Pub Date : 2025-12-10DOI: 10.1109/LSENS.2025.3642424
M. Jaswanth Kumar;Satyam Singh;Ahnaf Saneen;Della Thomas
A digital twin (DT)-based framework for autonomous drone navigation in GPS-denied indoor environments is presented in this letter. A real-time virtual replica of the drone enables precise control, trajectory optimization, and feedback. A 224 cm × 224 cm arena with ArUco markers defines the coordinate system, while an overhead camera and OpenCV provide vision-based localization. The ESP32-controlled drone uses the YOLOv11-nano model for obstacle detection and a lightweight transformer model, Depth Anything (Lihe-Young-small-hf), for monocular depth estimation—eliminating the need for LiDAR or stereo sensors. Detected obstacles are mapped into a 3-D grid for Dijkstra-based path planning. Real-time synchronization between the physical drone and its DT is achieved via message queuing telemetry transport (MQTT) within a robot operating system–Gazebo environment. The proposed DT system achieves an RMS trajectory deviation of approximately 0.015 m, representing an order-of-magnitude improvement compared with DT-based uncrewed aerial vehicle (UAV) navigation studies under similar experimental conditions, and maintains stable detection accuracy (mean average precision ≈ 0.994) throughout the maneuver. The proposed system offers a scalable low-cost solution for indoor UAV autonomy with potential applications in warehouse automation, disaster management, and intelligent surveillance.
在这封信中提出了一个基于数字孪生（DT）的框架，用于在gps拒绝的室内环境中自主无人机导航。无人机的实时虚拟复制品能够实现精确控制、轨迹优化和反馈。带有ArUco标记的224 cm × 224 cm竞技场定义了坐标系统，而头顶摄像机和OpenCV提供基于视觉的定位。esp32控制的无人机使用YOLOv11-nano模型进行障碍物检测，并使用轻型变压器模型Depth Anything （lieh - young -small-hf）进行单眼深度估计，从而消除了对激光雷达或立体传感器的需求。检测到的障碍物被映射到三维网格中，用于基于dijkstra的路径规划。物理无人机与其DT之间的实时同步是通过机器人操作系统gazebo环境中的消息队列遥测传输（MQTT）实现的。所提出的DT系统的RMS轨迹偏差约为0.015 m，与相似实验条件下基于DT的无人机（UAV）导航研究相比，提高了一个数量级，并且在整个机动过程中保持稳定的探测精度（平均精度≈0.994）。该系统为室内无人机自主提供了可扩展的低成本解决方案，在仓库自动化、灾害管理和智能监视方面具有潜在的应用前景。
{"title":"Digital Twin for Drone Indoor Autonomous Navigation","authors":"M. Jaswanth Kumar;Satyam Singh;Ahnaf Saneen;Della Thomas","doi":"10.1109/LSENS.2025.3642424","DOIUrl":"https://doi.org/10.1109/LSENS.2025.3642424","url":null,"abstract":"A digital twin (DT)-based framework for autonomous drone navigation in GPS-denied indoor environments is presented in this letter. A real-time virtual replica of the drone enables precise control, trajectory optimization, and feedback. A 224 cm × 224 cm arena with ArUco markers defines the coordinate system, while an overhead camera and OpenCV provide vision-based localization. The ESP32-controlled drone uses the YOLOv11-nano model for obstacle detection and a lightweight transformer model, Depth Anything (Lihe-Young-small-hf), for monocular depth estimation—eliminating the need for LiDAR or stereo sensors. Detected obstacles are mapped into a 3-D grid for Dijkstra-based path planning. Real-time synchronization between the physical drone and its DT is achieved via message queuing telemetry transport (MQTT) within a robot operating system–Gazebo environment. The proposed DT system achieves an RMS trajectory deviation of approximately 0.015 m, representing an order-of-magnitude improvement compared with DT-based uncrewed aerial vehicle (UAV) navigation studies under similar experimental conditions, and maintains stable detection accuracy (mean average precision ≈ 0.994) throughout the maneuver. The proposed system offers a scalable low-cost solution for indoor UAV autonomy with potential applications in warehouse automation, disaster management, and intelligent surveillance.","PeriodicalId":13014,"journal":{"name":"IEEE Sensors Letters","volume":"10 2","pages":"1-4"},"PeriodicalIF":2.2,"publicationDate":"2025-12-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145929579","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-09DOI: 10.1109/LSENS.2025.3642035
Navneet Gandhi;P. N. Kondekar
