IEEE Sensors Journal最新文献

{"title":"A Lightweight Network With Dual-Mixup Data Augmentation for Fault Diagnosis of Rotating Machinery Under Unknown Operating Conditions","authors":"Ximing Sun;Yifei Yang;Minjia Tan;Jin Zhu;Haiguo Jing","doi":"10.1109/JSEN.2025.3592948","DOIUrl":"https://doi.org/10.1109/JSEN.2025.3592948","url":null,"abstract":"In practical industrial scenarios, achieving an end-to-end intelligent fault diagnosis system remains a challenge due to the discrepancies in feature-space distributions caused by varying working conditions and limited computational resources. Existing domain generalization (DG) methods often overlook the tradeoff between model performance and deployment efficiency. Although lightweight convolutional networks provide high inference efficiency, they exhibit limitations in feature representation and generalization performance. Therefore, this article proposes a lightweight network architecture based on dual-mixup data augmentation (LNDMDA) to enhance the adaptability of fault diagnosis models in cross-domain environments. First, a dual-mixup data augmentation mechanism combining cross-domain confusion and intradomain confusion is constructed to significantly enhance the generalization capability of the model in complex domain transfer scenarios. Second, a lightweight feature extraction module combining partial convolution (PConv) and depthwise separable convolution (DSC) is designed to preserve critical diagnostic features of the signal while reducing computational complexity. In addition, this article introduces a lightweight subspace division attention mechanism, which divides the feature map into multiple subspaces and calculates attention weights for each subspace separately, thereby efficiently capturing multiscale fault features. Finally, comparative experiments on three public datasets demonstrate that LNDMDA consistently outperforms existing lightweight models and DG methods across multiple transfer tasks, highlighting its advantages of compact model size, low computational cost, and excellent generalization performance.","PeriodicalId":447,"journal":{"name":"IEEE Sensors Journal","volume":"25 17","pages":"34196-34208"},"PeriodicalIF":4.3,"publicationDate":"2025-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144934545","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}

{"title":"Design and Simulation of Piezoelectric Acoustic Emission Sensor","authors":"Chunli Wang;Pu Hu;Pengfei Lv;Yi Liu;Shun Zuo;Yanjun Zhang","doi":"10.1109/JSEN.2025.3592934","DOIUrl":"https://doi.org/10.1109/JSEN.2025.3592934","url":null,"abstract":"Piezoelectric acoustic emission (AE) sensors, characterized by their simple structure, high sensitivity, and wide frequency bandwidth, demonstrate significant potential for applications in acoustic measurement, ultrasonic testing, and biomedical fields. In this study, epoxy resin doped with tungsten powder was adopted and used as the backing layer material, to enhance sensor sensitivity and reduce noise through material optimization. Based on the simulation model using COMSOL software, the influence mechanisms of the backing layer’s material composition and structural parameters on sensor performance were thoroughly investigated. Investigations revealed that both the mixing ratio of epoxy resin-tungsten powder and the backing layer thickness significantly influenced sensor sensitivity. Moreover, optimized backing layer designs effectively absorb and transmit AE signal energy while minimizing noise interference, thereby enhancing the overall sensor performance. Furthermore, the consistency between simulation results and experimental validation further demonstrated the effectiveness and accuracy of COMSOL-based simulation analysis in sensor design optimization.","PeriodicalId":447,"journal":{"name":"IEEE Sensors Journal","volume":"25 17","pages":"32117-32126"},"PeriodicalIF":4.3,"publicationDate":"2025-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144990122","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}

