Search Results (48)

Reliable analysis of remote photoplethysmography (rPPG) signals depends on identifying physiologically plausible pulses. Traditional approaches rely on clustering self-similar pulses, which can discard valid variability. Automating pulse quality assessment could capture the true underlying morphology while preserving physiological variability. In this manuscript, individual rPPG pulses were manually labelled as plausible, borderline and implausible and used to train multilayer perceptron classifiers. Two independent datasets were used to ensure strict separation between training and test data: the Vision-MD dataset (4036 facial videos from 1270 participants) and a clinical laboratory dataset (235 videos from 58 participants). Vision-MD data were used for model development with an 80/20 training–validation split and 5-fold cross-validation, while the clinical dataset served exclusively as an independent test set. A three-class model was evaluated achieving $F_1$-scores of 0.92, 0.24 and 0.79 respectively. Recall was highest for plausible and implausible pulses but lower for borderline pulses. To test separability, three pairwise binary classifiers were trained, with ROC-AUC > 0.89 for all three category pairs. When combining borderline and implausible pulses into a single class, the binary classifier achieved an $F_1$-score of 0.93 for the plausible category. Finally, usability analysis showed that automated labelling identified more usable pulses per signal than the previously used agglomerative clustering method, while preserving physiological variability. Full article

This study investigates a non-destructive, compact pulse-echo ultrasonic method that combines an external transmitter with a single receiving sensor to identify different surface treatments applied to cementitious materials. The primary objective was to evaluate whether treatment-induced acoustic changes could be reliably quantified using time-domain signal parameters. Three types of surface conditions were examined: untreated reference specimens (R), specimens treated with a standard lithium silicate solution (A), and those treated with an enriched formulation containing hexylene glycol (B) intended to enhance pore sealing via gelation. A broadband piezoelectric receiver collected the backscattered echoes, from which the maximum amplitude, root mean square (RMS) voltage, signal energy, and effective duration were extracted. Receiver operating characteristic (ROC) analysis was conducted to quantify the discriminative power of each parameter. The results showed excellent classification performance between groups involving the B-treatment (AUC ≥ 0.96), whereas the R vs. A comparison yielded moderate separation (AUC ≈ 0.61). Optimal cut-off values were established using the Youden index, with sensitivity and specificity exceeding 96% in the best-performing scenarios. The results demonstrate that a single-receiver, one-sided pulse-echo arrangement coupled with straightforward amplitude metrics provides a rapid, cost-effective, and field-adaptable tool for the quality control of silicate-surface treatments. By translating laboratory ultrasonics into a practical on-site protocol, this study helps close the gap between the experimental characterisation and real-world implementation of surface-treatment verification. Full article

In complex urban electromagnetic environments, wireless positioning signals are subject to various types of interference, including narrowband, chirp, and pulse jamming. Traditional generic suppression methods struggle to achieve global optimization tailored to specific interference mechanisms. This paper proposes a classification-driven differentiated jamming suppression (CDDJ) method, which adaptively selects the optimal mitigation strategy by pre-identifying interference types and integrating classification confidence levels. First, the theoretical bounds of the output carrier-to-noise ratio ($C/N_{0_{out}}$) under typical interference scenarios are derived, characterizing the performance distribution of anti-jamming efficiency ($\mathsf{\Gamma }$). Then, a mapping relationship between interference categories and their corresponding suppression strategies is established, along with decision criteria for strategy switching based on signal quality evaluation metrics. Finally, an OpenMax-Lite rejection layer is designed to handle low-confidence inputs, identify unknown jamming using the Weibull distribution, and implement a broadband conservative suppression policy. Simulation results demonstrate that the proposed method exhibits significant advantages across different interference types. Under high JSR conditions, the signal recovery rate improves by over 10% and 8% compared to that of the WPT and KLT methods, respectively. In terms of SINR performance, the proposed approach outperforms the AFF, TDPB, and FDPB methods by 1.5 dB, 1.1 dB, and 5.3 dB, respectively, thereby enhancing the reliability of wireless positioning in complex environments. Full article

