Sensors, Volume 24, Issue 6 (March-2 2024) – 306 articles

Respiratory rate ($f_R$) monitoring through wearable devices is crucial in several scenarios, providing insights into well-being and sports performance while minimizing interference with daily activities. Strain sensors embedded into garments stand out but require thorough investigation for optimal deployment. Optimal sensor positioning is often overlooked, and when addressed, the quality of the respiratory signal is neglected. Additionally, sensor metrological characterization after sensor integration is often omitted. In this study, we present the design, development, and feasibility assessment of a smart t-shirt embedded with two flexible sensors for $f_R$ monitoring. Guided by a motion capture system, optimal sensor design and position on the chest wall were defined, considering both signal magnitude and quality. The sensors were developed, embedded into the wearable system, and metrologically characterized, demonstrating a remarkable response to both static (sensitivity 9.4$\mathsf{\Omega }\cdot {\%}^{-1}$ and 9.1$\mathsf{\Omega }\cdot {\%}^{-1}$for sensor A and sensor B, respectively) and cyclic loads (min. hysteresis span 20.4% at 36 bpm obtained for sensor A). The feasibility of the wearable system was assessed on healthy volunteers both under static and dynamic conditions (such as running, walking, and climbing stairs). A mean absolute error of 0.32 bpm was obtained by averaging all subjects and tests using the combination of the two sensors. This value was lower than that obtained using both sensor A (0.53 bpm) and sensor B (0.78 bpm) individually. Our study highlights the importance of signal amplitude and quality in optimal sensor placement evaluation, as well as the characterization of the embedded sensors for metrological assessment. Full article

The increasing focus on the development of positioning techniques reflects the growing interest in applications and services based on indoor positioning. Many applications necessitate precise indoor positioning or tracking of individuals and assets, leading to rapid growth in products based on these technologies in certain market sectors. Ultrasonic systems have already proven effective in achieving the desired positioning accuracy and refresh rates. The typical signal used in ultrasonic positioning systems for estimating the range between the target and reference points is the linear chirp. Unfortunately, it can undergo shape aberration due to the effects of acoustic diffraction when the aperture exceeds a certain limit. The extent of the aberration is influenced by the shape and size of the transducer, as well as the angle at which the transducer is observed by the receiver. This aberration also affects the shape of the cross-correlation, causing it to lose its easily detectable characteristic of a single global peak, which typically corresponds to the correct lag associated with the signal’s time of arrival. In such instances, cross-correlation techniques yield results with a significantly higher error than anticipated. In fact, the correct lag no longer corresponds to the peak of the cross-correlation. In this study, an alternative technique to global peak detection is proposed, leveraging the inherent symmetry observed in the shape of the aberrated cross-correlation. The numerical simulations, performed using the academic acoustic simulation software Field II, conducted using a typical ultrasonic chirp and ultrasonic emitter, compare the classical and the proposed range techniques in a standard office room. The analysis includes the effects of acoustical reflection in the room and of the acoustic noise at different levels of power. The results demonstrate that the proposed technique enables accurate range estimation even in the presence of severe cross-correlation shape aberrations and for signal-to-noise ratio levels common in office and room environments, even in presence of typical reflections. This allows the use of emitting transducers with a much larger aperture than that allowed by the classical cross-correlation technique. Consequently, it becomes possible to have greater acoustic power available, leading to improved signal-to-noise ratio (SNR). Full article

Hyperspectral image classification remains challenging despite its potential due to the high dimensionality of the data and its limited spatial resolution. To address the limited data samples and less spatial resolution issues, this research paper presents a two-scale module-based CTNet (convolutional transformer network) for the enhancement of spatial and spectral features. In the first module, a virtual RGB image is created from the HSI dataset to improve the spatial features using a pre-trained ResNeXt model trained on natural images, whereas in the second module, PCA (principal component analysis) is applied to reduce the dimensions of the HSI data. After that, spectral features are improved using an EAVT (enhanced attention-based vision transformer). The EAVT contained a multiscale enhanced attention mechanism to capture the long-range correlation of the spectral features. Furthermore, a joint module with the fusion of spatial and spectral features is designed to generate an enhanced feature vector. Through comprehensive experiments, we demonstrate the performance and superiority of the proposed approach over state-of-the-art methods. We obtained AA (average accuracy) values of 97.87%, 97.46%, 98.25%, and 84.46% on the PU, PUC, SV, and Houston13 datasets, respectively. Full article