Random variability is a critical concern in aggressively scaled devices, as it directly impacts yield, reliability, and sensing performance. This letter investigates the combined influence of metal grain granularity (MGG) and polarization gradient within the ferroelectric (FE) layer on the reliability of a proposed junctionless (JL) negative capacitance (NC) FinFET (JLNC-FinFET)-based hydrogen gas (H2) sensor. A previously fabricated JL-FinFET serves as the baseline structure for this study. Variability induced by MGG, dictated by grain size (G) and crystallographic orientation, is further intensified by the rise in gate-induced drain leakage current due to the spatial (nonuniform) distribution of polarization inside the FE layer, leading to stronger electrostatic fluctuations and reduced sensing stability. A palladium catalytic gate facilitates hydrogen diffusion, forming an interfacial dipole layer that modulates the gate work function and alters the sensor response. Device characteristics are evaluated for hydrogen concentrations ranging from 1.00 to 1.02 ppm using 3-D Sentaurus TCAD simulations, providing new insights into reliability-aware modeling of JLNC-FinFET-based gas sensors.
{"title":"Analysis of Polarization Gradient Effect on MGG-Induced Reliability Variations in JLNC-FinFET H2 Gas Sensors","authors":"Navneet Gandhi;P. N. Kondekar","doi":"10.1109/LSENS.2025.3642035","DOIUrl":"https://doi.org/10.1109/LSENS.2025.3642035","url":null,"abstract":"Random variability is a critical concern in aggressively scaled devices, as it directly impacts yield, reliability, and sensing performance. This letter investigates the combined influence of metal grain granularity (MGG) and polarization gradient within the ferroelectric (FE) layer on the reliability of a proposed junctionless (JL) negative capacitance (NC) FinFET (JLNC-FinFET)-based hydrogen gas (H<sub>2</sub>) sensor. A previously fabricated JL-FinFET serves as the baseline structure for this study. Variability induced by MGG, dictated by grain size (G) and crystallographic orientation, is further intensified by the rise in gate-induced drain leakage current due to the spatial (nonuniform) distribution of polarization inside the FE layer, leading to stronger electrostatic fluctuations and reduced sensing stability. A palladium catalytic gate facilitates hydrogen diffusion, forming an interfacial dipole layer that modulates the gate work function and alters the sensor response. Device characteristics are evaluated for hydrogen concentrations ranging from 1.00 to 1.02 ppm using 3-D Sentaurus TCAD simulations, providing new insights into reliability-aware modeling of JLNC-FinFET-based gas sensors.","PeriodicalId":13014,"journal":{"name":"IEEE Sensors Letters","volume":"10 1","pages":"1-4"},"PeriodicalIF":2.2,"publicationDate":"2025-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145830970","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-09DOI: 10.1109/LSENS.2025.3642276
Vennila Preethi Samuel;Gowri Annasamy
Creatinine, a waste product derived from protein and muscle metabolism, is a significant biomarker of kidney related diseases, with normal clinical values ranging in micromolar concentrations. The limitations of traditional kidney diagnostic modalities include the usage of more reagents, larger interference, and the requirement of expertise for diagnosis and interpretation. Recently, surface-enhanced Raman spectroscopy (SERS) has been explored for onsite detection of biomolecules with minimal sample preparation, achieving single molecule sensitivity. However, defining distinct Raman characteristic peaks from the complex structure of individual biomolecules and enhancing the weak Raman signal for single molecule detection are challenging. Therefore, this study focuses on the automated detection of creatinine using Raman spectral peaks obtained from a graphene oxide gold nanocomposite (GOAu)-coated SERS substrate. The GOAu substrate enhances the weak Raman signal, allowing for the identification of inherent peaks of creatinine at 604, 678, 836, and 904 cm−1. In addition, a deep learning feedforward neural network, utilizing rectified linear unit (ReLU) activation, was additionally employed to enable the classification and detection of ultra-low creatinine concentrations with a limit of detection (LoD) of 1 pM, where characteristic Raman peaks are not clearly distinct due to low signal-to-noise, and achieved an accuracy of 98%. This promotes the proposed sensor ultra-sensitive detection of creatinine, offering early diagnosis of kidney-related diseases.
