Sensors

Issues of fairness and consistency in Taekwondo poomsae evaluation have often occurred due to the lack of an objective evaluation method. This study proposes a three-dimensional (3D) convolutional neural network–based action recognition model for an objective evaluation of Taekwondo poomsae. The model exhibits robust recognition performance regardless of variations in the viewpoints by reducing the discrepancy between the training and test images. It uses 3D skeletons of poomsae unit actions collected using a full-body motion-capture suit to generate synthesized two-dimensional (2D) skeletons from desired viewpoints. The 2D skeletons obtained from diverse viewpoints form the training dataset, on which the model is trained to ensure consistent recognition performance regardless of the viewpoint. The performance of the model was evaluated against various test datasets, including projected 2D skeletons and RGB images captured from diverse viewpoints. Comparison of the performance of the proposed model with those of previously reported action recognition models demonstrated the superiority of the proposed model, underscoring its effectiveness in recognizing and classifying Taekwondo poomsae actions. Full article

This paper investigates the direction of arrival (DOA) estimation of coherent signals with a moving coprime array (MCA). Spatial smoothing techniques are often used to deal with the covariance matrix of coherent signals, but they cannot be used in sparse arrays. Therefore, super-resolution algorithms such as multiple signal classification (MUSIC) cannot be applied in the DOA estimation of coherent signals in sparse arrays. In this study, we propose an enhanced spatial smoothing method specifically designed for MCA. Firstly, we combine the signals received by the MCA at different times, which can be regarded as a sparse array with a larger number of array sensors. Secondly, we describe how to compute the covariance matrix, derive the signal subspace by eigenvalue decomposition, and prove that the signal subspace is also equivalent to a received signal. Thirdly, we apply enhanced spatial smoothing to the signal subspace and construct a rank recovered covariance matrix. Finally, the DOA of coherent signals are well estimated by the MUSIC algorithm. The simulation results validate the improved performance of the proposed algorithm compared with traditional methods, particularly in scenarios with low signal-to-noise ratios. Full article

The behavior of multicamera interference in 3D images (e.g., depth maps), which is based on infrared (IR) light, is not well understood. In 3D images, when multicamera interference is present, there is an increase in the amount of zero-value pixels, resulting in a loss of depth information. In this work, we demonstrate a framework for synthetically generating direct and indirect multicamera interference using a combination of a probabilistic model and ray tracing. Our mathematical model predicts the locations and probabilities of zero-value pixels in depth maps that contain multicamera interference. Our model accurately predicts where depth information may be lost in a depth map when multicamera interference is present. We compare the proposed synthetic 3D interference images with controlled 3D interference images captured in our laboratory. The proposed framework achieves an average root mean square error (RMSE) of 0.0625, an average peak signal-to-noise ratio (PSNR) of 24.1277 dB, and an average structural similarity index measure (SSIM) of 0.9007 for predicting direct multicamera interference, and an average RMSE of 0.0312, an average PSNR of 26.2280 dB, and an average SSIM of 0.9064 for predicting indirect multicamera interference. The proposed framework can be used to develop and test interference mitigation techniques that will be crucial for the successful proliferation of these devices. Full article

Two-needle 3D stochastic microsensors based on boron- and nitrogen-decorated gra-phenes, modified with N-(2-mercapto-1H-benzo[d]imidazole-5-yl), were designed and used for the molecular recognition and quantification of CA 72-4, CA 19-9, CEA and CA 125 biomarkers in biological samples such as whole blood, urine, saliva and tumoral tissue. The NBGr-2 sensor yielded lower limits of determination. For CEA, the LOD was 4.10 × 10⁻¹⁵ s⁻¹ g⁻¹ mL, while for CA72-4, the LOD was 4.00 × 10⁻¹¹ s⁻¹ U⁻¹ mL. When the NBGr-1 sensor was employed, the best results were obtained for CA12-5 and CA19-9, with values of LODs of 8.37 × 10⁻¹⁴ s⁻¹ U⁻¹ mL and 2.09 × 10⁻¹³ s⁻¹ U⁻¹ mL, respectively. High sensitivities were obtained when both sensors were employed. Broad linear concentration ranges favored their determination from very low to higher concentrations in biological samples, ranging from 8.37 × 10⁻¹⁴ to 8.37 × 10³ s⁻¹ U⁻¹ mL for CA12-5 when using the NBGr-1 sensor, and from 4.10 × 10⁻¹⁵ to 2.00 × 10⁻⁷ s⁻¹ g⁻¹ mL for CEA when using the NBGr-2 sensor. Student’s t-test showed that there was no significant difference between the results obtained utilizing the two microsensors for the screening tests, at a 99% confidence level, with the results obtained being lower than the tabulated values. Full article