The increase in life expectancy, and the consequent growth of the elderly population, represents a major challenge to guarantee adequate health and social care. The proposed system aims to provide a tool that automates the evaluation of gait and balance, essential to prevent falls in older people. Through an RGB-D camera, it is possible to capture and digitally represent certain parameters that describe how users carry out certain human motions and poses. Such individual motions and poses are actually related to items included in many well-known gait and balance evaluation tests. According to that information, therapists, who would not need to be present during the execution of the exercises, evaluate the results of such tests and could issue a diagnosis by storing and analyzing the sequences provided by the developed system. The system was validated in a laboratory scenario, and subsequently a trial was carried out in a nursing home with six residents. Results demonstrate the usefulness of the proposed system and the ease of objectively evaluating the main items of clinical tests by using the parameters calculated from information acquired with the RGB-D sensor. In addition, it lays the future foundations for creating a Cloud-based platform for remote fall risk assessment and its integration with a mobile assistant robot, and for designing Artificial Intelligence models that can detect patterns and identify pathologies for enabling therapists to prevent falls in users under risk. Full article

In visual-based localization, prior research falls short in addressing challenges for the Internet of Things with limited computational resources. The dominant state-of-the-art models are based on separate feature extractors and descriptors without consideration of the constraints of small hardware, the issue of inconsistent image scale, or the presence of multi-objects. We introduce “OOPose”, a real-time object-oriented pose estimation framework that leverages dense features from off-the-shelf object detection neural networks. It balances between pixel-matching accuracy and processing speed, enhancing overall performance. When input images share a comparable set of features, their matching accuracy is substantially heightened, while the reduction in image size facilitates faster processing but may compromise accuracy. OOPose resizes both the original library and cropped query object images to a width of 416 pixels. This adjustment results in a 2.4-fold improvement in pose accuracy and an 8.6-fold increase in processing speed. Moreover, OOPose eliminates the need for traditional sparse point extraction and description processes by capitalizing on dense network backbone features and selecting the detected query objects and sources of object library images, ensuring not only 1.3 times more accurate results but also three times greater stability compared to real-time sparse ORB matching algorithms. Beyond enhancements, we demonstrated the feasibility of OOPose in an autonomous mobile robot, enabling self-localization with a single camera at 10 FPS on a single CPU. It proves the cost-effectiveness and real-world applicability of OOPose for small embedded devices, setting the stage for potential markets and providing end-users with distinct advantages. Full article

Molding sand mixtures used in the foundry industry consist of various sands (quartz sands, chromite sands, etc.) and additives such as bentonite. The optimum control of the processes involved in using the mixtures and in their regeneration after the casting requires an efficient in-line monitoring method that is not available today. We are investigating whether such a method can be based on electrical impedance spectroscopy (EIS). To establish a database, we have characterized various sand mixtures by EIS in the frequency range from 0.5 kHz to 1 MHz under laboratory conditions. Attempts at classifying the different molding sand mixtures by support vector machines (SVM) show encouraging results. Already high assignment accuracies (above 90%) could even be improved with suitable feature selection (sequential feature selection). At the same time, the standard uncertainty of the SVM results is low, i.e., data assigned to a class by the presented SVMs have a high probability of being assigned correctly. The application of EIS with subsequent evaluation by machine learning (machine-learning-enhanced EIS, MLEIS) in the field of bulk material monitoring in the foundry industry appears possible. Full article

Dosimetry based on gas detectors operating in the recombination and saturation region provides unique research opportunities but requires high-quality electrometers with a measuring range below 1 pA (10⁻¹² A). The standard approach in electrometry is to strive to increase the accuracy and precision of the measurement, ignoring the importance of its duration. The article presents an algorithm for the measurement of low current values (from 100 fA) that allows both a fast measurement (with a step of 2.3 ms) and high accuracy (measurement error below 0.1%), depending on the measurement conditions and the expected results. A series of tests and validations of the algorithm were carried out in a measurement system with a Keithley 6517B electrometer and a REM-2 recombination chamber under conditions of constant and time-varying radiation fields. The result of the work is a set of parameters that allow for the optimisation of the operation of the algorithm, maximising the quality of the measurements according to needs and the expected results. The algorithm can be used in low current measurement systems, e.g., for dosimetry of mixed radiation fields using recombination methods and chambers. Full article