{"title":"Ultrasensitive Detection of Creatinine Using Deep Learning-Integrated Graphene Oxide Gold Nanocomposites SERS Sensor","authors":"Vennila Preethi Samuel;Gowri Annasamy","doi":"10.1109/LSENS.2025.3642276","DOIUrl":"https://doi.org/10.1109/LSENS.2025.3642276","url":null,"abstract":"Creatinine, a waste product derived from protein and muscle metabolism, is a significant biomarker of kidney related diseases, with normal clinical values ranging in micromolar concentrations. The limitations of traditional kidney diagnostic modalities include the usage of more reagents, larger interference, and the requirement of expertise for diagnosis and interpretation. Recently, surface-enhanced Raman spectroscopy (SERS) has been explored for onsite detection of biomolecules with minimal sample preparation, achieving single molecule sensitivity. However, defining distinct Raman characteristic peaks from the complex structure of individual biomolecules and enhancing the weak Raman signal for single molecule detection are challenging. Therefore, this study focuses on the automated detection of creatinine using Raman spectral peaks obtained from a graphene oxide gold nanocomposite (GOAu)-coated SERS substrate. The GOAu substrate enhances the weak Raman signal, allowing for the identification of inherent peaks of creatinine at 604, 678, 836, and 904 cm<sup>−1</sup>. In addition, a deep learning feedforward neural network, utilizing rectified linear unit (ReLU) activation, was additionally employed to enable the classification and detection of ultra-low creatinine concentrations with a limit of detection (LoD) of 1 pM, where characteristic Raman peaks are not clearly distinct due to low signal-to-noise, and achieved an accuracy of 98%. This promotes the proposed sensor ultra-sensitive detection of creatinine, offering early diagnosis of kidney-related diseases.","PeriodicalId":13014,"journal":{"name":"IEEE Sensors Letters","volume":"10 1","pages":"1-4"},"PeriodicalIF":2.2,"publicationDate":"2025-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145830832","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-08DOI: 10.1109/LSENS.2025.3641436
Alexander Kozlov;Sergey Fedorov;Fedor Kapralov
We share experimental results and statistical analysis for 36 stock ST Microelectronics ISM330DHCX industrial-grade 6-axis MEMS inertial measurement units calibration over a time span of two years. We analyze long-term variation of accelerometer and gyroscope biases, scaling and axial misalignment. Our data confirm that all error parameters remain well below specifications, and whithin them, there exist rare statistically significant long-term deviations.