Activity monitoring of living creatures based on the structural vibration of ambient objects is a promising method. For vibration measurement, multi-axial inertial measurement units (IMUs) offer a high sampling rate and a small size compared to geophones, but have higher intrinsic noise. This work proposes a sensing device that combines a single six-axis IMU with a beam structure to enable measurement of small vibrations. The beam structure is integrated into the PCB of the sensing device and connects the IMU to the ambient object. The beam is designed with finite element method (FEM) and optimized to maximize the vibration amplitude. Furthermore, the beam oscillation creates simultaneous translation and rotation of the IMU, which is measured with its accelerometers and gyroscopes. On this basis, a novel sensor fusion algorithm is presented that adaptively combines IMU data in the wavelet domain to reduce intrinsic sensor noise. In experimental evaluation, the proposed sensing device using a beam structure achieves a 6.2-times-higher vibration amplitude and an increase in signal energy of 480% when compared to a directly mounted IMU without a beam. The sensor fusion algorithm provides a noise reduction of 5.6% by fusing accelerometer and gyroscope data at 103 Hz. Full article

The Internet of Things (IoT) has significantly benefited several businesses, but because of the volume and complexity of IoT systems, there are also new security issues. Intrusion detection systems (IDSs) guarantee both the security posture and defense against intrusions of IoT devices. IoT systems have recently utilized machine learning (ML) techniques widely for IDSs. The primary deficiencies in existing IoT security frameworks are their inadequate intrusion detection capabilities, significant latency, and prolonged processing time, leading to undesirable delays. To address these issues, this work proposes a novel range-optimized attention convolutional scattered technique (ROAST-IoT) to protect IoT networks from modern threats and intrusions. This system uses the scattered range feature selection (SRFS) model to choose the most crucial and trustworthy properties from the supplied intrusion data. After that, the attention-based convolutional feed-forward network (ACFN) technique is used to recognize the intrusion class. In addition, the loss function is estimated using the modified dingo optimization (MDO) algorithm to ensure the maximum accuracy of classifier. To evaluate and compare the performance of the proposed ROAST-IoT system, we have utilized popular intrusion datasets such as ToN-IoT, IoT-23, UNSW-NB 15, and Edge-IIoT. The analysis of the results shows that the proposed ROAST technique did better than all existing cutting-edge intrusion detection systems, with an accuracy of 99.15% on the IoT-23 dataset, 99.78% on the ToN-IoT dataset, 99.88% on the UNSW-NB 15 dataset, and 99.45% on the Edge-IIoT dataset. On average, the ROAST-IoT system achieved a high AUC-ROC of 0.998, demonstrating its capacity to distinguish between legitimate data and attack traffic. These results indicate that the ROAST-IoT algorithm effectively and reliably detects intrusion attacks mechanism against cyberattacks on IoT systems. Full article

Patients affected with type 1 diabetes and a non-negligible number of patients with type 2 diabetes are insulin dependent. Both the injection technique and the choice of the most suitable needle are fundamental for allowing them to have a good injection experience. The needles may differ in several parameters, from the length and diameter, up to the forces required to perform the injection and to some geometrical parameters of the needle tip (e.g., number of facets or bevels). The aim of the research is to investigate whether an increased number of bevels could decrease forces and energy involved in the insertion–extraction cycle, thus potentially allowing patients to experience lower pain. Two needle variants, namely, 31 G × 5 mm and 32 G × 4 mm, are considered, and experimental tests are carried out to compare 3-bevels with 5-bevels needles for both the variants. The analysis of the forces and energy for both variants show that the needles with 5 bevels require a statistically significant lower drag or sliding force (p-value = 0.040 for the 31 G × 5 mm needle and p-value < 0.001 for 32 G × 4 mm), extraction force (p-value < 0.001 for both variants), and energy (p-value < 0.001 for both variants) during the insertion–extraction cycle. As a result, 3-bevels needles do not have the same functionality of 5-bevels needles, show lower capacity of drag and extraction, and can potentially be related to more painful injection experience for patients. Full article

