Marine gearboxes operating long-term in high-temperature, high-humidity, and high-salinity mist marine environments are highly susceptible to corrosion faults, posing significant threats to the reliability and safety of shipboard equipment. Ultrasonic testing (UT) technology, with its noncontact and remote capabilities, is well-suited for inspecting complex workpieces in such adverse conditions. To address the limitations of existing quantitative corrosion detection methods for gearboxes, while fully leveraging the nondestructiveness and information integrity advantages of ultrasonic nondestructive testing technology, this article proposes the wavelet time–frequency attention fusion network (WAFN), an ultrasonic signal-based method for gearbox corrosion detection. The method first constructs an optimized four-channel parallel ConvNeXt network (multichannel time–frequency feature extraction (MCFE) module) for deep feature extraction. Subsequently, a Transformer encoder module is introduced to fuse global features and capture cross-channel spatial dependencies. Then, a symmetric multichannel cross-attention feature fusion (CAFF) module realizes adaptive weighted fusion of local and global features. Finally, a supervised collaborative contrast loss (SCCL) training mechanism is designed, combining feature loss and classification loss to pull features of the same corrosion level closer while pushing features of different levels apart. This effectively mitigates interference from intraclass variations and blurred interclass feature boundaries inherent in quantitative data, achieving quantitative nondestructive detection of gearbox corrosion. Experimental results show that the proposed model achieves higher comprehensive accuracy on the two datasets and in the supplementary experiments with actual plates, verifying the effectiveness of this method.
{"title":"Corrosion Detection of Gearbox Based on Wavelet Time–Frequency Attention Fusion Network","authors":"Tianyu Gao;Yongjiang Li;Jingli Yang;Yongqi Chang;Xiaopeng Fan;Meiyan Zhang","doi":"10.1109/TIM.2025.3644553","DOIUrl":"https://doi.org/10.1109/TIM.2025.3644553","url":null,"abstract":"Marine gearboxes operating long-term in high-temperature, high-humidity, and high-salinity mist marine environments are highly susceptible to corrosion faults, posing significant threats to the reliability and safety of shipboard equipment. Ultrasonic testing (UT) technology, with its noncontact and remote capabilities, is well-suited for inspecting complex workpieces in such adverse conditions. To address the limitations of existing quantitative corrosion detection methods for gearboxes, while fully leveraging the nondestructiveness and information integrity advantages of ultrasonic nondestructive testing technology, this article proposes the wavelet time–frequency attention fusion network (WAFN), an ultrasonic signal-based method for gearbox corrosion detection. The method first constructs an optimized four-channel parallel ConvNeXt network (multichannel time–frequency feature extraction (MCFE) module) for deep feature extraction. Subsequently, a Transformer encoder module is introduced to fuse global features and capture cross-channel spatial dependencies. Then, a symmetric multichannel cross-attention feature fusion (CAFF) module realizes adaptive weighted fusion of local and global features. Finally, a supervised collaborative contrast loss (SCCL) training mechanism is designed, combining feature loss and classification loss to pull features of the same corrosion level closer while pushing features of different levels apart. This effectively mitigates interference from intraclass variations and blurred interclass feature boundaries inherent in quantitative data, achieving quantitative nondestructive detection of gearbox corrosion. Experimental results show that the proposed model achieves higher comprehensive accuracy on the two datasets and in the supplementary experiments with actual plates, verifying the effectiveness of this method.","PeriodicalId":13341,"journal":{"name":"IEEE Transactions on Instrumentation and Measurement","volume":"75 ","pages":"1-13"},"PeriodicalIF":5.9,"publicationDate":"2025-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145904313","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-29DOI: 10.1109/TIM.2025.3648107
Hyo-Jeong Kim;Jae-Hyun Park;Yeong-Jin Choe;Kyung-Tae Kim
Indoor human localization (IHL) using radar is an essential technology for various applications. However, accurately identifying human presence indoors is complicated by strong reflections from static clutter and multipath propagation, which can produce ghost targets and reduce localization precision. To overcome these limitations, we introduce a novel framework that utilizes periodic micro-movements, such as those from human respiration, using multiple-input–multiple-output frequency-modulated continuous-wave (MIMO FMCW) radar system to distinguish stationary individuals from surrounding clutter. The proposed method analyzes Doppler signals in the 2-D range-angle domain and employs a curvature-based feature to capture directional variations of Doppler trajectories over slow time. Without requiring prior knowledge of the environment or training data, this approach effectively highlights human-induced motion patterns. Experimental results using a commercial MIMO FMCW radar system demonstrate that the proposed method effectively suppresses ghost targets caused by multipath propagation and achieves accurate localization of stationary human targets in indoor environments.