Chronic pain is prevalent and disproportionately impacts adults with a lower quality of life. Although subjective self-reporting is the “gold standard” for pain assessment, tools are needed to objectively monitor and account for inter-individual differences. This study introduced a novel framework to objectively classify pain intensity levels using physiological signals during Quantitative Sensory Testing sessions. Twenty-four participants participated in the study wearing physiological sensors (blood volume pulse (BVP), galvanic skin response (GSR), electromyography (EMG), respiration rate (RR), skin temperature (ST), and pupillometry). This study employed two analysis plans. Plan 1 utilized a grid search methodology with a 10-fold cross-validation framework to optimize time windows (1–5 s) and machine learning hyperparameters for pain classification tasks. The optimal time windows were identified as 3 s for the pressure session, 2 s for the pinprick session, and 1 s for the cuff session. Analysis Plan 2 implemented a leave-one-out design to evaluate the individual contribution of each sensor modality. By systematically excluding one sensor’s features at a time, the performance of these sensor sets was compared to the full model using Wilcoxon signed-rank tests. BVP emerged as a critical sensor, significantly influencing performance in both pinprick and cuff sessions. Conversely, GSR, RR, and pupillometry demonstrated stimulus-specific sensitivity, significantly contributing to the cuff session but with limited influence in other sessions. EMG and ST showed minimal impact across all sessions, suggesting they are non-critical and suitable for reducing sensor redundancy. These findings advance the design of sensor configurations for personalized pain management. Future research will focus on refining sensor integration and addressing stimulus-specific physiological responses. Full article

Cementitious materials are susceptible to damage not only from mechanical loading, but also from environmental (physical, chemical, and biological) factors. For Textile-Reinforced Cementitious (TRC) composites, durability poses a significant challenge, and a reliable method to assess long-term performance is still lacking. Among various durability attacks, freeze–thaw can induce internal cracking within the cementitious matrix, and weaken the textile–matrix bond. Such cracks result from hydraulic, osmotic, and crystallization pressure arising from the thermal cycles, leading to a reduction in the stiffness in the TRC composites. Early detection of freeze–thaw deterioration can significantly reduce the cost of repair, which is only possible through periodic, full-field monitoring of the composite. Full-field monitoring provides a comprehensive view of the damage distribution, offering valuable insights into the causes and progression of damage. The crack location, size, and pattern give more information than that offered by single-point measurement. While visual inspections are commonly employed for crack assessment, they are often time-consuming. Technological advances now enable crack pattern classification based on high-quality surface images; however, these methods only provide information limited to the surface. Elastic wave-based non-destructive testing (NDT) methods are highly sensitive to the material’s mechanical properties, and therefore are widely used for damage monitoring. On the other hand, electromagnetic wave-based NDTs offer the advantage of fast, non-contact measurements. Micro- and millimeter wave frequencies offer a balance of high resolution and wave penetration, although they have not yet been sufficiently explored for detecting damage in cementitious composites. In this study, TRC specimens were subjected to up to 150 freeze–thaw cycles and monitored using a combination of active elastic and electromagnetic wave-based NDT mapping methods. For this purpose, transmission measurements were conducted at multiple points, with ultrasonic pulse velocity (UPV) employed as a benchmark and, for the first time, millimeter wave (MMW) spectrometry applied. This multi-modal mapping approach enabled the tracking of damage progression, and the identification of degraded zones. Full article