The digestion of protein into peptide fragments reduces the size and complexity of protein molecules. Peptide fragments can be analyzed with higher sensitivity (often > 10² fold) and resolution using MALDI-TOF mass spectrometers, leading to improved pattern recognition by common machine learning algorithms. In turn, enhanced sensitivity and specificity for bacterial sorting and/or disease diagnosis may be obtained. To test this hypothesis, four exemplar case studies have been pursued in which samples are sorted into dichotomous groups by machine learning (ML) software based on MALDI-TOF spectra. Samples were analyzed in ‘intact’ mode in which the proteins present in the sample were not digested with protease prior to MALDI-TOF analysis and separately after the standard overnight tryptic digestion of the same samples. For each case, sensitivity (sens), specificity (spc), and the Youdin index (J) were used to assess the ML model performance. The proteolytic digestion of samples prior to MALDI-TOF analysis substantially enhanced the sensitivity and specificity of dichotomous sorting. Two exceptions were when substantial differences in chemical composition between the samples were present and, in such cases, both ‘intact’ and ‘digested’ protocols performed similarly. The results suggest proteolytic digestion prior to analysis can improve sorting in MALDI/ML-based workflows and may enable improved biomarker discovery. However, when samples are easily distinguishable protein digestion is not necessary to obtain useful diagnostic results. Full article

In recent years, unmanned aerial vehicles (UAVs) have become a valuable platform for many applications, including communication networks. UAV-enabled wireless communication faces challenges in complex urban and dynamic environments. UAVs can suffer from power limitations and path losses caused by non-line-of-sight connections, which may hamper communication performance. To address these issues, reconfigurable intelligent surfaces (RIS) have been proposed as helpful technologies to enhance UAV communication networks. However, due to the high mobility of UAVs, complex channel environments, and dynamic RIS configurations, it is challenging to estimate the link quality of ground users. In this paper, we propose a link quality estimation model using a gated recurrent unit (GRU) to assess the link quality of ground users for a multi-user RIS-assisted UAV-enabled wireless communication system. Our proposed framework uses a time series of user channel data and RIS phase shift information to estimate the quality of the link for each ground user. The simulation results showed that the proposed GRU model can effectively and accurately estimate the link quality of ground users in the RIS-assisted UAV-enabled wireless communication network. Full article

Spectral efficiency is a crucial metric in wireless communication systems, as it defines how much information can be transmitted over a given amount of spectrum resources. Non-orthogonal multiple access (NOMA) is a promising technology that has captured the interest of the wireless research community because of its capacity to enhance spectral efficiency. NOMA allows multiple users to share the same frequency band and time slot by assigning different power levels and modulation schemes to different users. Furthermore, channel assignment is a critical challenge in OFDMA-NOMA systems that must be addressed to achieve optimal performance. In this context, we propose a solution for both channel and power assignment based on channel condition by splitting the problem into two parts: first, we introduce a novel algorithm to solve the channel user allocation problem, which we refer to as Channel User Sorting and Filling (CUSF). Then, we solve the power allocation problem in two steps: we apply the water filling algorithm at the power assignment and then we implement the Fractional Transmit Power Control (FTPC) algorithm in the NOMA power assignment. Full article

In recent years, the use of electroencephalography (EEG) has grown as a tool for diagnostic and brain function monitoring, being a simple and non-invasive method compared with other procedures like histological sampling. Typically, in order to extract functional brain responses from EEG signals, prolonged and repeated stimuli are needed because of the artifacts generated in recordings which adversely impact the stimulus-response analysis. To mitigate the artifact effect, correlation analysis (CA) methods are applied in the literature, where the predominant approaches focus on enhancing stimulus-response correlations through the use of linear analysis methods like canonical correlation analysis (CCA). This paper introduces a novel CA framework based on a neural network with a loss function specifically designed to maximize correlation between EEG and speech stimuli. Compared with other deep learning CA approaches (DCCAs) in the literature, this framework introduces a single multilayer perceptron (MLP) network instead of two networks for each stimulus. To validate the proposed approach, a comparison with linear CCA (LCCA) and DCCA was performed, using a dataset containing the EEG traces of subjects listening to speech stimuli. The experimental results show that the proposed method improves the overall Pearson correlation by 10.56% compared with the state-of-the-art DCCA method. Full article

Ciphertext policy–attribute-based encryption (CP-ABE), which provides fine-grained access control and ensures data confidentiality, is widely used in data sharing. However, traditional CP-ABE schemes often choose to outsource data to untrusted third-party cloud service providers for storage or to verify users’ access rights through third parties, which increases the risk of privacy leakage and also suffers from the problem of opaque permission verification. This paper proposes an access control scheme based on blockchain and CP-ABE, which is based on multiple authorization centers and supports policy updating. In addition, blockchain technology’s distributed, decentralized, and tamper-proof features are utilized to solve the trust crisis problem in the data-sharing process. Security analysis and performance evaluation show that the proposed scheme improves the computational efficiency by 18%, 26%, and 68% compared to previous references. The proposed scheme also satisfies the indistinguishability under chosen-plaintext attack (IND-CPA). Full article