{"title":"Empirical Long-Term Stability of Industrial-Grade STM ISM330DHCX MEMS Inertial Sensor Calibration Parameters","authors":"Alexander Kozlov;Sergey Fedorov;Fedor Kapralov","doi":"10.1109/LSENS.2025.3641436","DOIUrl":"https://doi.org/10.1109/LSENS.2025.3641436","url":null,"abstract":"We share experimental results and statistical analysis for 36 stock ST Microelectronics ISM330DHCX industrial-grade 6-axis MEMS inertial measurement units calibration over a time span of two years. We analyze long-term variation of accelerometer and gyroscope biases, scaling and axial misalignment. Our data confirm that all error parameters remain well below specifications, and whithin them, there exist rare statistically significant long-term deviations.","PeriodicalId":13014,"journal":{"name":"IEEE Sensors Letters","volume":"10 1","pages":"1-4"},"PeriodicalIF":2.2,"publicationDate":"2025-12-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145830907","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-08DOI: 10.1109/LSENS.2025.3641837
Jitendra B. Zalke;Dinesh R. Rotake;Khushi N. Mahule;Madhura A. Ambadkar;Aditi P. Wanjari;Manashwi A. Patle;Mangesh B. Thakre
This study presents a facile green synthesis approach for developing a copper oxide (CuO) porous microflowers (PMFs) based enzymatic glucose biosensor, functionalized with ZnO nanofibers. The CuO-PMFs were synthesized using an ecofriendly method, utilizing plant extracts as reducing agents, ensuring biocompatibility and minimizing environmental impact. These CuO PMFs were then integrated with ZnO nanofibers, known for their excellent electron mobility and high surface area, to enhance the biosensor's performance. The hybrid nanomaterials were employed to immobilize glucose oxidase (GOx) enzymes, facilitating the efficient electrochemical detection of glucose on printed circuit board (PCB) based interdigitated electrodes (IDEs). The resulting biosensor was tested for its impedance change, which showed the linear range of 10–250 µM, demonstrated sensitivity of 58.131 KΩ µM−1 cm−2, a low detection limit of 117 nM, and percentage relative standard deviation of 1.56% showing good stability, making it suitable for monitoring glucose levels in biomedical applications. The green synthesis route not only contributes to sustainability but also provides a cost-effective and scalable method for fabricating high-performance biosensors, offering significant potential for noninvasive glucose monitoring in diabetic care.
{"title":"Green Synthesis of CuO Porous Microflowers on PCB-Based Interdigitated Electrodes for Noninvasive Glucose Sensing","authors":"Jitendra B. Zalke;Dinesh R. Rotake;Khushi N. Mahule;Madhura A. Ambadkar;Aditi P. Wanjari;Manashwi A. Patle;Mangesh B. Thakre","doi":"10.1109/LSENS.2025.3641837","DOIUrl":"https://doi.org/10.1109/LSENS.2025.3641837","url":null,"abstract":"This study presents a facile green synthesis approach for developing a copper oxide (CuO) porous microflowers (PMFs) based enzymatic glucose biosensor, functionalized with ZnO nanofibers. The CuO-PMFs were synthesized using an ecofriendly method, utilizing plant extracts as reducing agents, ensuring biocompatibility and minimizing environmental impact. These CuO PMFs were then integrated with ZnO nanofibers, known for their excellent electron mobility and high surface area, to enhance the biosensor's performance. The hybrid nanomaterials were employed to immobilize glucose oxidase (GOx) enzymes, facilitating the efficient electrochemical detection of glucose on printed circuit board (PCB) based interdigitated electrodes (IDEs). The resulting biosensor was tested for its impedance change, which showed the linear range of 10–250 µM, demonstrated sensitivity of 58.131 KΩ µM<sup>−1</sup> cm<sup>−2</sup>, a low detection limit of 117 nM, and percentage relative standard deviation of 1.56% showing good stability, making it suitable for monitoring glucose levels in biomedical applications. The green synthesis route not only contributes to sustainability but also provides a cost-effective and scalable method for fabricating high-performance biosensors, offering significant potential for noninvasive glucose monitoring in diabetic care.","PeriodicalId":13014,"journal":{"name":"IEEE Sensors Letters","volume":"10 1","pages":"1-4"},"PeriodicalIF":2.2,"publicationDate":"2025-12-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145830943","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This work presents a novel algorithm for failure prediction in manufacturing processes using online unsupervised learning based on Kullback–Leibler divergence (KLD). The proposed method continuously monitors sensor data by comparing the probability distributions of a test window against those of a reference window to detect deviations that signal potential system degradation. These distributions are modeled as multivariate Gaussians to capture interdependencies between sensor signals. The algorithm is applied to real-world data from an electric arc furnace in the steel industry, demonstrating its ability to predict failures without prior offline training. Experimental results reveal that multivariate KLD analysis offers a more favorable balance between early fault detection and false alarm rates than univariate approaches. The method provides a lightweight, data-efficient, and practical solution for predictive maintenance in industrial settings where labeled failure data is limited or unavailable.