Designed to meet the demands of AGV global optimal path planning and dynamic obstacle avoidance, this paper proposes a combination of an improved A* algorithm and dynamic window method fusion algorithm. Firstly, the heuristic function is dynamically weighted to reduce the search scope and improve the planning efficiency; secondly, a path-optimization method is introduced to eliminate redundant nodes and redundant turning points in the path; thirdly, combined with the improved A* algorithm and dynamic window method, the local dynamic obstacle avoidance in the global optimal path is realized. Finally, the effectiveness of the proposed method is verified by simulation experiments. According to the results of simulation analysis, the path-planning time of the improved A* algorithm is 26.3% shorter than the traditional A* algorithm, the search scope is 57.9% less, the path length is 7.2% shorter, the number of path nodes is 85.7% less, and the number of turning points is 71.4% less. The fusion algorithm can evade moving obstacles and unknown static obstacles in different map environments in real time along the global optimal path. Full article

The adoption of Industry 4.0 technologies in manufacturing systems has accelerated in recent years, with a shift towards understanding operators’ well-being and resilience within the context of creating a human-centric manufacturing environment. In addition to measuring physical workload, monitoring operators’ cognitive workload is becoming a key element in maintaining a healthy and high-performing working environment in future digitalized manufacturing systems. The current approaches to the measurement of cognitive workload may be inadequate when human operators are faced with a series of new digitalized technologies, where their impact on operators’ mental workload and performance needs to be better understood. Therefore, a new method for measuring and determining the cognitive workload is required. Here, we propose a new method for determining cognitive-workload indices in a human-centric environment. The approach provides a method to define and verify the relationships between the factors of task complexity, cognitive workload, operators’ level of expertise, and indirectly, the operator performance level in a highly digitalized manufacturing environment. Our strategy is tested in a series of experiments where operators perform assembly tasks on a Wankel Engine block. The physiological signals from heart-rate variability and pupillometry bio-markers of 17 operators were captured and analysed using eye-tracking and electrocardiogram sensors. The experimental results demonstrate statistically significant differences in both cardiac and pupillometry-based cognitive load indices across the four task complexity levels (rest, low, medium, and high). Notably, these developed indices also provide better indications of cognitive load responding to changes in complexity compared to other measures. Additionally, while experts appear to exhibit lower cognitive loads across all complexity levels, further analysis is required to confirm statistically significant differences. In conclusion, the results from both measurement sensors are found to be compatible and in support of the proposed new approach. Our strategy should be useful for designing and optimizing workplace environments based on the cognitive load experienced by operators. Full article

Diabetic foot ulcers (DFUs) pose a significant challenge in diabetes care, demanding advanced approaches for effective prevention and management. Smart insoles using sensor technology have emerged as promising tools to address the challenges associated with DFU and neuropathy. By recognizing the pivotal role of smart insoles in successful prevention and healthcare management, this scoping review aims to present a comprehensive overview of the existing evidence regarding DFU studies related to smart insoles, offloading sensors, and actuator technologies. This systematic review identified and critically evaluated 11 key studies exploring both sensor technologies and offloading devices in the context of DFU care through searches in CINAHL, MEDLINE, and ScienceDirect databases. Predominantly, smart insoles, mobile applications, and wearable technologies were frequently utilized for interventions and patient monitoring in diabetic foot care. Patients emphasized the importance of these technologies in facilitating care management. The pivotal role of offloading devices is underscored by the majority of the studies exhibiting increased efficient monitoring, prevention, prognosis, healing rate, and patient adherence. The findings indicate that, overall, smart insoles and digital technologies are perceived as acceptable, feasible, and beneficial in meeting the specific needs of DFU patients. By acknowledging the promising outcomes, the present scoping review suggests smart technologies can potentially redefine DFU management by emphasizing accessibility, efficacy, and patient centricity. Full article

The acoustic tomography (AT) velocity field reconstruction technique has become a research hotspot in recent years due to its noninvasive nature, high accuracy, and real-time measurement advantages. However, most of the existing studies are limited to the reconstruction of the velocity field in a rectangular area, and there are very few studies on a circular area, mainly because the layout of acoustic transducers, selection of acoustic paths, and division of measured regions are more difficult in a circular area than in a rectangular area. Therefore, based on AT and using the reconstruction algorithm of the Markov function and singular value decomposition (MK-SVD), this paper proposes a measured regional division optimization algorithm for velocity field reconstruction in a circular area. First, an acoustic path distribution based on the multipath effect is designed to solve the problem of the limited emission angle of the acoustic transducer. On this basis, this paper proposes an adaptive optimization algorithm for measurement area division based on multiple sub-objectives. The steps are as follows: first, two optimization objectives, the condition number of coefficient matrix and the uniformity of acoustic path distribution, were designed. Then, the weights of each sub-objective are calculated using the coefficient of variation (CV). Finally, the measured regional division is optimized based on particle swarm optimization (PSO). The reconstruction effect of the algorithm and the anti-interference ability are verified through the reconstruction experiments of the model velocity field and the simulated velocity field. Full article

