide gauges, reducing the risk of derailments and maintenance costs. Additionally, the integration of machine learning and BIM highlights the potential for comprehensive digital solutions in railway asset management.

The Polarimetric Interferometric Synthetic Aperture Radar (PolInSAR) system is a combination of polarimetric SAR and interferometric SAR, which can simultaneously obtain the power information, polarimetric information, and interferometric information of land cover. Traditional land cover classification methods fail to fully utilize these information types, resulting in limited classification types and low accuracy. This paper proposes a PolInSAR land cover classification method that fuses power information, polarimetric information, and interferometric information, aiming to enrich the classification types and improve the classification accuracy. Firstly, the land cover is divided into strong scattering areas and weak scattering areas by using the power information to avoid the influence of weak scattering areas on the classification results. Then, the weak scattering areas are distinguished into shadows and water bodies by combining the interferometric information and image corners. For the strong scattering areas, the polarimetric information is utilized to distinguish vegetation, buildings, and bare soil. For the vegetation area, the concept of vegetation ground elevation is put forward. By combining with the anisotropy parameter, the vegetation is further subdivided into tall coniferous vegetation, short coniferous vegetation, tall broad-leaved vegetation, and short broad-leaved vegetation. The effectiveness of the method has been verified by the PolInSAR data obtained from the N-SAR system developed by Nanjing Research Institute of Electronics Technology. The overall classification accuracy reaches 90.2%, and the Kappa coefficient is 0.876.

A signal processing-based procedure is proposed for calibrating an experimental sensor-based test system used to identify material model parameters. A standard dynamic mechanical analyzer (DMA) sensorized test apparatus is considered, enabling the measurement of dynamic excitation and displacement response in a specimen under flexural conditions. To account for the dynamic contributions of the system frame and fixtures to the measured response, a novel calibration procedure is introduced, mainly differing from the techniques used in standard test applications. A multi-degree-of-freedom dynamic model of the instrument frame, coupled with the beam specimen under test, is considered, and a frame identification procedure is described. The proposed procedure requires measurements from at least three beam specimens made of a known material but with different geometries. It is shown that an accurate frame model can be identified using an algebraic numerical technique. It is shown that the accuracy of the material model identification can be improved by applying the proposed calibration technique. Some experimental application examples are presented and discussed.