{"title":"Suitable Matching Area Selection Algorithm Based on Terrain Triangulation Model for Terrain-Aided Inertial Navigation","authors":"Zhaonan Jiao;Shengwu Zhao;Yu Wang;Zhihong Deng","doi":"10.1109/JSEN.2025.3592681","DOIUrl":"https://doi.org/10.1109/JSEN.2025.3592681","url":null,"abstract":"The terrain suitable matching area (SMA) selection algorithm is one of the key technologies for underwater terrain-aided inertial navigation to achieve good navigation and positioning results. Within the SMA, the accuracy of positioning and the matching rate of the terrain matching algorithm can be improved. Traditional threshold selection algorithms usually use statistical information on various features of the terrain field as evaluation indicators. However, these methods are difficult to reflect the spatial information of underwater terrain, while misclassification often occurs in the selection process. This article proposes an SMA selection algorithm for the underwater terrain-aided inertial navigation based on terrain triangulation, which converts the digital terrain map (DTM) from a raster map to terrain triangulation model. This is a novel approach that extracts eight types of spatial features from a triangulation model to analyze the characteristics of terrain. The Gaussian mixture model (GMM) clustering method is used to automatically select and divide the SMA based on the feature parameter matrix. Simulation shows that the proposed algorithm can achieve three-class classification of the terrain field and the select more suitable-matching area, which totally cover the result of traditional and other mainstream algorithm. Marine experiment verified the proposed algorithm.","PeriodicalId":447,"journal":{"name":"IEEE Sensors Journal","volume":"25 17","pages":"33339-33349"},"PeriodicalIF":4.3,"publicationDate":"2025-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144934397","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}

{"title":"Adaptive Graph Reconstruction-Enhanced Multiscale GCN and DCN Fusion Network for Hyperspectral Image Classification","authors":"Kewen Qu;Mingming Ding;Huiyang Wang;Xiaojuan Luo;Wenxing Bao","doi":"10.1109/JSEN.2025.3592786","DOIUrl":"https://doi.org/10.1109/JSEN.2025.3592786","url":null,"abstract":"In hyperspectral image (HSI) classification research, graph convolutional networks (GCNs) and convolutional neural networks (CNNs) fusion networks have become a research hotspot due to their advantages in complementing spatial and spectral information. However, existing methods still face dual challenges. On the GCN side, performance heavily depends on the initial graph quality. Most methods rely on a single similarity metric, leading to inaccurate node relationships and limiting global information perception. On the CNN side, regular convolutional kernels struggle to efficiently capture local features in irregular regions, thus restricting the precise representation of spatial and spectral information. To overcome the aforementioned limitations, this article proposes an adaptive graph reconstruction enhanced multiscale GCN and DCN fusion network (GRGDFN) for HSI classification. Specifically, the dynamic spectral feature extraction module (DSFEM) harnesses adaptive sampling via lightweight deformable convolutional networks (DCNs) to capture fine-grained spectral features with spatially irregular distributions. Simultaneously, the multiscale spatial–spectral graph convolutional module (MSSGCM) employs dual-view dynamic adaptive perception (DVDAP) to capture long-range dependencies across multiple spatial scales, thereby enhancing the global perception ability of features. Furthermore, the hierarchical feature fusion module (HFFM) fuses complementary features from DSFEM (local fine-grained) and MSSGCM (global topological relationships), generating enriched joint representations. Subsequently, the graph structure reconstruction module based on geometric and label information (GSRM-GL) employs the fused features to dynamically optimize node connections, reconstructing the initial graph structure to compensate for its limitations in node similarity measurement, with the resulting graph then fed into the model for retraining to enhance classification accuracy. Finally, a spatial–spectral joint loss (SSJL) regularization function is proposed to further optimize classification accuracy while maintaining spatial–spectral consistency. Comprehensive experiments conducted on five representative HSI datasets demonstrate that the proposed method significantly outperforms state-of-the-art deep learning approaches in terms of classification accuracy and generalization capability, with substantial overall accuracy (OA) improvements of 0.7%–33.8% for Indian Pines (IP), 1.16%–38.14% for University of Pavia (UP), 2.26%–28.11% for Salinas, 1.02%–42.88% for Kennedy Space Center (KSC), and 2.43%–31.48% for WHU-Hi-HanChuan under identical sampling conditions.","PeriodicalId":447,"journal":{"name":"IEEE Sensors Journal","volume":"25 17","pages":"34070-34090"},"PeriodicalIF":4.3,"publicationDate":"2025-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144934421","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}

