Pub Date : 2025-01-03DOI: 10.1109/JSEN.2024.3523279
Xining Xu;Wei Liu
Switch rail is an important basic component of rail transportation. Due to its variable cross section structure, there are many guided wave modes that can propagate inside it. When the ultrasonic guided wave is used to detect the defect of switch rail, the defect echo is often superimposed with complex background signal, which is difficult to extract. To solve the problem that the time-domain baseline method needs complex temperature compensation algorithm and is difficult to be applied in engineering, this article explores a new method from the frequency domain. The Fourier transform is applied to the waveguide signal and the FFT result of a waveguide signal from a nondefective switch rail is selected as the baseline. The difference of the FFT result between the waveguide signal and the baseline is calculated by the algorithm designed, being defined as a frequency-domain operator. The results show that the frequency-domain baseline method has a comprehensive identification rate of 99.89% and that no temperature compensation is required for indoor switch rail detection. Based on this, this article proposes the wavelet baseline method that integrates time-domain and frequency-domain analysis. The 3-D waveform data of the guided wave is transformed by wavelet, the difference between the data to be recognized and the baseline data is calculated based on the corresponding segments by the algorithm designed, and the frequency-time operator is obtained. For indoor datasets, the comprehensive detection rate of the wavelet baseline method is 99.93%, and the defect discrimination is better than that of the frequency-domain baseline method. For outdoor test data collected within 28 days, the comprehensive detection rate of the wavelet baseline method is 99.2%. The defect detection experiment of the switch rail on the actual line with the accessory structure is also carried out. The results show that the wavelet baseline method can effectively identify the defects of the switch rail in service. The wavelet baseline method proposed in this article can identify the defects of the switch rail effectively by dividing the temperature interval without complicated temperature compensation algorithm, and has practical value in engineering and application.
{"title":"Method for Detecting Defects in Switch Rails Based on the Wavelet Baseline","authors":"Xining Xu;Wei Liu","doi":"10.1109/JSEN.2024.3523279","DOIUrl":"https://doi.org/10.1109/JSEN.2024.3523279","url":null,"abstract":"Switch rail is an important basic component of rail transportation. Due to its variable cross section structure, there are many guided wave modes that can propagate inside it. When the ultrasonic guided wave is used to detect the defect of switch rail, the defect echo is often superimposed with complex background signal, which is difficult to extract. To solve the problem that the time-domain baseline method needs complex temperature compensation algorithm and is difficult to be applied in engineering, this article explores a new method from the frequency domain. The Fourier transform is applied to the waveguide signal and the FFT result of a waveguide signal from a nondefective switch rail is selected as the baseline. The difference of the FFT result between the waveguide signal and the baseline is calculated by the algorithm designed, being defined as a frequency-domain operator. The results show that the frequency-domain baseline method has a comprehensive identification rate of 99.89% and that no temperature compensation is required for indoor switch rail detection. Based on this, this article proposes the wavelet baseline method that integrates time-domain and frequency-domain analysis. The 3-D waveform data of the guided wave is transformed by wavelet, the difference between the data to be recognized and the baseline data is calculated based on the corresponding segments by the algorithm designed, and the frequency-time operator is obtained. For indoor datasets, the comprehensive detection rate of the wavelet baseline method is 99.93%, and the defect discrimination is better than that of the frequency-domain baseline method. For outdoor test data collected within 28 days, the comprehensive detection rate of the wavelet baseline method is 99.2%. The defect detection experiment of the switch rail on the actual line with the accessory structure is also carried out. The results show that the wavelet baseline method can effectively identify the defects of the switch rail in service. The wavelet baseline method proposed in this article can identify the defects of the switch rail effectively by dividing the temperature interval without complicated temperature compensation algorithm, and has practical value in engineering and application.","PeriodicalId":447,"journal":{"name":"IEEE Sensors Journal","volume":"25 4","pages":"6836-6849"},"PeriodicalIF":4.3,"publicationDate":"2025-01-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143446210","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-03DOI: 10.1109/JSEN.2024.3523393
Anandita Bhardwaj;Sandeep Singh;Deepak Joshi