In this study, flexible nanocomposites made from PVDF-HFP reinforced with carbon nanotubes (CNTs) and graphene nanoplatelets (GNPs) are manufactured using a sonication and solvent casting method for monitoring purposes. More specifically, the effect of the volume batch under the sonication process is explored. For CNT-based composites, the electrical conductivity decreases as the batch volume increases due to less effective dispersion of the CNTs during the 30-min sonication. The maximum electrical conductivity achieved in this type of sensor is 1.44 ± 0.17 S/m. For the GNP-based nanocomposites, the lower the batch volume is, the more breakage of nanoplatelets is induced by sonication, and the electrical response decreases. This is also validated by AC analysis, where the characteristic frequencies are extracted. Here, the maximum electrical conductivity measured is 8.66 ± 1.76 S/m. The electromechanical results also show dependency on the batch volume. In the CNT-based nanocomposites, the higher gauge factor achieved corresponds to the batch size, where the sonication may be more effective because it leads to a dispersed pathway formed by aggregates connected by tunneling mechanisms. In contrast, in the CNT-based nanocomposites, the GF depends on the lateral size of the GNPs. The biggest GF of all sensors is achieved with the PVDF-HFP/GNP sensors, having a value of 69.36 × 10⁴ at 35% of strain, while the highest GF achieved with a PVDF-HFP/CNT sensor is 79.70 × 10³ at 70%. In addition, cycling tests show robust electromechanical response with cycling for two different strain percentages for each type of nanocomposite. The sensor with the highest sensitivity is selected for monitoring two joint movements as proof of the applicability of the sensors manufactured. Full article

In the era of Industry 5.0, advanced technologies like artificial intelligence (AI), robotics, big data, and the Internet of Things (IoT) offer promising avenues for economic growth and solutions to societal challenges. Digital twin technology is important for real-time three-dimensional space reproduction in this transition, and unmanned aerial vehicles (UAVs) can support it. While recent studies have explored the potential applications of UAVs in nonterrestrial networks (NTNs), bandwidth limitations have restricted their utility. This paper addresses these constraints by integrating millimeter wave (mmWave) technology into UAV networks for high-definition video transmission. Specifically, we focus on coordinating intelligent reflective surfaces (IRSs) and UAV networks to extend coverage while maintaining virtual line-of-sight (LoS) conditions essential for mmWave communication. We present a novel approach for integrating IRS into Beyond 5G/6G networks to enhance high-speed communication coverage. Our proposed IRS selection method ensures optimal communication paths between UAVs and user equipment (UE). We perform numerical analysis in a realistically modeled 3D urban environment to validate our approach. Our results demonstrate significant improvements in the received SNR for multiple UEs upon the introduction of IRSs, and they confirm the feasibility of coverage extension in mmWave UAV networks. Full article

Autonomous driving recognition technology that can quickly and accurately recognize even small objects must be developed in high-speed situations. This study proposes an object point extraction method using rule-based LiDAR ring data and edge triggers to increase both speed and performance. The LiDAR’s ring information is interpreted as a digital pulse to remove the ground, and object points are extracted by detecting discontinuous edges of the z value aligned with the ring ID and azimuth. A bounding box was simply created using DBSCAN and PCA to check recognition performance from the extracted object points. Verification of the results of removing the ground and extracting points through Ring Edge was conducted using SemanticKITTI and Waymo Open Dataset, and it was confirmed that both F1 scores were superior to RANSAC. In addition, extracting bounding boxes of objects also showed higher PDR index performance when verified in open datasets, virtual driving environments, and actual driving environments. Full article