{"title":"Failure Prediction in Manufacturing Processes Via Kullback–Leibler Divergence","authors":"Gianluca Tabella;Mohammed Ayalew Belay;Ismael Viejo;María Herrando;Pierluigi Salvo Rossi","doi":"10.1109/LSENS.2025.3641051","DOIUrl":"https://doi.org/10.1109/LSENS.2025.3641051","url":null,"abstract":"This work presents a novel algorithm for failure prediction in manufacturing processes using online unsupervised learning based on Kullback–Leibler divergence (KLD). The proposed method continuously monitors sensor data by comparing the probability distributions of a test window against those of a reference window to detect deviations that signal potential system degradation. These distributions are modeled as multivariate Gaussians to capture interdependencies between sensor signals. The algorithm is applied to real-world data from an electric arc furnace in the steel industry, demonstrating its ability to predict failures without prior offline training. Experimental results reveal that multivariate KLD analysis offers a more favorable balance between early fault detection and false alarm rates than univariate approaches. The method provides a lightweight, data-efficient, and practical solution for predictive maintenance in industrial settings where labeled failure data is limited or unavailable.","PeriodicalId":13014,"journal":{"name":"IEEE Sensors Letters","volume":"10 1","pages":"1-4"},"PeriodicalIF":2.2,"publicationDate":"2025-12-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145830807","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Flexible strain sensors with high sensitivity and a wide detection range are essential for next-generation healthcare and soft robotics. However, achieving a tradeoff between sensitivity and detection range is challenging in a strain sensor. This letter reports an interface engineering approach that leverages a dual conductive network to develop high-fidelity strain sensors. 3-Aminopropyltriethoxysilane, which features amine and siloxane functional groups, simultaneously binds to hybrid filler and Ecoflex. Therefore, the composite shows enhanced filler dispersion, stress transfer, and electrical signal stability under strain. This interfacial interaction enables the achievement of a sensitivity of ∼188.7 with three-zone linearity. The synergy of a dual-network conductive pathway and interfacial adhesion facilitates more repetitive cycles, resulting in a ∼75% reduction of hysteresis and a response time of approximately 450 ms. Furthermore, the high stability of >1000 cycles is attributed to the prevention of filler pullout and the maintenance of a conductive network during continuous testing. This strategy provides a scalable approach for designing next-generation flexible sensors with molecular-level interface engineering, enabling superior sensitivity, mechanical reliability, and real-time health monitoring.
{"title":"Interface-Engineered Hybrid Networks to Resolve the Trade-Off Between Sensitivity and Detection Range in Flexible Strain Sensors","authors":"Animesh Maji;Chinmoy Kuila;Naresh Chandra Murmu;Tapas Kuila","doi":"10.1109/LSENS.2025.3639618","DOIUrl":"https://doi.org/10.1109/LSENS.2025.3639618","url":null,"abstract":"Flexible strain sensors with high sensitivity and a wide detection range are essential for next-generation healthcare and soft robotics. However, achieving a tradeoff between sensitivity and detection range is challenging in a strain sensor. This letter reports an interface engineering approach that leverages a dual conductive network to develop high-fidelity strain sensors. 3-Aminopropyltriethoxysilane, which features amine and siloxane functional groups, simultaneously binds to hybrid filler and Ecoflex. Therefore, the composite shows enhanced filler dispersion, stress transfer, and electrical signal stability under strain. This interfacial interaction enables the achievement of a sensitivity of ∼188.7 with three-zone linearity. The synergy of a dual-network conductive pathway and interfacial adhesion facilitates more repetitive cycles, resulting in a ∼75% reduction of hysteresis and a response time of approximately 450 ms. Furthermore, the high stability of >1000 cycles is attributed to the prevention of filler pullout and the maintenance of a conductive network during continuous testing. This strategy provides a scalable approach for designing next-generation flexible sensors with molecular-level interface engineering, enabling superior sensitivity, mechanical reliability, and real-time health monitoring.","PeriodicalId":13014,"journal":{"name":"IEEE Sensors Letters","volume":"10 1","pages":"1-4"},"PeriodicalIF":2.2,"publicationDate":"2025-12-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145778258","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Accelerating the design and adoption of compact devices involving 3-D current-carrying architectures requires new and enhanced inspection methodologies to support critical device development and failure analysis. Vector magnetic field imaging with high spatio-temporal resolution is a promising approach for probing these architectures by revealing 3-D current paths. These 3-D current density maps can be obtained from magnetic field maps by solving the difficult 3-D current reconstruction problem. We present a novel incoherent convolutional dictionary learning (ICDL)-based method to process magnetic field maps acquired via nitrogen-vacancy center–based wide-field magnetic microscopy. The ICDL-based approach separates the composite magnetic field into layer-specific components within the 3-D stacked structure. Subsequently, a plug-and-play-based iterative approach jointly deconvolves each layer's magnetic field to estimate the underlying current sources. The results demonstrate an average improvement of $approx$ 2.7 dB in peak signal-to-noise ratio and $approx$ 3.7% in structural similarity index over conventional convolutional dictionary learning-based methods.