Cardiac prosthetic valve dysfunction (PVD) is a life-threatening complication of valve replacement surgery (VRS). It is therefore crucial to monitor the mechanical prosthetic heart valve (MPHV) functioning regularly. The standard diagnostic method, cine fluoroscopy (CF) involves X-ray exposure and may not be available to a large population. Therefore, a wearable modality like phonocardiogram (PCG) seems to be a promising alternative. The proposed work is a novel method to automate PCG-based PVD detection. 2-D convolutional neural network (CNN) is explored toward the automated classification of persistence spectrum images of the PCG. Persistence spectrum, a time-frequency representation, displays the duration for which a particular frequency is present. It enables the identification of the hidden components of a signal. This work explores persistence spectrum for PCG analysis. In all, 4215 PCG samples (2127 normal and 2088 PVD) were used for training and testing the CNN. Two AI interpretation techniques, occlusion maps and deep dream images (DDIs), are used to introduce interpretability in the DL model’s decision-making. The overall accuracy of 95.73 (SD = 7.62)% is achieved during fivefold cross-validation (CV) with the highest accuracy of 100% for three out of five folds. The performance during the leave-one-subject-out CV (LOSOCV) is 90.64 (SD = 27.98)%. Through AI interpretation, novel findings of MPHV’s PCG characteristics in the spectral domain, corresponding to cardiac events of normally functioning MPHV and PVD, are revealed, making the CNN decision more transparent. The novel explainable DL model may potentially address PVD-induced clinical burden in resource-constrained settings with no radiation exposure and can be used for screening.
{"title":"Phonocardiography-Based Automated Detection of Prosthetic Heart Valve Dysfunction Using Persistence Spectrum and Interpretable Deep CNN","authors":"Anandita Bhardwaj;Sandeep Singh;Deepak Joshi","doi":"10.1109/JSEN.2024.3523393","DOIUrl":"https://doi.org/10.1109/JSEN.2024.3523393","url":null,"abstract":"Cardiac prosthetic valve dysfunction (PVD) is a life-threatening complication of valve replacement surgery (VRS). It is therefore crucial to monitor the mechanical prosthetic heart valve (MPHV) functioning regularly. The standard diagnostic method, cine fluoroscopy (CF) involves X-ray exposure and may not be available to a large population. Therefore, a wearable modality like phonocardiogram (PCG) seems to be a promising alternative. The proposed work is a novel method to automate PCG-based PVD detection. 2-D convolutional neural network (CNN) is explored toward the automated classification of persistence spectrum images of the PCG. Persistence spectrum, a time-frequency representation, displays the duration for which a particular frequency is present. It enables the identification of the hidden components of a signal. This work explores persistence spectrum for PCG analysis. In all, 4215 PCG samples (2127 normal and 2088 PVD) were used for training and testing the CNN. Two AI interpretation techniques, occlusion maps and deep dream images (DDIs), are used to introduce interpretability in the DL model’s decision-making. The overall accuracy of 95.73 (SD = 7.62)% is achieved during fivefold cross-validation (CV) with the highest accuracy of 100% for three out of five folds. The performance during the leave-one-subject-out CV (LOSOCV) is 90.64 (SD = 27.98)%. Through AI interpretation, novel findings of MPHV’s PCG characteristics in the spectral domain, corresponding to cardiac events of normally functioning MPHV and PVD, are revealed, making the CNN decision more transparent. The novel explainable DL model may potentially address PVD-induced clinical burden in resource-constrained settings with no radiation exposure and can be used for screening.","PeriodicalId":447,"journal":{"name":"IEEE Sensors Journal","volume":"25 4","pages":"6869-6880"},"PeriodicalIF":4.3,"publicationDate":"2025-01-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143446216","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-03DOI: 10.1109/JSEN.2024.3523039
Jiaming He;Mingrui Li;Yangyang Wang;Hongyu Wang
Camera and IMU are widely used in robotics to achieve accurate and robust pose estimation. However, this fusion relies heavily on sufficient visual feature observations and precise inertial state variables. This article proposes PLE-SLAM, a real-time visual-inertial simultaneous localization and mapping (SLAM) for complex environments, which introduces line features to point-based SLAM and proposes an efficient IMU initialization method. First, we use parallel computing methods to extract point-line features and compute descriptors to ensure real-time performance. Adjacent short-line segments are merged into long-line segments for more stable tracking, and isolated short-line segments are directly eliminated. Second, to overcome rapid rotation and low-texture scenes, we estimate gyroscope bias by tightly coupling rotation preintegration and 2-D point-line observations without 3-D point cloud and vision-only rotation estimation. Accelerometer bias and gravity direction are solved by an analytical method, which is more efficient than nonlinear optimization. To improve the system’s robustness in complex environments, an improved method of dynamic feature elimination and a solution for loop detection and loop frames pose estimation using CNN and GNN are integrated into the system. The experimental results on public datasets demonstrate that PLE-SLAM achieves more than 20%~50% improvement in localization performance than ORB-SLAM3 and outperforms other state-of-the-art visual-inertial SLAM systems in most environments.