This work proposes an innovative method, based on the use of low-cost infrared thermography (IRT) instrumentation, to assess in real time the effectiveness of scoliosis braces. Establishing the effectiveness of scoliosis braces means deciding whether the pressure exerted by the brace on the patient’s back is adequate for the intended therapeutic purpose. Traditionally, the evaluation of brace effectiveness relies on empirical, qualitative assessments carried out by orthopedists during routine follow-up examinations. Hence, it heavily depends on the expertise of the orthopedists involved. In the state of the art, the only objective methods used to confirm orthopedists’ opinions are based on the evaluation of how scoliosis progresses over time, often exposing people to ionizing radiation. To address these limitations, the method proposed in this work aims to provide a real-time, objective assessment of the effectiveness of scoliosis braces in a non-harmful way. This is achieved by exploiting the thermoelastic effect and correlating temperature changes on the patient’s back with the mechanical pressure exerted by the braces. A system based on this method is implemented and then validated through an experimental study on 21 patients conducted at an accredited orthopedic center. The experimental results demonstrate a classification accuracy slightly below 70% in discriminating between adequate and inadequate pressure, which is an encouraging result for further advancement in view of the clinical use of such systems in orthopedic centers. Full article

Traditional ultrasonic testing uses a single probe or phased array probe to investigate and visualize defects by adapting certain imaging algorithms. The time-domain synthetic aperture focusing technique (T-SAFT) is an imaging algorithm that employs a single probe to scan along the test specimen in various positions, to generate inspection images with better resolution. Both the T-SAFT and phased array probes are contact methods with limited bandwidth. This work aims to combine the advantages of the T-SAFT and phased array in a noncontact way with the aid of laser ultrasonics. Here, a pulsed laser beam is employed to generate ultrasonic waves in both thermoelastic and ablation regimes, whereas the laser Doppler vibrometer is used to acquire the generated signals. These two lasers are focused on the test specimen and, to avoid the plasma and crater influence in the ablation regime, the transmission beam and reception beam are separated by 5 mm. By moving the test specimen with a step size of 0.5 mm, a 1D linear phased array (41 and 43 elements) with a pitch of 0.5 mm was synthesized, and three side-drilled holes (Ø 8 mm—thermoelastic regime, Ø 10 mm and Ø 2 mm—ablation regime) were introduced for inspection. The A-scan data obtained from these elements were processed via the T-SAFT algorithm to generate the inspection images in various grid sizes. The results showed that the defect reflections obtained in the ablation regime have better visibility than those from the thermoelastic regime. This is due to the high-amplitude signals obtained in the ablation regime, which pave the way for enhancing the pixel intensity of each grid. Moreover, the separation distance (5 mm) does not have any significant effect on the defect location during the reconstruction process. Full article

Due to its advantages of low latency, low power consumption, and high flexibility, FPGA-based acceleration technology has been more and more widely studied and applied in the field of computer vision in recent years. An FPGA-based feature extraction and tracking accelerator for real-time visual odometry (VO) and visual simultaneous localization and mapping (V-SLAM) is proposed, which can realize the complete acceleration processing capability of the image front-end. For the first time, we implement a hardware solution that combines features from accelerated segment test (FAST) feature points with Gunnar Farneback (GF) dense optical flow to achieve better feature tracking performance and provide more flexible technical route selection. In order to solve the scale invariance and rotation invariance lacking problems of FAST features, an efficient pyramid module with a five-layer thumbnail structure was designed and implemented. The accelerator was implemented on a modern Xilinx Zynq FPGA. The evaluation results showed that the accelerator could achieve stable tracking of features of violently shaking images and were consistent with the results from MATLAB code running on PCs. Compared to PC CPUs, which require seconds of processing time, the processing latency was greatly reduced to the order of milliseconds, making GF dense optical flow an efficient and practical technical solution on the edge side. Full article