With the development of IoT technology and 5G massive machine-type communication, the 3GPP standardization body considered as viable the integration of Narrowband Internet of Things (NB-IoT) in low Earth orbit (LEO) satellite-based architectures. However, the presence of the LEO satellite channel comes up with new challenges for the NB-IoT random access procedures and coverage enhancement mechanism. In this paper, an Adaptive Coverage Enhancement (ACE) method is proposed to meet the requirement of random access parameter configurations for diverse applications. Based on stochastic geometry theory, an expression of random access channel (RACH) success probability is derived for LEO satellite-based NB-IoT networks. On the basis of a power consumption model of the NB-IoT terminal, a multi-objective optimization problem is formulated to trade-off RACH success probability and power consumption. To solve this multi-objective optimization problem, we employ the Non-dominated Sorting Genetic Algorithms-II (NSGA-II) method to obtain the Pareto-front solution set. According to different application requirements, we also design a random access parameter configuration method to minimize the power consumption under the constraints of RACH success probability requirements. Simulation results show that the maximum number of repetitions and back-off window size have a great influence on the system performance and their value ranges should be set within [4, 18] and [0, 2048]. The power consumption of coverage enhancement with ACE is about 58% lower than that of the 3GPP proposed model. All this research together provides good reference for the scale deployment of NB-IoT in LEO satellite networks. Full article

Imaging using scattering media is a very important yet challenging technology. As one of the most widely used scattering imaging methods, speckle autocorrelation technology has important applications in several fields. However, traditional speckle autocorrelation imaging methods usually use iterative phase recovery algorithms to obtain the Fourier phase of hidden objects, posing issues such as large data calculation volumes and uncertain reconstruction results. Here, we propose a single-shot scattering imaging method based on the bispectrum truncation method. The bispectrum analysis is utilized for hidden object phase recovery, the truncation method is used to avoid the computation of redundant data when calculating the bispectrum data, and the method is experimentally verified. The experimental results show that our method does not require uncertain iterative calculations and can reduce the bispectrum data computation by more than 80% by adjusting the truncation factor without damaging the imaging quality, which greatly improves imaging efficiency. This method paves the way for rapid imaging through scattering media and brings benefits for imaging in dynamic situations. Full article

Aero engines are the key power source for aerospace vehicles. Cermet turbine blades are the guarantee for the new-generation fighters to improve aero-engine overall performance. X-ray non-destructive reconstruction can obtain the internal structure and morphology of cermet turbine blades. However, the beam hardening effect causes artifacts in objects and affects the reconstruction quality, which is an issue that needs to be solved urgently. This study proposes a hardening-correction framework for industrial computed tomography (ICT) images based on iterative linear fitting. First, an iterative binarization was performed to improve the penetration length accuracy of the forward projection. Then, the proposed linear fitting technology combined with the Hermite function model is derived and analyzed to obtain suitable parameters of blade data. Finally, the fitting curves of the blade data, using the proposed method and the traditional polynomial fitting method, were analyzed and compared and were used to correct the engine turbine blade projection data to reconstruct different groups of tomographic images. Different groups of tomographic images were analyzed using three quantitative image quality evaluation indicators. The results show that the root-mean-square error (RMSE) of the tomographic image obtained by the proposed framework is 0.0133, which is lower than that of the compared method. The peak signal-to-noise ratio (PSNR) is 37.7050 dB and the feature structural similarity (FSIM) is 0.9881, which are both higher than that of the compared method. The proposed method improves the hardening-artifact-correction capability and can obtain higher-quality images, which provides new ideas for the development of imaging and detection of new-generation aero-engine turbine blades. Full article

Optical biosensors have a significant impact on various aspects of our lives. In many applications of optical biosensors, fluidic chambers play a crucial role in facilitating controlled fluid delivery. It is essential to achieve complete liquid replacement in order to obtain accurate and reliable results. However, the configurations of fluidic chambers vary across different optical biosensors, resulting in diverse fluidic volumes and flow rates, and there are no standardized guidelines for liquid replacement. In this paper, we utilize COMSOL Multiphysics, a finite element analysis software, to investigate the optimal fluid volume required for two types of fluidic chambers in the context of the oblique–incidence reflectivity difference (OI-RD) biosensor. We found that the depth of the fluidic chamber is the most crucial factor influencing the required liquid volume, with the volume being a quadratic function of the depth. Additionally, the required fluid volume is also influenced by the positions on the substrate surface bearing samples, while the flow rate has no impact on the fluid volume. Full article