{"title":"Incoherent Convolutional Dictionary Learning-Based 3-D Current Reconstruction From Magnetic Field Imaging","authors":"Saurabh Sahu;Prabhat Anand;Anuj Bathla;Kasturi Saha;M Girish Chandra","doi":"10.1109/LSENS.2025.3639475","DOIUrl":"https://doi.org/10.1109/LSENS.2025.3639475","url":null,"abstract":"Accelerating the design and adoption of compact devices involving 3-D current-carrying architectures requires new and enhanced inspection methodologies to support critical device development and failure analysis. Vector magnetic field imaging with high spatio-temporal resolution is a promising approach for probing these architectures by revealing 3-D current paths. These 3-D current density maps can be obtained from magnetic field maps by solving the difficult 3-D current reconstruction problem. We present a novel incoherent convolutional dictionary learning (ICDL)-based method to process magnetic field maps acquired via nitrogen-vacancy center–based wide-field magnetic microscopy. The ICDL-based approach separates the composite magnetic field into layer-specific components within the 3-D stacked structure. Subsequently, a plug-and-play-based iterative approach jointly deconvolves each layer's magnetic field to estimate the underlying current sources. The results demonstrate an average improvement of <inline-formula><tex-math>$approx$</tex-math></inline-formula> 2.7 dB in peak signal-to-noise ratio and <inline-formula><tex-math>$approx$</tex-math></inline-formula> 3.7% in structural similarity index over conventional convolutional dictionary learning-based methods.","PeriodicalId":13014,"journal":{"name":"IEEE Sensors Letters","volume":"10 1","pages":"1-4"},"PeriodicalIF":2.2,"publicationDate":"2025-12-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145886610","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This letter presents a comprehensive study of the source of nonlinearities in a novel z-axis microelectromechanical systems (MEMS) accelerometer fabricated using a two-silicon-layer fabrication process. The device features a unique mechanical architecture that converts the out-of-plane motion of the proof mass into linear in-plane displacement of the sensing frames, enabling efficient capacitive readout. Initial experimental characterization revealed an unexpected nonlinearity, exceeding predictions of the ideal mechanical model. To investigate the origin of this behavior, a detailed 3-D finite element method (FEM) analysis was performed, incorporating fabrication-induced effects such as substrate deformation and residual stresses. Simulations demonstrated that substrate deformation has negligible impact within the operational range, while residual prestresses on the structural silicon layer strongly influence the device response, producing nonlinearity levels consistent with experimental measurements. The close agreement between FEM predictions and experimental data validates the model and identifies residual prestresses on the structural silicon layer as the dominant factor affecting the device linearity. These insights provide a clear pathway for future design optimization, suggesting that careful control of residual stress and potential structural modifications can significantly improve the performance, linearity, and reliability of subsequent generations of z-axis MEMS accelerometers.