{"title":"Enhancing Multicamera-Based 3-D Detection in Low-Light Driving Scenarios With Self-Supervised Illumination Estimation","authors":"Yalong Ma;Xuan Liu;Zhongxia Xiong;Ziying Yao;Xinkai Wu","doi":"10.1109/JSEN.2025.3592706","DOIUrl":"https://doi.org/10.1109/JSEN.2025.3592706","url":null,"abstract":"Vision-based multisensor system for bird’s eye view (BEV) 3-D perception is gaining attention as an alternative to high-cost multi-LiDAR systems and has achieved notable success. However, there is a significant safety concern for future image-based BEV autonomous driving in low-light conditions (such as nighttime) while the limited research on BEV detectors for these scenes. In this article, we attempt to enhance low-light BEV perception with illumination-guided feature fusion. We propose RetBEV, which uses illumination information generated based on the retinex theory to enhance the model’s robustness in varying lighting conditions. Additionally, to address the illumination estimation discontinuity from multiview images that can adversely affect detection, we propose the multiview self-balancing retinex (MVB-retinex), a self-supervised learning mechanism which balances illumination estimation by leveraging overlapping regions between adjacent images. Notably, RetBEV is a plug-and-play module that can be applied to many image-based BEV detector methods and does not require any additional ground truth (GT) supervision. We conduct extensive experiments on the nuScenes dataset, validating our algorithm in nighttime and daytime scenes. Compared to the baseline, our algorithm achieves a 2.34% increase in mean average precision (mAP) on the validation set with minimal computational cost, especially showing a 3.60% improvement in nighttime scene. The experiments demonstrate that our RetBEV effectively improves detection performance in low-light conditions and enhances performance under normal illumination, indicating increased robustness of the BEV detector.","PeriodicalId":447,"journal":{"name":"IEEE Sensors Journal","volume":"25 17","pages":"34187-34195"},"PeriodicalIF":4.3,"publicationDate":"2025-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144934504","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}

{"title":"The Limiting Mechanism and Methods to Expand the Range of the Sensitivity Control in Six-Electrode MET Seismic Sensor","authors":"Anna V. Akinina;Ivan V. Egorov;Alexander S. Bugaev;Vadim M. Agafonov","doi":"10.1109/JSEN.2025.3577088","DOIUrl":"https://doi.org/10.1109/JSEN.2025.3577088","url":null,"abstract":"A design of the six-electrode seismic sensor based on molecular-electronic technology (MET) was proposed earlier. The sensitive electrochemical cell of this sensor consists of two cathodes, two anodes, and two additional electrodes. The distinctive feature of the six-electrode sensor is the possibility to control the sensitivity by electrical signals. Meanwhile, the range in which the sensitivity could be changed is very limited. In this article, an investigation has been conducted into the mechanism constraining the sensitivity control range, employing a combination of experimental measurements and theoretical analysis. The findings reveal that the application of a positive voltage between the anodes and additional electrodes increases the conversion factor. This phenomenon arises from the generation of active ions on the anodes, concomitant with their consumption on the additional electrodes. As the applied voltage increases, the current flowing through additional electrodes becomes constrained by a diffusion factor, thereby restricting the production of active ions on the anodes. Methods of expanding the sensitivity control range involve increasing the surface area of additional electrodes and/or decreasing the surface area of the anodes.","PeriodicalId":447,"journal":{"name":"IEEE Sensors Journal","volume":"25 15","pages":"28576-28584"},"PeriodicalIF":4.3,"publicationDate":"2025-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144758229","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}