{"title":"PLE-SLAM: A Visual-Inertial SLAM Based on Point-Line Features and Efficient IMU Initialization","authors":"Jiaming He;Mingrui Li;Yangyang Wang;Hongyu Wang","doi":"10.1109/JSEN.2024.3523039","DOIUrl":"https://doi.org/10.1109/JSEN.2024.3523039","url":null,"abstract":"Camera and IMU are widely used in robotics to achieve accurate and robust pose estimation. However, this fusion relies heavily on sufficient visual feature observations and precise inertial state variables. This article proposes PLE-SLAM, a real-time visual-inertial simultaneous localization and mapping (SLAM) for complex environments, which introduces line features to point-based SLAM and proposes an efficient IMU initialization method. First, we use parallel computing methods to extract point-line features and compute descriptors to ensure real-time performance. Adjacent short-line segments are merged into long-line segments for more stable tracking, and isolated short-line segments are directly eliminated. Second, to overcome rapid rotation and low-texture scenes, we estimate gyroscope bias by tightly coupling rotation preintegration and 2-D point-line observations without 3-D point cloud and vision-only rotation estimation. Accelerometer bias and gravity direction are solved by an analytical method, which is more efficient than nonlinear optimization. To improve the system’s robustness in complex environments, an improved method of dynamic feature elimination and a solution for loop detection and loop frames pose estimation using CNN and GNN are integrated into the system. The experimental results on public datasets demonstrate that PLE-SLAM achieves more than 20%~50% improvement in localization performance than ORB-SLAM3 and outperforms other state-of-the-art visual-inertial SLAM systems in most environments.","PeriodicalId":447,"journal":{"name":"IEEE Sensors Journal","volume":"25 4","pages":"6801-6811"},"PeriodicalIF":4.3,"publicationDate":"2025-01-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143446246","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-03DOI: 10.1109/JSEN.2024.3523201
Lei Wang;Rui-Qi Wang;Hui Xiao;Chengyang Luo
In this article, a new differential composite probe with low-profile and high-sensitivity is developed. The proposed probe includes a U-shaped loop as a main element, a pair of U-shaped loops on the outermost layers and a pair of long U-shaped loops on the inner layers as parasitic elements, and two stripline terminated with two sub-miniature A (SMA) connectors. Traditionally, single-loop probes are used to sense electric- and magnetic-field components simultaneously. Here, two pairs of U-shaped loops are introduced as parasitic elements into a single-loop probe to enhance the amount of received electromagnetic signals. Moreover, these parasitic U-shaped loops on the inner layers of the probe are replaced with parasitic long U-shaped loops, thus significantly improving the detection sensitivity. Unlike the detection loops of traditional probes, which are all located on the inner layers, a pair of parasitic U-shaped loops, and the ground planes are printed together on the outermost layers in the proposed probe to reduce the probe’s profile. Because of this new design, the proposed probe only requires a five-layer printed circuit board (PCB) structure instead of a seven-layer PCB structure. A commercial high-frequency electromagnetic simulation software is used to design and simulate the proposed composite probe. A five-layer PCB structure is used to manufacture a prototype, and a self-developed near-field scanning test system with a standard microstrip line is utilized to characterize and calibrate the probe. As last, the simulation and measurement results are compared to verify the design rationality. The comparison results reveal that the designed differential composite probe has high detection sensitivity, a low profile, and the ability of measure electromagnetic-field components.