The Kenyan part of the East African Rift System hosts several geothermal fields for energy production. Changes in the extraction rate of geothermal fluids and the amount of water re-injected into the system affect reservoir pressure and production capacity over time. Understanding the balance of production, natural processes and the response of the geothermal system requires long-term monitoring. The presence of a geothermal system at depth is often accompanied by surface manifestations, such as hot water springs and fumaroles, which have the potential for monitoring subsurface activity. Two thermal camera timelapse systems were developed and installed as part of a multi-sensor observatory in Kenya to capture fumarole activity over time. These cameras are an aggregation of a camera unit, a control unit, and a battery charged by a solar panel, and they monitor fumarole activity on an hourly basis, with a deep sleep of the system in between recordings. The article describes the choice of hardware and software, presents the data that the cameras acquire, and discusses the system’s performance and possible improvement points. Full article

Crop leaf length, perimeter, and area serve as vital phenotypic indicators of crop growth status, the measurement of which is important for crop monitoring and yield estimation. However, processing a leaf point cloud is often challenging due to cluttered, fluctuating, and uncertain points, which culminate in inaccurate measurements of leaf phenotypic parameters. To tackle this issue, the RKM-D point cloud method for measuring leaf phenotypic parameters is proposed, which is based on the fusion of improved Random Sample Consensus with a ground point removal (R) algorithm, the K-means clustering (K) algorithm, the Moving Least Squares (M) method, and the Euclidean distance (D) algorithm. Pepper leaves were obtained from three growth periods on the 14th, 28th, and 42nd days as experimental subjects, and a stereo camera was employed to capture point clouds. The experimental results reveal that the RKM-D point cloud method delivers high precision in measuring leaf phenotypic parameters. (i) For leaf length, the coefficient of determination (R²) surpasses 0.81, the mean absolute error (MAE) is less than 3.50 mm, the mean relative error (MRE) is less than 5.93%, and the root mean square error (RMSE) is less than 3.73 mm. (ii) For leaf perimeter, the R² surpasses 0.82, the MAE is less than 7.30 mm, the MRE is less than 4.50%, and the RMSE is less than 8.37 mm. (iii) For leaf area, the R² surpasses 0.97, the MAE is less than 64.66 mm², the MRE is less than 4.96%, and the RMSE is less than 73.06 mm². The results show that the proposed RKM-D point cloud method offers a robust solution for the precise measurement of crop leaf phenotypic parameters. Full article

We aimed to determine the feasibility, test–retest reliability and long-term stability of a novel method for assessing the force (torque)-velocity (cadence) profile and maximal dynamic force (MDF) during leg-pedaling using a friction-loaded isoinertial cycle ergometer and a high-precision power-meter device. Fifty-two trained male cyclists completed a progressive loading test up to the one-repetition maximum (1RM) on a cycle ergometer. The MDF was defined as the force attained at the cycle performed with the 1RM-load. To examine the test–retest reliability and long-term stability of torque–cadence values, the progressive test was repeated after 72 h and also after 10 weeks of aerobic and strength training. The participants’ MDF averaged 13.4 ± 1.3 N·kg⁻¹, which was attained with an average pedal cadence of 21 ± 3 rpm. Participants’ highest power output value was attained with a cadence of 110 ± 16 rpm (52 ± 5% MDF). The relationship between the MDF and cadence proved to be very strong (R² = 0.978) and independent of the cyclists’ MDF (p = 0.66). Cadence values derived from this relationship revealed a very high test–retest repeatability (mean SEM = 4 rpm, 3.3%) and long-term stability (SEM = 3 rpm, 2.3%); despite increases in the MDF following the 10-week period. Our findings support the validity, reliability and long-term stability of this method for the assessment of the torque–cadence profile and MDF in cyclists. Full article

Speaker recognition is a challenging problem in behavioral biometrics that has been rigorously investigated over the last decade. Although numerous supervised closed-set systems inherit the power of deep neural networks, limited studies have been made on open-set speaker recognition. This paper proposes a self-supervised open-set speaker recognition that leverages the geometric properties of speaker distribution for accurate and robust speaker verification. The proposed framework consists of a deep neural network incorporating a wider viewpoint of temporal speech features and Laguerre–Voronoi diagram-based speech feature extraction. The deep neural network is trained with a specialized clustering criterion that only requires positive pairs during training. The experiments validated that the proposed system outperformed current state-of-the-art methods in open-set speaker recognition and cluster representation. Full article