The hindering of Global Navigation Satellite Systems (GNSS) signal reception by jamming and spoofing attacks degrades the signal quality. Careful attention needs to be paid when post-processing the signal under these circumstances before feeding the signal into the GNSS receiver’s post-processing stage. The identification of the time domain statistical attributes and the spectral domain characteristics play a vital role in analyzing the behaviour of the signal characteristics under various kinds of jamming attacks, spoofing attacks, and multipath scenarios. In this paper, the signal records of five disruptions (pure, continuous wave interference (CWI), multi-tone continuous wave interference (MCWI), multipath (MP), spoofing, pulse, and chirp) are examined, and the most influential features in both the time and frequency domains are identified with the help of explainable AI (XAI) models. Different Machine learning (ML) techniques were employed to assess the importance of the features to the model’s prediction. From the statistical analysis, it has been observed that the usage of the SHapley Additive exPlanations (SHAP) and local interpretable model-agnostic explanations (LIME) models in GNSS signals to test the types of disruption in unknown GNSS signals, using only the best-correlated and most important features in the training phase, provided a better classification accuracy in signal prediction compared to traditional feature selection methods. This XAI model reveals the black-box ML model’s output prediction and provides a clear explanation of the specific signal occurrences based on the individual feature contributions. By using this black-box revealer, we can easily analyze the behaviour of the GNSS ground-station signals and employ fault detection and resilience diagnosis in GNSS post-processing. Full article

Increasingly, humans are exposed to different activities at work, at home, and in general in their daily lives that generate episodes of stress. In many cases, these episodes could produce disorders in their health and reduce their quality of life. For this reason, it is crucial to implement mechanisms that can detect stress in individuals and develop applications that provide feedback through various activities to help reduce stress levels. Physiological parameters, such as galvanic skin response (GSR) and heart rate (HR) are indicative of stress-related changes. There exist methodologies that use wearable sensors to measure these stress levels. In this study, a sensor of blood volume pulse (BVP) and an electrocardiography (ECG) sensor were utilized to obtain metrics like heart rate variability (HRV) and pulse arrival time (PAT). Their features were extracted, processed, and analyzed for anxiety detection. The classification performance was evaluated using decision trees, a support vector machine (SVM), and meta-classifiers to accurately distinguish between “stressed” and “non-stressed” states. We obtained the best results with the SVM classifier using all the features. Additionally, we found that the ECG AD8232 sensor provided more reliable data compared to the photoplethysmography (PPG) signal obtained from the MAX30100 sensor. Therefore, the ECG is a more accurate tool for assessing emotional states related to stress and anxiety. Full article

Radar radiation source recognition is critical for the reliable operation of radar communication systems. However, in increasingly complex electromagnetic environments, traditional identification methods face significant limitations. These methods often struggle with high noise levels and diverse modulation types, making it difficult to maintain accuracy, especially when the Signal-to-Noise Ratio (SNR) is low or the available training data are limited. These difficulties are further intensified by the necessity to generalize in environments characterized by a substantial quantity of noisy, low-quality signal samples while being constrained by a limited number of desirable high-quality training samples. To more effectively address these issues, this paper proposes a novel approach utilizing Model-Agnostic Meta-Learning (MAML) to enhance model adaptability in few-shot learning scenarios, allowing the model to quickly learn with limited data and optimize parameters effectively. Furthermore, a multimodal fusion neural network, DCFANet, is designed, incorporating residual blocks, squeeze and excitation blocks, and a multi-scale CNN, to fuse I/Q waveform data and time–frequency image data for more comprehensive feature extraction. Our model enables more robust signal recognition, even when the signal quality is severely degraded by noise or when only a few examples of a signal type are available. Testing on 13 intra-pulse modulated signals in an Additive White Gaussian Noise (AWGN) environment across SNRs ranging from −20 to 10 dB demonstrated the approach’s effectiveness. Particularly, under a $5-way5-shot$ setting, the model achieves high classification accuracy even at −10dB SNR. Our research underscores the model’s ability to address the key challenges of radar emitter signal recognition in low-SNR and data-scarce conditions, demonstrating its strong adaptability and effectiveness in complex, real-world electromagnetic environments. Full article