{"title":"Curvature Variance Method for Indoor Human Localization Using MIMO FMCW Radar","authors":"Hyo-Jeong Kim;Jae-Hyun Park;Yeong-Jin Choe;Kyung-Tae Kim","doi":"10.1109/TIM.2025.3648107","DOIUrl":"https://doi.org/10.1109/TIM.2025.3648107","url":null,"abstract":"Indoor human localization (IHL) using radar is an essential technology for various applications. However, accurately identifying human presence indoors is complicated by strong reflections from static clutter and multipath propagation, which can produce ghost targets and reduce localization precision. To overcome these limitations, we introduce a novel framework that utilizes periodic micro-movements, such as those from human respiration, using multiple-input–multiple-output frequency-modulated continuous-wave (MIMO FMCW) radar system to distinguish stationary individuals from surrounding clutter. The proposed method analyzes Doppler signals in the 2-D range-angle domain and employs a curvature-based feature to capture directional variations of Doppler trajectories over slow time. Without requiring prior knowledge of the environment or training data, this approach effectively highlights human-induced motion patterns. Experimental results using a commercial MIMO FMCW radar system demonstrate that the proposed method effectively suppresses ghost targets caused by multipath propagation and achieves accurate localization of stationary human targets in indoor environments.","PeriodicalId":13341,"journal":{"name":"IEEE Transactions on Instrumentation and Measurement","volume":"75 ","pages":"1-10"},"PeriodicalIF":5.9,"publicationDate":"2025-12-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145904268","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-29DOI: 10.1109/TIM.2025.3647997
Jiaxin Yang;Kun Feng;Yuan Liu;Yongjia Peng
Accurate extraction of second-order cyclostationary (CS2) components in bearing fault diagnosis requires effective suppression of first-order cyclostationary (CS1) interference. Under strong cyclostationary disturbances and non-Gaussian noise, conventional filtering methods often lack robustness to impulsive noise and fail to sufficiently suppress CS1 components. To address this, a generalized correntropy-driven adaptive joint filtering framework is proposed, integrating adaptive filtering and blind deconvolution. In the first stage, an adaptive filter based on generalized correntropy and an energy-guided dynamic step-size mechanism is used to remove CS1 and enhance noise robustness. A multigroup candidate fault frequency (MGCFF) extraction strategy then identifies dominant fault frequencies, which are passed to the second-stage generalized Gaussian cyclostationary blind deconvolution (CYCBD$beta $ ) filter. The filter length is adaptively optimized using a sparsity-based performance-efficiency ratio (PER) to amplify CS2 features. The proposed method enables adaptive parameter tuning and coordinated filtering, ensuring accurate extraction of weak fault features under complex, noisy conditions. Simulation, experimental, and engineering validations confirm its superiority over conventional approaches in feature enhancement and robustness.