{"title":"IEEE Sensors Council","authors":"","doi":"10.1109/JSEN.2025.3589920","DOIUrl":"https://doi.org/10.1109/JSEN.2025.3589920","url":null,"abstract":"","PeriodicalId":447,"journal":{"name":"IEEE Sensors Journal","volume":"25 15","pages":"C3-C3"},"PeriodicalIF":4.3,"publicationDate":"2025-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11106320","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144750945","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Guest Editorial: Special Issue on Sensors for Sustainable Future","authors":"Sofia Sandhu;Danilo Demarchi;Victor Grimblatt;Shiv Gobind Singh;Ravinder Dahiya","doi":"10.1109/JSEN.2025.3586113","DOIUrl":"https://doi.org/10.1109/JSEN.2025.3586113","url":null,"abstract":"","PeriodicalId":447,"journal":{"name":"IEEE Sensors Journal","volume":"25 15","pages":"27791-27791"},"PeriodicalIF":4.3,"publicationDate":"2025-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11106293","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144758418","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Tampering Detection Framework for Digital Audio Signals Under Low-SNR Conditions","authors":"Bing Li;Junfeng Duan;Wei Qiu;He Yin;Wenxuan Yao","doi":"10.1109/JSEN.2025.3592826","DOIUrl":"https://doi.org/10.1109/JSEN.2025.3592826","url":null,"abstract":"Extracting the electric network frequency (ENF) from digital audio signals is a crucial method in audio forensics. However, the ENF signal is highly susceptible to noise, making it difficult to establish an effective matching relationship with the reference frequency. This poses a significant challenge to the effectiveness of audio forensic frameworks. To address this problem, a low signal-to-noise ratio (SNR) digital audio tampering forensics (DATF) framework is proposed in this article. First, an improved chirp <italic>Z</i>-transform (ICZT) method is proposed to extract the ENF signal from audio under low-SNR conditions. Subsequently, a dual-sampling isolation forest (DSIF) method is proposed to identify potential outliers by integrating bootstrap sampling with conventional random sampling. This approach enhances the perception of local data variations, thereby improving anomaly detection accuracy. Finally, a real-world digital audio dataset collected from low-SNR conditions is employed for tampering forensics. The experimental results demonstrate that the proposed DATF framework exhibits superior performance in ENF extraction, outlier detection, and tampering forensics compared with several state-of-the-art methods.","PeriodicalId":447,"journal":{"name":"IEEE Sensors Journal","volume":"25 17","pages":"32157-32166"},"PeriodicalIF":4.3,"publicationDate":"2025-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144990216","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}

{"title":"Development of Scalable Torque Sensor for Radial Fans","authors":"Eren Kaya;Bünyamin Öztürk;Sezcan Yılmaz","doi":"10.1109/JSEN.2025.3592855","DOIUrl":"https://doi.org/10.1109/JSEN.2025.3592855","url":null,"abstract":"To evaluate fan performance, key parameters such as airflow rate, pressure, and temperature at both the inlet and outlet, along with motor speed and power, are typically measured while the fan is operated at specific speeds by an electric motor. In theory, the shaft torque of the motor can be calculated from the measured values. However, in practice, the shaft torque of the motor varies depending on motor efficiency and fan efficiency. The complex relationship between motor efficiency and fan efficiency necessitates the mechanical measurement of braking torque in a motor driven fan system. This article proposes a scalable measurement unit that can be used for mechanical torque measurement in radial fans. In this context, the proposed fan torque measurement unit is built as a prototype by performing kinematic and dynamic analysis. For the testing of the unit, forward curved blade radial fans preferred in braking applications were used. The electrical voltage and current of the dc motor were measured and the power consumption values were calculated accordingly. Furthermore, a vane-like air intake mechanism is implemented to adjust the air flow. This allowed the demonstration that varying speed–torque characteristics can be achieved using the same fan-motor configuration. Overall, the results show that torque estimation based solely on electrical power measurement is unreliable, underscoring the importance of direct mechanical torque measurement. The study offers important contributions to the braking torque measurement of the fans and to the efficiency analysis of fan-motor systems.","PeriodicalId":447,"journal":{"name":"IEEE Sensors Journal","volume":"25 17","pages":"32426-32435"},"PeriodicalIF":4.3,"publicationDate":"2025-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144990250","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}