The modulation classification technology for radar intra-pulse signals is important in the electronic countermeasures field. As the high quality labeled radar signals are difficult to be captured in the real applications, the signal modulation classification base on the limited number of labeled samples is playing a more and more important role. To relieve the requirement of the labeled samples, many self-supervised learning (SeSL) models exist. However, as they cannot fully explore the information of the labeled samples and rely significantly on the unlabeled samples, highly time-consuming processing of the pseudo-labels of the unlabeled samples is caused. To solve these problems, a supervised learning (SL) model, using the contrastive learning (CL) method (SL-CL), is proposed in this paper, which achieves a high classification accuracy, even adopting limited number of labeled training samples. The SL-CL model uses a two-stage training structure, in which the CL method is used in the first stage to effectively capture the features of samples, then the multilayer perceptron is applied in the second stage for the classification. Especially, the supervised contrastive loss is constructed to fully exploring the label information, which efficiently increases the classification accuracy. In the experiments, the SL-CL outperforms the comparison models in the situation of limited number of labeled samples available, which reaches 94% classification accuracy using 50 samples per class at 5 dB SNR. Full article

Accurately classifying the intra-pulse modulations of radar emitter signals is important for radar systems and can provide necessary information for relevant military command strategy and decision making. As strong additional white Gaussian noise (AWGN) leads to a lower signal-to-noise ratio (SNR) of received signals, which results in a poor classification accuracy on the classification models based on deep neural networks (DNNs), in this paper, we propose an effective denoising method based on a denoising diffusion probabilistic model (DDPM) for increasing the quality of signals. Trained with denoised signals, classification models can classify samples denoised by our method with better accuracy. The experiments based on three DNN classification models using different modal input, with undenoised data, data denoised by the convolutional denoising auto-encoder (CDAE), and our method’s denoised data, are conducted with three different conditions. The extensive experimental results indicate that our proposed method could denoise samples with lower values of the SNR, and that it is more effective for increasing the accuracy of DNN classification models for radar emitter signal intra-pulse modulations, where the average accuracy is increased from around 3 to 22 percentage points based on three different conditions. Full article

A gearbox compound fault intelligent diagnosis method based on the period group sparse model is proposed for the problem that the fault features are coupled with each other and the fault components are superimposed on each other and difficult to be separated in the gearbox compound fault signal. Firstly, a binary sequence is constructed to embed the fault pulse period as a priori knowledge into the group sparse model to decouple and separate the composite fault signal while maintaining the amplitude and sparsity of the extracted features. Secondly, the wavelet packet energy features of the decoupled signals are extracted to improve the data quality while enhancing the characterization ability of the dictionary in the classification model. Finally, the wavelet packet energy features are imported into the sparse dictionary classification model, and the fault diagnosis is completed by outputting the fault categories using the self-driven characteristics of the data. The results show that the fault identification accuracy using the proposed method is 97%. In addition, the experimental validation under different states and working conditions with different rotational speeds allows the superiority and effectiveness of the algorithm in this paper to be tested and has the feasibility of a practical application in engineering. Full article

This study aimed to investigate the impact of Pulsed Electric Fields (PEF) technology in the extraction of olive oil on a pilot scale, using the “Galega Vulgar” olive variety as raw material. The extraction assisted by PEF had a malaxation time of 30 min and was compared with the traditional process of 45 min of malaxation. The main quality parameters of olive oil and the PEF’s cost-benefit assessment were performed. The incorporation of PEF in olive oil production reduced the malaxation stage by 33% without compromising the yield or extra-virgin classification. This efficiency leads to a potential 12.3% increase in annual olive oil production, with a 12.3% and 36.8% rise in revenue and gross profit, respectively. For small-scale production, the considerable upfront investment required for PEF equipment may be a challenge in terms of return on investment. In this scenario, opting for a renting scheme is the best economic solution, especially given the seasonal nature of olive oil production. In medium- to large-scale production, the investment in PEF is a sound investment since it is possible to achieve, with an equipment cost of EUR 450,000 and a production output of 5 tons per hour, an annual ROI of 20%. Full article