{"title":"An Improved Generalized Correntropy-Driven Adaptive Cyclostationary Blind Deconvolution Method for Bearing Fault Diagnosis Under Strong Interference","authors":"Jiaxin Yang;Kun Feng;Yuan Liu;Yongjia Peng","doi":"10.1109/TIM.2025.3647997","DOIUrl":"https://doi.org/10.1109/TIM.2025.3647997","url":null,"abstract":"Accurate extraction of second-order cyclostationary (CS2) components in bearing fault diagnosis requires effective suppression of first-order cyclostationary (CS1) interference. Under strong cyclostationary disturbances and non-Gaussian noise, conventional filtering methods often lack robustness to impulsive noise and fail to sufficiently suppress CS1 components. To address this, a generalized correntropy-driven adaptive joint filtering framework is proposed, integrating adaptive filtering and blind deconvolution. In the first stage, an adaptive filter based on generalized correntropy and an energy-guided dynamic step-size mechanism is used to remove CS1 and enhance noise robustness. A multigroup candidate fault frequency (MGCFF) extraction strategy then identifies dominant fault frequencies, which are passed to the second-stage generalized Gaussian cyclostationary blind deconvolution (CYCBD<inline-formula> <tex-math>$beta $ </tex-math></inline-formula>) filter. The filter length is adaptively optimized using a sparsity-based performance-efficiency ratio (PER) to amplify CS2 features. The proposed method enables adaptive parameter tuning and coordinated filtering, ensuring accurate extraction of weak fault features under complex, noisy conditions. Simulation, experimental, and engineering validations confirm its superiority over conventional approaches in feature enhancement and robustness.","PeriodicalId":13341,"journal":{"name":"IEEE Transactions on Instrumentation and Measurement","volume":"75 ","pages":"1-15"},"PeriodicalIF":5.9,"publicationDate":"2025-12-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145904293","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In this study, the design and optimization of a dipole electromagnet pole tip are presented, with the aim of compacting the structure and enhancing the magnetic field quality across the Tanabe optimal width. The optimization process was conducted by introducing adjustable control points at the pole edges of this beam-bending precision instrument, using the genetic algorithm (GA) and particle swarm optimization (PSO) techniques, coupled with a 2-D finite element method magnetics (FEMM) simulation code. The simulation results demonstrated a 15.9% improvement in the horizontal good field region (GFR) compared to the existing dipole in the ES150 ion accelerator’s mass spectrometer, achieving the field uniformity $(1times 10^{-3})$ . Additionally, the proposed approach led to a 12.3% reduction in the dipole electromagnet’s overall dimensions, while also reducing magnetic saturation at the edges by 60% and enhancing manufacturability. Subsequent mass spectrometry of hydrogen ion beams confirmed the presence of $text {H}^{+},text {H}_{2}^{+},~text {and} ~text {H}_{3}^{+}$ ions with relative abundances of 72.2%, 8.9%, and 18.9%, respectively. The results validate the accuracy, effectiveness, and potential of the proposed design for future development in advanced diagnostic instrumentation.
{"title":"Development of a Miniaturized Dipole Electromagnet for GFR Enhancement in an Accelerator Mass Spectrometer","authors":"Mohsen Dehghan;Fereydoun Abbasi Davani;Shahin Sanaye Hajari;Reza Ghaderi;Farshad Ghasemi","doi":"10.1109/TIM.2025.3643079","DOIUrl":"https://doi.org/10.1109/TIM.2025.3643079","url":null,"abstract":"In this study, the design and optimization of a dipole electromagnet pole tip are presented, with the aim of compacting the structure and enhancing the magnetic field quality across the Tanabe optimal width. The optimization process was conducted by introducing adjustable control points at the pole edges of this beam-bending precision instrument, using the genetic algorithm (GA) and particle swarm optimization (PSO) techniques, coupled with a 2-D finite element method magnetics (FEMM) simulation code. The simulation results demonstrated a 15.9% improvement in the horizontal good field region (GFR) compared to the existing dipole in the ES150 ion accelerator’s mass spectrometer, achieving the field uniformity <inline-formula> <tex-math>$(1times 10^{-3})$ </tex-math></inline-formula>. Additionally, the proposed approach led to a 12.3% reduction in the dipole electromagnet’s overall dimensions, while also reducing magnetic saturation at the edges by 60% and enhancing manufacturability. Subsequent mass spectrometry of hydrogen ion beams confirmed the presence of <inline-formula> <tex-math>$text {H}^{+},text {H}_{2}^{+},~text {and} ~text {H}_{3}^{+}$ </tex-math></inline-formula> ions with relative abundances of 72.2%, 8.9%, and 18.9%, respectively. The results validate the accuracy, effectiveness, and potential of the proposed design for future development in advanced diagnostic instrumentation.","PeriodicalId":13341,"journal":{"name":"IEEE Transactions on Instrumentation and Measurement","volume":"75 ","pages":"1-8"},"PeriodicalIF":5.9,"publicationDate":"2025-12-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145904262","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Active noise control (ANC) systems in vehicle cabins are conventionally validated through real-time prototypes in road tests. However, the escalating complexity of ANC algorithms and the inherent variability of real-world testing conditions frequently lead to high experimental costs and inefficiencies, significantly impeding the broader implementation of in-vehicle ANC systems. This challenge underscores a significant deficiency in the field: the absence of a high-fidelity approach to facilitate the validation of algorithms under reproducible conditions. This article introduces a sound field reproduction (SFR) method based on the multichannel pressure matching least squares (MCPMLSs) algorithm for in-vehicle ANC system evaluation. The SFR-ANC system is further proposed, integrating a feedback-equalization-optimized SFR subsystem with dual parallel ANC subsystems powered by broadband and narrowband adaptive algorithms. The hybrid system was implemented through a loudspeaker array and active headrest integration, designed to achieve accurate in-vehicle sound environment reproduction and localized cancellation. The system’s performance was evaluated using ear zone noise collected inside a real vehicle as disturbance signals within an acoustic laboratory. Experimental results validate the system’s accuracy in reproducing authentic noise environments and the noise canceling performances both in stable and transient conditions. The proposed approach establishes a reproducible testing protocol for standardized subjective-objective assessment of in-vehicle ANC performance.