Dental 3D modeling plays a pivotal role in digital dentistry, offering precise tools for treatment planning, implant placement, and prosthesis customization. Traditional methods rely on physical plaster casts, which pose challenges in storage, accessibility, and accuracy, fueling interest in digitization using 3D computed tomography (CT) imaging. We introduce a method that can reduce both artifacts simultaneously. To validate the proposed method, we carried out CT scan experiments using plaster dental casts created from dental impressions. After the artifact correction, the CT image quality was greatly improved in terms of image uniformity, contrast-to-noise ratio (CNR), and edge sharpness. We examined the correction effects on the accuracy of the 3D models generated from the CT images. As referenced to the 3D models derived from the optical scan data, the root mean square (RMS) errors were reduced by 8.8~71.7% for three dental casts of different sizes and shapes. Our method offers a solution to challenges posed by artifacts in CT scanning of plaster dental casts, leading to enhanced 3D model accuracy. This advancement holds promise for dental professionals seeking precise digital modeling for diverse applications in dentistry. Full article

Clinical screening tests for balance and mobility often fall short of predicting fall risk. Cognitive distractors and unpredictable external stimuli, common in busy natural environments, contribute to this risk, especially in older adults. Less is known about the effects of upper sensory–motor coordination, such as coordinating one’s hand with an external stimulus. We combined movement sonification and affordable inertial motion sensors to develop a task for the precise measurement and manipulation of full-body interaction with stimuli in the environment. In a double-task design, we studied how a supra-postural activity affected quiet stance. The supra-postural task consisted of rhythmic synchronization with a repetitive auditory stimulus. The stimulus was attentionally demanding because it was being modulated continuously. The participant’s hand movement was sonified in real time, and their goal was to synchronize their hand movement with the stimulus. In the unpredictable condition, the tempo changed at random points in the trial. A separate sensor recorded postural fluctuations. Young healthy adults were compared to older adult (OA) participants without known risk of falling. The results supported the hypothesis that supra-postural coordination would entrain postural control. The effect was stronger in OAs, supporting the idea that diminished reserve capacities reduce the ability to isolate postural control from sensory–motor and cognitive activity. Full article

Energy efficiency and security issues are the main concerns in wireless sensor networks (WSNs) because of limited energy resources and the broadcast nature of wireless communication. Therefore, how to improve the energy efficiency of WSNs while enhancing security performance has attracted widespread attention. In order to solve this problem, this paper proposes a new deep reinforcement learning (DRL)-based strategy, i.e., DeepNR strategy, to enhance the energy efficiency and security performance of WSN. Specifically, the proposed DeepNR strategy approximates the Q-value by designing a deep neural network (DNN) to adaptively learn the state information. It also designs DRL-based multi-level decision-making to learn and optimize the data transmission paths in real time, which eventually achieves accurate prediction and decision-making of the network. To further enhance security performance, the DeepNR strategy includes a defense mechanism that responds to detected attacks in real time to ensure the normal operation of the network. In addition, DeepNR adaptively adjusts its strategy to cope with changing network environments and attack patterns through deep learning models. Experimental results show that the proposed DeepNR outperforms the conventional methods, demonstrating a remarkable 30% improvement in network lifespan, a 25% increase in network data throughput, and a 20% enhancement in security measures. Full article

Near-infrared (NIR) spectroscopy is widely used as a nondestructive evaluation (NDE) tool for predicting wood properties. When deploying NIR models, one faces challenges in ensuring representative training data, which large datasets can mitigate but often at a significant cost. Machine learning and deep learning NIR models are at an even greater disadvantage because they typically require higher sample sizes for training. In this study, NIR spectra were collected to predict the modulus of elasticity (MOE) of southern pine lumber (training set = 573 samples, testing set = 145 samples). To account for the limited size of the training data, this study employed a generative adversarial network (GAN) to generate synthetic NIR spectra. The training dataset was fed into a GAN to generate 313, 573, and 1000 synthetic spectra. The original and enhanced datasets were used to train artificial neural networks (ANNs), convolutional neural networks (CNNs), and light gradient boosting machines (LGBMs) for MOE prediction. Overall, results showed that data augmentation using GAN improved the coefficient of determination (R²) by up to 7.02% and reduced the error of predictions by up to 4.29%. ANNs and CNNs benefited more from synthetic spectra than LGBMs, which only yielded slight improvement. All models showed optimal performance when 313 synthetic spectra were added to the original training data; further additions did not improve model performance because the quality of the datapoints generated by GAN beyond a certain threshold is poor, and one of the main reasons for this can be the size of the initial training data fed into the GAN. LGBMs showed superior performances than ANNs and CNNs on both the original and enhanced training datasets, which highlights the significance of selecting an appropriate machine learning or deep learning model for NIR spectral-data analysis. The results highlighted the positive impact of GAN on the predictive performance of models utilizing NIR spectroscopy as an NDE technique and monitoring tool for wood mechanical-property evaluation. Further studies should investigate the impact of the initial size of training data, the optimal number of generated synthetic spectra, and machine learning or deep learning models that could benefit more from data augmentation using GANs. Full article