In ultrashort-pulsed laser processing, surface modification is subject to complex laser and scanning parameter studies. In addition, quality assurance systems for monitoring surface modification are still lacking. Automated laser processing routines featuring machine learning (ML) can help overcome these limitations, but they are largely absent in the literature and still lack practical applications. This paper presents a new methodology for machine learning classification of self-organized surface structures based on light microscopic images. For this purpose, three application-relevant types of self-organized surface structures are fabricated using a 300 fs laser system on hot working tool steel and stainless-steel substrates. Optical images of the hot working tool steel substrates were used to learn a classification algorithm based on the open-source tool Teachable Machine from Google. The trained classification algorithm achieved very high accuracy in distinguishing the surface types for the hot working steel substrate learned on, as well as for surface structures on the stainless-steel substrate. In addition, the algorithm also achieved very high accuracy in classifying the images of a specific structure class captured at different optical magnifications. Thus, the methodology proposed represents a simple and robust automated classification of surface structures that can be used as a basis for further development of quality assurance systems, automated process parameter recommendation, and inline laser parameter control. Full article

During 14–16 March 2021, a large-scale dust storm event occurred in the northern region of China, and it was considered the most intense event in the past decade. This study employs observation data for PM2.5 and PM10 from the air quality monitoring station, the HYSPLIT model, ground-based polarized Lidar networks, AGRI payload data from Fengyun satellites and CALIPSO satellite Lidar data to jointly explore and scrutinize the three-dimensional spatial and temporal characteristics of aerosol transport. Firstly, by integrating meteorological data for PM2.5 and PM10, the air quality is assessed across six stations within the Lidar network during the dust storm. Secondly, employing a backward trajectory tracking model, the study elucidates sources of dust at the Lidar network sites. Thirdly, deploying a newly devised portable infrared 1064 nm Lidar and a pulsed 532 nm Lidar, a ground-based Lidar observation network is established for vertical probing of transboundary dust transport within the observed region. Finally, by incorporating cloud imagery from Fengyun satellites and CALIPSO satellite Lidar data, this study revealed the classification of dust and the height distribution of dust layers at pertinent sites within the Lidar observation network. The findings affirm that the eastward movement and southward compression of the intensifying Mongolian cyclone led to severe dust storm weather in western and southern Mongolia, as well as Inner Mongolia, further transporting dust into northern, northwestern, and northeastern parts of China. This dust event wielded a substantial impact on a broad expanse in northern China, manifesting in localized dust storms in Inner Mongolia, Beijing, Gansu, and surrounding areas. In essence, the dust emanated from the deserts in Mongolia and northwest China, encompassing both deserts and the Gobi region. The amalgamation of ground-based and spaceborne Lidar observations conclusively establishes that the distribution height of dust in the source region ranged from 3 to 5 km. Influenced by high-pressure systems, the protracted transport of dust over extensive distances prompted a gradual reduction in its distribution height owing to sedimentation. The comprehensive analysis of pertinent research data and information collectively affirms the precision and efficacy of the three-dimensional aerosol monitoring conducted by the ground-based Lidar network within the region. Full article

The phase circulating current (PCC) of the parallel three-phase inverter systems dramatically affects the power quality and conversion efficiency of the power grid. In this paper, a composite suppression strategy is proposed to solve the PCC issue by using the sliding mode control (SMC) approach and improved virtual impedance droop control. Taking the commonly used 2-group parallel three-phase inverter as an example, an inter- and intra-classification model is established by analyzing the sources of PCC. In order to suppress the inter-PCC, the traditional virtual impedance droop control is given, following the improved substitute by combining SMC. And the variables of the bus voltage, Q-U loop, P-f loop, and the virtual-induced reactance are also introduced for the robust control of the impedance droop. On the other side, a SMC-based suppression approach is designed to solve the issue of the intra-PCC. Its idea is to introduce a regulation factor for the space vector pulse width modulation (SVPWM) so that the zero-sequence voltage can be eliminated and the influence of the intra-PCC can be relieved. Comparative simulations and experiments validate the effectiveness of the methods proposed in this paper. Full article