{"title":"A Study on Active Noise Control in Reproduced In-Vehicle Sound Environment","authors":"Rubin Li;Xu Zheng;Bo Wan;Yong Yu;Xuexian Liu;Chi Liu;Yi Qiu","doi":"10.1109/TIM.2025.3648526","DOIUrl":"https://doi.org/10.1109/TIM.2025.3648526","url":null,"abstract":"Active noise control (ANC) systems in vehicle cabins are conventionally validated through real-time prototypes in road tests. However, the escalating complexity of ANC algorithms and the inherent variability of real-world testing conditions frequently lead to high experimental costs and inefficiencies, significantly impeding the broader implementation of in-vehicle ANC systems. This challenge underscores a significant deficiency in the field: the absence of a high-fidelity approach to facilitate the validation of algorithms under reproducible conditions. This article introduces a sound field reproduction (SFR) method based on the multichannel pressure matching least squares (MCPMLSs) algorithm for in-vehicle ANC system evaluation. The SFR-ANC system is further proposed, integrating a feedback-equalization-optimized SFR subsystem with dual parallel ANC subsystems powered by broadband and narrowband adaptive algorithms. The hybrid system was implemented through a loudspeaker array and active headrest integration, designed to achieve accurate in-vehicle sound environment reproduction and localized cancellation. The system’s performance was evaluated using ear zone noise collected inside a real vehicle as disturbance signals within an acoustic laboratory. Experimental results validate the system’s accuracy in reproducing authentic noise environments and the noise canceling performances both in stable and transient conditions. The proposed approach establishes a reproducible testing protocol for standardized subjective-objective assessment of in-vehicle ANC performance.","PeriodicalId":13341,"journal":{"name":"IEEE Transactions on Instrumentation and Measurement","volume":"75 ","pages":"1-16"},"PeriodicalIF":5.9,"publicationDate":"2025-12-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145982371","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-25DOI: 10.1109/TIM.2025.3648093
Chihoon Kim;Myungjoo Kang;Munseob Lee
Plenoptic systems represent a significant advancement in imaging technology, enabling sophisticated 3-D image capture with a single exposure. This study presents the development of a microscope system for microscale applications, based on plenoptic principles. The proposed system incorporates a microlens array (MLA) designed and fabricated with minimal errors through optical simulation and a custom-designed lens jig. Precise optical alignment was performed to optimize the system performance, ensuring improved spatial resolution and depth of field (DOF) by matching the numerical apertures (NAs) of the MLA and the tube lens. The system achieved a spatial resolution of $12.4~mu $ m, with contrast ratios of 90% and 97% in the horizontal and vertical directions, respectively. Additionally, the DOF was enhanced by $1.6times $ , increasing from the theoretical value of 630–$1000~mu $ m. The analysis revealed that high-frequency components were more sensitive to variations in the DOF relative to the spatial frequency, whereas low-frequency components maintained high clarity over an extended range.