{"title":"A New Differential Composite Probe With Outstanding Advantages of High Sensitivity, Multiple Components, and Low Profile","authors":"Lei Wang;Rui-Qi Wang;Hui Xiao;Chengyang Luo","doi":"10.1109/JSEN.2024.3523201","DOIUrl":"https://doi.org/10.1109/JSEN.2024.3523201","url":null,"abstract":"In this article, a new differential composite probe with low-profile and high-sensitivity is developed. The proposed probe includes a U-shaped loop as a main element, a pair of U-shaped loops on the outermost layers and a pair of long U-shaped loops on the inner layers as parasitic elements, and two stripline terminated with two sub-miniature A (SMA) connectors. Traditionally, single-loop probes are used to sense electric- and magnetic-field components simultaneously. Here, two pairs of U-shaped loops are introduced as parasitic elements into a single-loop probe to enhance the amount of received electromagnetic signals. Moreover, these parasitic U-shaped loops on the inner layers of the probe are replaced with parasitic long U-shaped loops, thus significantly improving the detection sensitivity. Unlike the detection loops of traditional probes, which are all located on the inner layers, a pair of parasitic U-shaped loops, and the ground planes are printed together on the outermost layers in the proposed probe to reduce the probe’s profile. Because of this new design, the proposed probe only requires a five-layer printed circuit board (PCB) structure instead of a seven-layer PCB structure. A commercial high-frequency electromagnetic simulation software is used to design and simulate the proposed composite probe. A five-layer PCB structure is used to manufacture a prototype, and a self-developed near-field scanning test system with a standard microstrip line is utilized to characterize and calibrate the probe. As last, the simulation and measurement results are compared to verify the design rationality. The comparison results reveal that the designed differential composite probe has high detection sensitivity, a low profile, and the ability of measure electromagnetic-field components.","PeriodicalId":447,"journal":{"name":"IEEE Sensors Journal","volume":"25 4","pages":"7562-7568"},"PeriodicalIF":4.3,"publicationDate":"2025-01-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143446293","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-03DOI: 10.1109/JSEN.2024.3523335
Yu Wang;Shujie Liu;Shuai Lv;Gengshuo Liu
Slurry pumps are crucial in the mining industry, as their performance and reliability directly affect the efficiency and safety of mining production systems. However, existing remaining useful life (RUL) prediction models face challenges due to the scarcity of degradation samples caused by the difficulty of obtaining degradation data in industrial settings, and their inability to provide prediction result confidence intervals (CIs). This article proposes a meta transformer with uncertainty quantification (MTUQ) based on an approximate Bayesian framework. The model enhances the capability to quickly adapt to new tasks in few-shot scenarios through a dual-loop meta-learning strategy, addressing the issue of sample sparsity. Additionally, random subnetwork sampling (RSNS) is proposed to achieve approximate Bayesian posterior distribution and combines Kernel density estimation (KDE) to quantify the model’s prediction uncertainty. Experimental results on the few-shot RUL prediction of slurry pumps in actual production scenarios demonstrate that MTUQ outperforms baseline methods in handling sparse samples and quantifying uncertainty, improving its prediction accuracy and reliability.