{"title":"Understanding the Nonlinear Behavior of a New z-Axis MEMS Accelerometer With In-Plane Readout","authors":"Yassine Banani;Christian Padovani;Giacomo Langfelder;Gabriele Gattere;Valentina Zega","doi":"10.1109/LSENS.2025.3638964","DOIUrl":"https://doi.org/10.1109/LSENS.2025.3638964","url":null,"abstract":"This letter presents a comprehensive study of the source of nonlinearities in a novel <italic>z</i>-axis microelectromechanical systems (MEMS) accelerometer fabricated using a two-silicon-layer fabrication process. The device features a unique mechanical architecture that converts the out-of-plane motion of the proof mass into linear in-plane displacement of the sensing frames, enabling efficient capacitive readout. Initial experimental characterization revealed an unexpected nonlinearity, exceeding predictions of the ideal mechanical model. To investigate the origin of this behavior, a detailed 3-D finite element method (FEM) analysis was performed, incorporating fabrication-induced effects such as substrate deformation and residual stresses. Simulations demonstrated that substrate deformation has negligible impact within the operational range, while residual prestresses on the structural silicon layer strongly influence the device response, producing nonlinearity levels consistent with experimental measurements. The close agreement between FEM predictions and experimental data validates the model and identifies residual prestresses on the structural silicon layer as the dominant factor affecting the device linearity. These insights provide a clear pathway for future design optimization, suggesting that careful control of residual stress and potential structural modifications can significantly improve the performance, linearity, and reliability of subsequent generations of <italic>z</i>-axis MEMS accelerometers.","PeriodicalId":13014,"journal":{"name":"IEEE Sensors Letters","volume":"10 1","pages":"1-4"},"PeriodicalIF":2.2,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11271566","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145778318","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Radar-based respiratory measurement is a promising tool for the noncontact detection of sleep apnea. Our team has reported that apnea events can be accurately detected using the statistical characteristics of the amplitude of respiratory displacement. However, apnea and hypopnea events are often followed by irregular breathing, reducing the detection accuracy. This study proposes a new method to overcome this performance degradation by repeatedly applying the detection method to radar data sets corresponding to multiple overlapping time intervals. Averaging the detected classes over multiple time intervals gives an analog value between 0 and 1, which can be interpreted as the probability of apnea and hypopnea events occurring. We show that the proposed method can mitigate the effect of irregular breathing that occurs after apnea and hypopnea events. The performance was validated using overnight recordings from seven patients, showing a 1.4-fold improvement in apnea and hypopnea event detection compared with the nonoverlapping method.
{"title":"Accurate Radar-Based Detection of Sleep Apnea Using Overlapping Time-Interval Averaging","authors":"Kodai Hasegawa;Shigeaki Okumura;Hirofumi Taki;Hironobu Sunadome;Satoshi Hamada;Susumu Sato;Kazuo Chin;Takuya Sakamoto","doi":"10.1109/LSENS.2025.3639141","DOIUrl":"https://doi.org/10.1109/LSENS.2025.3639141","url":null,"abstract":"Radar-based respiratory measurement is a promising tool for the noncontact detection of sleep apnea. Our team has reported that apnea events can be accurately detected using the statistical characteristics of the amplitude of respiratory displacement. However, apnea and hypopnea events are often followed by irregular breathing, reducing the detection accuracy. This study proposes a new method to overcome this performance degradation by repeatedly applying the detection method to radar data sets corresponding to multiple overlapping time intervals. Averaging the detected classes over multiple time intervals gives an analog value between 0 and 1, which can be interpreted as the probability of apnea and hypopnea events occurring. We show that the proposed method can mitigate the effect of irregular breathing that occurs after apnea and hypopnea events. The performance was validated using overnight recordings from seven patients, showing a 1.4-fold improvement in apnea and hypopnea event detection compared with the nonoverlapping method.","PeriodicalId":13014,"journal":{"name":"IEEE Sensors Letters","volume":"10 1","pages":"1-4"},"PeriodicalIF":2.2,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11271637","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145886660","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","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：