The accuracy and efficacy of medical treatment would be greatly improved by the continuous and real-time monitoring of protein biomarkers. Identification of cancer biomarkers in patients with solid malignant tumors is receiving increasing attention. Existing techniques for detecting cancer proteins, such as the enzyme-linked immunosorbent assay, require a lot of work, are not multiplexed, and only allow for single-time point observations. In order to get one step closer to clinical usage, a dynamic platform for biosensing the cancer biomarker CD44 using a single-mode optical fiber-based ball resonator biosensor was designed, constructed and evaluated in this work. The main novelty of the work is an in-depth study of the capability of an in-house fabricated optical fiber biosensor for in situ detection of a cancer biomarker (CD44 protein) by conducting several types of experiments. The main results of the work are as follows: (1) Calibration of the fabricated fiber-optic ball resonator sensors in both static and dynamic conditions showed similar sensitivity to the refractive index change demonstrating its usefulness as a biosensing platform for dynamic measurements; (2) The fabricated sensors were shown to be insensitive to pressure changes further confirming their utility as an in situ sensor; (3) The sensor’s packaging and placement were optimized to create a better environment for the fabricated ball resonator’s performance in blood-mimicking environment; (4) Incubating increasing protein concentrations with antibody-functionalized sensor resulted in nearly instantaneous signal change indicating a femtomolar detection limit in a dynamic range from 7.1 aM to 16.7 nM; (5) The consistency of the obtained signal change was confirmed by repeatability studies; (6) Specificity experiments conducted under dynamic conditions demonstrated that the biosensors are highly selective to the targeted protein; (7) Surface morphology studies by AFM measurements further confirm the biosensor’s exceptional sensitivity by revealing a considerable shift in height but no change in surface roughness after detection. The biosensor’s ability to analyze clinically relevant proteins in real time with high sensitivity offers an advancement in the detection and monitoring of malignant tumors, hence improving patient diagnosis and health status surveillance. Full article

We have investigated a polarization property of the (specularly) reflected light from an aluminum grating, coated with a palladium (Pd) thin-film on its surface. The polarization property, which is associated with surface plasmon resonance (SPR), and occurs in the Pd thin-film on the aluminum grating in a conical mounting, is observed as a rapid change in the normalized Stokes parameter $s_3,$ around the resonance angle, ${\theta }_{\mathrm{s}\mathrm{p}}$, at which point, SPR occurs. The sensing technique used the rapid change in $s_3$ to allow us to successfully detect a small change in the complex refractive index of the Pd thin-film layer upon exposure to hydrogen gas, with a concentration near the lower explosion level. Experimental results showed that the sensing technique provided a sensitive and stable response when the Pd thin-film layer was exposed to gas mixtures containing hydrogen at concentrations of 1 to 4% (by volume) in nitrogen. Full article

Lateral flow immunoassay (LFIA) is extensively utilized for point-of-care testing due to its ease of operation, cost-effectiveness, and swift results. This study investigates the flow dynamics and reaction mechanisms in LFIA by developing a three-dimensional model using the Richards equation and porous media transport, and employing numerical simulations through the finite element method. The study delves into the transport and diffusion behaviors of each reaction component in both sandwich LFIA and competitive LFIA under non-uniform flow conditions. Additionally, the impact of various parameters (such as reporter particle concentration, initial capture probe concentrations for the T-line and C-line, and reaction rate constants) on LFIA performance is analyzed. The findings reveal that, in sandwich LFIA, optimizing parameters like increasing reporter particle concentration and initial capture probe concentration for the T-line, as well as adjusting reaction rate constants, can effectively enhance detection sensitivity and broaden the working range. Conversely, in competitive LFIA, the effects are inverse. This model offers valuable insights for the design and enhancement of LFIA assays. Full article