{"title":"Few-Shot Probabilistic RUL Prediction With Uncertainty Quantification of Slurry Pumps","authors":"Yu Wang;Shujie Liu;Shuai Lv;Gengshuo Liu","doi":"10.1109/JSEN.2024.3523335","DOIUrl":"https://doi.org/10.1109/JSEN.2024.3523335","url":null,"abstract":"Slurry pumps are crucial in the mining industry, as their performance and reliability directly affect the efficiency and safety of mining production systems. However, existing remaining useful life (RUL) prediction models face challenges due to the scarcity of degradation samples caused by the difficulty of obtaining degradation data in industrial settings, and their inability to provide prediction result confidence intervals (CIs). This article proposes a meta transformer with uncertainty quantification (MTUQ) based on an approximate Bayesian framework. The model enhances the capability to quickly adapt to new tasks in few-shot scenarios through a dual-loop meta-learning strategy, addressing the issue of sample sparsity. Additionally, random subnetwork sampling (RSNS) is proposed to achieve approximate Bayesian posterior distribution and combines Kernel density estimation (KDE) to quantify the model’s prediction uncertainty. Experimental results on the few-shot RUL prediction of slurry pumps in actual production scenarios demonstrate that MTUQ outperforms baseline methods in handling sparse samples and quantifying uncertainty, improving its prediction accuracy and reliability.","PeriodicalId":447,"journal":{"name":"IEEE Sensors Journal","volume":"25 4","pages":"6122-6132"},"PeriodicalIF":4.3,"publicationDate":"2025-01-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143422824","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-03DOI: 10.1109/JSEN.2024.3523269
Sirui Wang;Guiling Sun;Liang Dong;Bowen Zheng
Multispectral pedestrian detection, which combines visible and infrared images, has demonstrated significant advantages under various lighting and weather conditions, making it a highly focused research topic in recent years. The red, green, blue (RGB)-thermal (RGB-T) modality essentially provides different descriptions of the same scene, encompassing both modality-specific and modality-consistent information. However, most existing approaches overlook the differences between these two types of information during feature fusion, leading to insufficient feature representation. To address this, we propose an illumination-aware (IA) octave fusion framework (OctNet) for RGB-T pedestrian detection. Specifically, we introduce an illumination-aware octave fusion (IA-OctFuse) module, which utilizes frequency domain analysis to separate modality-complementary target features from redundant background features. Additionally, an IA mechanism is incorporated to adaptively balance the contributions of different modalities, producing highly discriminative RGB-T fused features. Then, a multi-head dilated-convolution enhancement (MHDE) module is designed to deeply explore the spatial self-similarity of fused features, further enhancing feature representation. Extensive experiments and comparisons show that the proposed OctNet achieves state-of-the-art performance on publicly available KAIST and LLVIP pedestrian detection datasets.
{"title":"OctNet: Illumination-Aware Octave Fusion and Feature Enhancement for Multispectral Pedestrian Detection","authors":"Sirui Wang;Guiling Sun;Liang Dong;Bowen Zheng","doi":"10.1109/JSEN.2024.3523269","DOIUrl":"https://doi.org/10.1109/JSEN.2024.3523269","url":null,"abstract":"Multispectral pedestrian detection, which combines visible and infrared images, has demonstrated significant advantages under various lighting and weather conditions, making it a highly focused research topic in recent years. The red, green, blue (RGB)-thermal (RGB-T) modality essentially provides different descriptions of the same scene, encompassing both modality-specific and modality-consistent information. However, most existing approaches overlook the differences between these two types of information during feature fusion, leading to insufficient feature representation. To address this, we propose an illumination-aware (IA) octave fusion framework (OctNet) for RGB-T pedestrian detection. Specifically, we introduce an illumination-aware octave fusion (IA-OctFuse) module, which utilizes frequency domain analysis to separate modality-complementary target features from redundant background features. Additionally, an IA mechanism is incorporated to adaptively balance the contributions of different modalities, producing highly discriminative RGB-T fused features. Then, a multi-head dilated-convolution enhancement (MHDE) module is designed to deeply explore the spatial self-similarity of fused features, further enhancing feature representation. Extensive experiments and comparisons show that the proposed OctNet achieves state-of-the-art performance on publicly available KAIST and LLVIP pedestrian detection datasets.","PeriodicalId":447,"journal":{"name":"IEEE Sensors Journal","volume":"25 4","pages":"7584-7595"},"PeriodicalIF":4.3,"publicationDate":"2025-01-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143438408","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-03DOI: 10.1109/JSEN.2024.3523272
Xiao-Xiao Liao;Hong Yang;Qiang Wu;Juan Liu;Yingying Hu;Yue Zhang;Wei-Qing Liu;Yue Fu;Andrew R. Pike;Bin Liu
The advancement of artificial intelligence technology has led to the widespread adoption of deep learning techniques within spectral analysis over recent years. In this study, we introduce an advanced demodulation approach utilizing a 1-D convolutional neural network (1D-CNN) for feature extraction and the analysis of spectral signals from surface plasmon resonance (SPR) fiber refractive index (RI) sensors featuring a multimode-no-core-multimode (MNM) structure while simultaneously forecasting changes in RI due to environmental factors. Through segmentation-based predictive training on spectral signals, our approach achieves an average prediction accuracy exceeding 98%, even at low resolutions. Experimental findings demonstrate superior demodulation performance using our intelligent demodulation technique based on 1D-CNN compared to conventional methods. Furthermore, our method is adaptable across diverse and intricate structures enabling observation of parameter correlations spanning their entire range, thereby enhancing measurement capabilities within SPR sensing systems with significant potential applications.