With advances in the development of autonomous vehicles (AVs), more attention has been paid to the effects caused by adverse weather conditions. It is well known that the performance of self-driving vehicles is reduced when they are exposed to stressors that impair visibility or cause water or snow accumulation on sensor surfaces. This paper proposes a model to quantify weather precipitation, such as rain and snow, perceived by moving vehicles based on outdoor data. The modeling covers a wide range of parameters, such as varying the wind direction and realistic particle size distributions. The model allows the calculation of precipitation intensity on inclined surfaces of different orientations and on a circular driving path. The modeling results were partially validated against direct measurements carried out using a test vehicle. The model outputs showed a strong correlation with the experimental data for both rain and snow. Mitigation strategies for heavy precipitation on vehicles can be developed, and correlations between precipitation rate and accumulation level can be traced using the presented analytical model. A dimensional analysis of the problem highlighted the critical parameters that can help the design of future experiments. The obtained results highlight the importance of the angle of the sensing surface for the perceived precipitation level. The proposed model was used to analyze optimal orientations for minimization of the precipitation flux, which can help to determine the positioning of sensors on the surface of autonomous vehicles. Full article

Stuttering, a prevalent neurodevelopmental disorder, profoundly affects fluent speech, causing involuntary interruptions and recurrent sound patterns. This study addresses the critical need for the accurate classification of stuttering types. The researchers introduce “TranStutter”, a pioneering Convolution-free Transformer-based DL model, designed to excel in speech disfluency classification. Unlike conventional methods, TranStutter leverages Multi-Head Self-Attention and Positional Encoding to capture intricate temporal patterns, yielding superior accuracy. In this study, the researchers employed two benchmark datasets: the Stuttering Events in Podcasts Dataset (SEP-28k) and the FluencyBank Interview Subset. SEP-28k comprises 28,177 audio clips from podcasts, meticulously annotated into distinct dysfluent and non-dysfluent labels, including Block (BL), Prolongation (PR), Sound Repetition (SR), Word Repetition (WR), and Interjection (IJ). The FluencyBank subset encompasses 4144 audio clips from 32 People Who Stutter (PWS), providing a diverse set of speech samples. TranStutter’s performance was assessed rigorously. On SEP-28k, the model achieved an impressive accuracy of 88.1%. Furthermore, on the FluencyBank dataset, TranStutter demonstrated its efficacy with an accuracy of 80.6%. These results highlight TranStutter’s significant potential in revolutionizing the diagnosis and treatment of stuttering, thereby contributing to the evolving landscape of speech pathology and neurodevelopmental research. The innovative integration of Multi-Head Self-Attention and Positional Encoding distinguishes TranStutter, enabling it to discern nuanced disfluencies with unparalleled precision. This novel approach represents a substantial leap forward in the field of speech pathology, promising more accurate diagnostics and targeted interventions for individuals with stuttering disorders. Full article

The integration of renewable energy sources, electric vehicles, and other electrical assets has introduced complexities in monitoring and controlling power networks. Consequently, numerous grid nodes have been equipped with sensors and complex measurement systems to enhance network observability. Additionally, real-time power network simulators have become crucial tools for predicting and estimating the behavior of electrical quantities at different network components, such as nodes, branches, and assets. In this paper, a new user-friendly model for Rogowski coils is presented and validated. The model’s simplicity stems from utilizing information solely from the Rogowski coil datasheet. By establishing the input/output relationship, the output of the Rogowski coil is obtained. The effectiveness and accuracy of the proposed model are tested using both simulations and commercially available Rogowski coils. The results confirm that the model is simple, accurate, and easily implementable in various simulation environments for a wide range of applications and purposes. Full article

Synthetic aperture radar (SAR) sensor often produces a shadow in pairs with the target due to its slant-viewing imaging. As a result, shadows in SAR images can provide critical discriminative features for classifiers, such as target contours and relative positions. However, shadows possess unique properties that differ from targets, such as low intensity and sensitivity to depression angles, making it challenging to extract depth features from shadows directly using convolutional neural networks (CNN). In this paper, we propose a new SAR image-classification framework to utilize target and shadow information comprehensively. First, we design a SAR image segmentation method to extract target regions and shadow masks. Second, based on SAR projection geometry, we propose a data-augmentation method to compensate for the geometric distortion of shadows due to differences in depression angles. Finally, we introduce a feature-enhancement module (FEM) based on depthwise separable convolution (DSC) and convolutional block attention module (CBAM), enabling deep networks to fuse target and shadow features adaptively. The experimental results on the Moving and Stationary Target Acquisition and Recognition (MSTAR) dataset show that when only using target and shadow information, the published deep-learning models can still achieve state-of-the-art performance after embedding the FEM. Full article