{"title":"Advanced Optical Design and Evaluation of a Plenoptic Microscope for Extended 3-D Imaging Capabilities","authors":"Chihoon Kim;Myungjoo Kang;Munseob Lee","doi":"10.1109/TIM.2025.3648093","DOIUrl":"https://doi.org/10.1109/TIM.2025.3648093","url":null,"abstract":"Plenoptic systems represent a significant advancement in imaging technology, enabling sophisticated 3-D image capture with a single exposure. This study presents the development of a microscope system for microscale applications, based on plenoptic principles. The proposed system incorporates a microlens array (MLA) designed and fabricated with minimal errors through optical simulation and a custom-designed lens jig. Precise optical alignment was performed to optimize the system performance, ensuring improved spatial resolution and depth of field (DOF) by matching the numerical apertures (NAs) of the MLA and the tube lens. The system achieved a spatial resolution of <inline-formula> <tex-math>$12.4~mu $ </tex-math></inline-formula>m, with contrast ratios of 90% and 97% in the horizontal and vertical directions, respectively. Additionally, the DOF was enhanced by <inline-formula> <tex-math>$1.6times $ </tex-math></inline-formula>, increasing from the theoretical value of 630–<inline-formula> <tex-math>$1000~mu $ </tex-math></inline-formula>m. The analysis revealed that high-frequency components were more sensitive to variations in the DOF relative to the spatial frequency, whereas low-frequency components maintained high clarity over an extended range.","PeriodicalId":13341,"journal":{"name":"IEEE Transactions on Instrumentation and Measurement","volume":"75 ","pages":"1-10"},"PeriodicalIF":5.9,"publicationDate":"2025-12-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145904332","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-25DOI: 10.1109/TIM.2025.3647988
Yanju Ji;Weiyi Wang;Shilin Qiu;Yibing Yu;Huaishi Liu
The superposition of the primary field in time-domain electromagnetic (TDEM) surveys can severely degrade data quality, especially during the early stage after the transmitting current is turned off, resulting in shallow detection blind spots. To address this problem, this article proposes a cm-level coil-based method for equivalent measurement and removal of the primary field. By driving the cm-level coil with a proportionally reduced current, an equivalent primary field is generated while the induced secondary field remains negligible. The receiving system directly measures this equivalent primary field, enabling its subtraction from the observational data. The method was validated through field experiments using $4times 4$ m and $50times 50$ m transmitting coils, where residual primary fields from the bucking device and the complete primary field of the large loop were effectively removed. Consequently, the effective observation window was extended by 150 and $278~mu $ s, respectively. The method significantly enhances shallow detection capability and exhibits broad applicability across TDEM system scales.
{"title":"Equivalent Measurement and Removal of Primary Field Effect Using a cm-Level Coil in Time-Domain Electromagnetic Surveys","authors":"Yanju Ji;Weiyi Wang;Shilin Qiu;Yibing Yu;Huaishi Liu","doi":"10.1109/TIM.2025.3647988","DOIUrl":"https://doi.org/10.1109/TIM.2025.3647988","url":null,"abstract":"The superposition of the primary field in time-domain electromagnetic (TDEM) surveys can severely degrade data quality, especially during the early stage after the transmitting current is turned off, resulting in shallow detection blind spots. To address this problem, this article proposes a cm-level coil-based method for equivalent measurement and removal of the primary field. By driving the cm-level coil with a proportionally reduced current, an equivalent primary field is generated while the induced secondary field remains negligible. The receiving system directly measures this equivalent primary field, enabling its subtraction from the observational data. The method was validated through field experiments using <inline-formula> <tex-math>$4times 4$ </tex-math></inline-formula> m and <inline-formula> <tex-math>$50times 50$ </tex-math></inline-formula> m transmitting coils, where residual primary fields from the bucking device and the complete primary field of the large loop were effectively removed. Consequently, the effective observation window was extended by 150 and <inline-formula> <tex-math>$278~mu $ </tex-math></inline-formula>s, respectively. The method significantly enhances shallow detection capability and exhibits broad applicability across TDEM system scales.","PeriodicalId":13341,"journal":{"name":"IEEE Transactions on Instrumentation and Measurement","volume":"75 ","pages":"1-10"},"PeriodicalIF":5.9,"publicationDate":"2025-12-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145904302","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
To address the challenge of small and sparse datasets, this study presents a first attempt to integrate slag mechanism knowledge with Shapley additive explanations (SHAP)-guided stacking and data generation for accurately predicting endpoint quality in the converter steelmaking process. The slag mechanism knowledge is adopted to increase key characteristic variables and the interpretability of prediction results. The missing rate and abnormal values in the smelting data are hierarchically identified and processed by combining the interquartile range (IQR) method and smelting experience. The conditional tabular generative adversarial network (CTGAN) is used to expand the training dataset for alleviating the overfitting problem while avoiding data leaking and distribution shifts. The cumulative contribution rate of SHAP is used to screen out KNN, TabTransformer, RF, TabNet, and ET as the base models (Layer-0). The Ridge is used as a meta-learner (Layer-1) to alleviate the multicollinearity through L2 regularization. By comparing the ablation experiments and existing networks, the superiority of the proposed model is verified by the production data from a true steel plant. The experimental results illustrate that the prediction $R^{2}$ of the proposed model in terms of phosphorus (P) content and endpoint temperature are 0.932 and 0.948, respectively, and other evaluation indicators are also significantly better than the comparison models. The proposed modeling technology lays a foundation for optimization of the converter steelmaking process.