{"title":"Convolutional Neural Network-Enabled Optical Fiber SPR Sensors for RI Prediction","authors":"Xiao-Xiao Liao;Hong Yang;Qiang Wu;Juan Liu;Yingying Hu;Yue Zhang;Wei-Qing Liu;Yue Fu;Andrew R. Pike;Bin Liu","doi":"10.1109/JSEN.2024.3523272","DOIUrl":"https://doi.org/10.1109/JSEN.2024.3523272","url":null,"abstract":"The advancement of artificial intelligence technology has led to the widespread adoption of deep learning techniques within spectral analysis over recent years. In this study, we introduce an advanced demodulation approach utilizing a 1-D convolutional neural network (1D-CNN) for feature extraction and the analysis of spectral signals from surface plasmon resonance (SPR) fiber refractive index (RI) sensors featuring a multimode-no-core-multimode (MNM) structure while simultaneously forecasting changes in RI due to environmental factors. Through segmentation-based predictive training on spectral signals, our approach achieves an average prediction accuracy exceeding 98%, even at low resolutions. Experimental findings demonstrate superior demodulation performance using our intelligent demodulation technique based on 1D-CNN compared to conventional methods. Furthermore, our method is adaptable across diverse and intricate structures enabling observation of parameter correlations spanning their entire range, thereby enhancing measurement capabilities within SPR sensing systems with significant potential applications.","PeriodicalId":447,"journal":{"name":"IEEE Sensors Journal","volume":"25 4","pages":"6371-6379"},"PeriodicalIF":4.3,"publicationDate":"2025-01-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143430409","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-03DOI: 10.1109/JSEN.2024.3523290
Wenbo Wang;Wei Wang;Xixin Yu;Weibin Zhang
In the landscape of recent technological advancements, the advent of 4-D millimeter-wave radar has ushered in a new era of data quality improvements, showcasing the potential to rival, or even surpass, Lidar systems. Despite its innovative prowess, the lower density and accuracy of 4-D millimeter-wave radar’s point clouds, in comparison to those generated by Lidar, pose significant limitations to the technology’s broader application. Addressing these constraints, our research introduces a comprehensive, end-to-end methodology for augmenting point cloud data through a fusion of monocular camera imagery and 4-D millimeter-wave radar. First, the monocular image is transformed into a pseudo-point cloud. Subsequently, features from both the radar-generated point clouds and the pseudo-point clouds are independently extracted and merged using two distinct feature extraction modules. To refine this process further, a novel loss function is designed, taking into account the global and local feature consistency between the reconstructed point cloud and the Lidar raw point cloud. The experimental results, particularly within the realms of object detection, illustrate a marked enhancement in point cloud quality over the baseline provided by native 4-D millimeter-wave radar outputs. Additionally, the application of this method to simultaneous localization and mapping (SLAM) demonstrates a significant improvement in accuracy, achieving a level of performance that is competitive with Lidar. Notably, the proposed method is low computational demand, enabling real-time inference within a mere 30 ms on resource-constrained platforms such as the NVIDIA Jetson Nano 2G.
{"title":"C4RFNet: Camera and 4D-Radar Fusion Network for Point Cloud Enhancement","authors":"Wenbo Wang;Wei Wang;Xixin Yu;Weibin Zhang","doi":"10.1109/JSEN.2024.3523290","DOIUrl":"https://doi.org/10.1109/JSEN.2024.3523290","url":null,"abstract":"In the landscape of recent technological advancements, the advent of 4-D millimeter-wave radar has ushered in a new era of data quality improvements, showcasing the potential to rival, or even surpass, Lidar systems. Despite its innovative prowess, the lower density and accuracy of 4-D millimeter-wave radar’s point clouds, in comparison to those generated by Lidar, pose significant limitations to the technology’s broader application. Addressing these constraints, our research introduces a comprehensive, end-to-end methodology for augmenting point cloud data through a fusion of monocular camera imagery and 4-D millimeter-wave radar. First, the monocular image is transformed into a pseudo-point cloud. Subsequently, features from both the radar-generated point clouds and the pseudo-point clouds are independently extracted and merged using two distinct feature extraction modules. To refine this process further, a novel loss function is designed, taking into account the global and local feature consistency between the reconstructed point cloud and the Lidar raw point cloud. The experimental results, particularly within the realms of object detection, illustrate a marked enhancement in point cloud quality over the baseline provided by native 4-D millimeter-wave radar outputs. Additionally, the application of this method to simultaneous localization and mapping (SLAM) demonstrates a significant improvement in accuracy, achieving a level of performance that is competitive with Lidar. Notably, the proposed method is low computational demand, enabling real-time inference within a mere 30 ms on resource-constrained platforms such as the NVIDIA Jetson Nano 2G.","PeriodicalId":447,"journal":{"name":"IEEE Sensors Journal","volume":"25 4","pages":"7596-7610"},"PeriodicalIF":4.3,"publicationDate":"2025-01-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143438356","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-03DOI: 10.1109/JSEN.2024.3523343
Bin Li;Xin Jiang;Yirui Du;Yanzuo Yu;Ruonan Zhang
In recent years, human activity recognition (HAR) based on Wi-Fi channel state information (CSI) has received widespread attention due to its non-intrusive and privacy-preserving nature. However, many CSI activity recognition models based on traditional methods and deep learning face two major challenges: first, most studies rely on commercial Wi-Fi network cards, which usually have only three RF ports, resulting in limited spatiotemporal resolution of the acquired CSI; second, some of the studies require complex CSI processing, which increases the network parameters, significantly lengthens the recognition time and raises the deployment costs. To this end, this study develops a lightweight high-resolution recognition model Wi-SSR based on Wi-Fi. To improve the spatiotemporal resolution of CSI, we introduce array antennas and solve the problem of coherent signals that are difficult to distinguish by communication algorithms. The lightweight CSI processing strategy proposed by Wi-SSR is able to efficiently extract the main relevant features while compressing the model size. We combine 3-D convolution with a convolutional block attention module (CBAM) to extract activity-related information from CSI and employ knowledge distillation to migrate the features learned from this model to a simple model. Extensive experimental results show that our system outperforms other deep learning models in terms of efficiency, with recognition accuracy up to 98.6% on six different types of human activities.
{"title":"Wi-SSR: Wi-Fi-Based Lightweight High-Resolution Model for Human Activity Recognition","authors":"Bin Li;Xin Jiang;Yirui Du;Yanzuo Yu;Ruonan Zhang","doi":"10.1109/JSEN.2024.3523343","DOIUrl":"https://doi.org/10.1109/JSEN.2024.3523343","url":null,"abstract":"In recent years, human activity recognition (HAR) based on Wi-Fi channel state information (CSI) has received widespread attention due to its non-intrusive and privacy-preserving nature. However, many CSI activity recognition models based on traditional methods and deep learning face two major challenges: first, most studies rely on commercial Wi-Fi network cards, which usually have only three RF ports, resulting in limited spatiotemporal resolution of the acquired CSI; second, some of the studies require complex CSI processing, which increases the network parameters, significantly lengthens the recognition time and raises the deployment costs. To this end, this study develops a lightweight high-resolution recognition model Wi-SSR based on Wi-Fi. To improve the spatiotemporal resolution of CSI, we introduce array antennas and solve the problem of coherent signals that are difficult to distinguish by communication algorithms. The lightweight CSI processing strategy proposed by Wi-SSR is able to efficiently extract the main relevant features while compressing the model size. We combine 3-D convolution with a convolutional block attention module (CBAM) to extract activity-related information from CSI and employ knowledge distillation to migrate the features learned from this model to a simple model. Extensive experimental results show that our system outperforms other deep learning models in terms of efficiency, with recognition accuracy up to 98.6% on six different types of human activities.","PeriodicalId":447,"journal":{"name":"IEEE Sensors Journal","volume":"25 4","pages":"6556-6571"},"PeriodicalIF":4.3,"publicationDate":"2025-01-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143446245","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-03DOI: 10.1109/JSEN.2024.3523479
Qing Shi;Shilun Feng;Jianlong Zhao
This article proposed a silicon-on-insulator (SOI)-based optofluidic 1-D photonic crystal slab biosensor array structure for multiple tumor markers detection. The array consists of multiple expandable sensing branches, each composed of a nanobeam resonator transducer with excellent detection limit, a filter with low sidelobe jitter, and a microfluidics roof. Using the 3-D finite-difference time-domain (FDTD) method, a 1-D photonic crystal slot nanobeam resonator transducer consisting of a circular hole array linearly decreasing from the center to both ends was obtained. Under the influence of absorption loss of biological solution, the transducer works in the communication E-band, with the Q-value up to 10487, refractive index sensitivity of 355 nm/RIU, and refractive index detection limit of $2.61times 10^{-{5}}$ RIU, corresponding to the detection of fg/mL carcinoembryonic antigen, which can be directly used for the detection of tumor marker under the capture of antibody probes in microfluidics chip. By optimizing the apertures on both sides of 1-D photonic crystals with a tapered shape, a cutoff filter with low sidelobe jitter can effectively filter out the high-order resonant peaks of the transducer, forming a large free spectral range (FSR). More importantly, the aforementioned sensing branch can be extended into arrays based on the frequency band effect of photonic crystals. This article provided the expansion method and examples to verify that the extended branches have equally excellent detection performance and analyzed the reasons why the sensing array has high MEMS preparation robustness. The array structure provides a good choice for label-free point-of-care detection of multiple tumor markers.
{"title":"Engineering Design of an Expandable 1-D Photonic Crystal Slab Biosensor Array for Joint Detection of Multiple Tumor Markers","authors":"Qing Shi;Shilun Feng;Jianlong Zhao","doi":"10.1109/JSEN.2024.3523479","DOIUrl":"https://doi.org/10.1109/JSEN.2024.3523479","url":null,"abstract":"This article proposed a silicon-on-insulator (SOI)-based optofluidic 1-D photonic crystal slab biosensor array structure for multiple tumor markers detection. The array consists of multiple expandable sensing branches, each composed of a nanobeam resonator transducer with excellent detection limit, a filter with low sidelobe jitter, and a microfluidics roof. Using the 3-D finite-difference time-domain (FDTD) method, a 1-D photonic crystal slot nanobeam resonator transducer consisting of a circular hole array linearly decreasing from the center to both ends was obtained. Under the influence of absorption loss of biological solution, the transducer works in the communication E-band, with the Q-value up to 10487, refractive index sensitivity of 355 nm/RIU, and refractive index detection limit of <inline-formula> <tex-math>$2.61times 10^{-{5}}$ </tex-math></inline-formula> RIU, corresponding to the detection of fg/mL carcinoembryonic antigen, which can be directly used for the detection of tumor marker under the capture of antibody probes in microfluidics chip. By optimizing the apertures on both sides of 1-D photonic crystals with a tapered shape, a cutoff filter with low sidelobe jitter can effectively filter out the high-order resonant peaks of the transducer, forming a large free spectral range (FSR). More importantly, the aforementioned sensing branch can be extended into arrays based on the frequency band effect of photonic crystals. This article provided the expansion method and examples to verify that the extended branches have equally excellent detection performance and analyzed the reasons why the sensing array has high MEMS preparation robustness. The array structure provides a good choice for label-free point-of-care detection of multiple tumor markers.","PeriodicalId":447,"journal":{"name":"IEEE Sensors Journal","volume":"25 4","pages":"5986-5994"},"PeriodicalIF":4.3,"publicationDate":"2025-01-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143422844","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：