{"title":"Stacking Network Fusing Slag Knowledge and Meta-Learner Decision for Predicting Quality of Converter Steelmaking Process With Sparse Samples","authors":"Shijian Dong;Tianyu Yu;Jiahao Liu;Zhaojie Wang;Xiaoqing Jiang","doi":"10.1109/TIM.2025.3648070","DOIUrl":"https://doi.org/10.1109/TIM.2025.3648070","url":null,"abstract":"To address the challenge of small and sparse datasets, this study presents a first attempt to integrate slag mechanism knowledge with Shapley additive explanations (SHAP)-guided stacking and data generation for accurately predicting endpoint quality in the converter steelmaking process. The slag mechanism knowledge is adopted to increase key characteristic variables and the interpretability of prediction results. The missing rate and abnormal values in the smelting data are hierarchically identified and processed by combining the interquartile range (IQR) method and smelting experience. The conditional tabular generative adversarial network (CTGAN) is used to expand the training dataset for alleviating the overfitting problem while avoiding data leaking and distribution shifts. The cumulative contribution rate of SHAP is used to screen out KNN, TabTransformer, RF, TabNet, and ET as the base models (Layer-0). The Ridge is used as a meta-learner (Layer-1) to alleviate the multicollinearity through L2 regularization. By comparing the ablation experiments and existing networks, the superiority of the proposed model is verified by the production data from a true steel plant. The experimental results illustrate that the prediction <inline-formula> <tex-math>$R^{2}$ </tex-math></inline-formula> of the proposed model in terms of phosphorus (P) content and endpoint temperature are 0.932 and 0.948, respectively, and other evaluation indicators are also significantly better than the comparison models. The proposed modeling technology lays a foundation for optimization of the converter steelmaking process.","PeriodicalId":13341,"journal":{"name":"IEEE Transactions on Instrumentation and Measurement","volume":"75 ","pages":"1-16"},"PeriodicalIF":5.9,"publicationDate":"2025-12-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145904343","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Global navigation satellite system (GNSS) carrier phase measurement is highly vulnerable to signal attenuation, multipath, and blockage in urban environments, which significantly degrades the availability of precise GNSS positioning. Long coherent integration (LCI) serves as an effective approach to suppress thermal noise and mitigate multipath interferences within phase-locked loops (PLLs); however, its performance is constrained by the dynamic stress resulting from satellite–receiver motions. This study proposes a GNSS/inertial navigation system (INS)/odometer (ODO) deeply coupled (GIO-DC) system with LCI PLLs. An (ODO) distance increment measurement model is integrated with a MEMS IMU to estimate and compensate for the PLLs’ dynamic stress with enhanced accuracy and reliability, thereby enabling extended coherent integration time. In addition, a four-quadrant phase discriminator is adopted to expand the PLL pull-in range and reduce the likelihood of cycle slips. Field tests on a wheeled vehicle in typical urban complex environments were conducted to evaluate the performance of the GIO-DC system from multiple perspectives. The results confirmed the superiority of the proposed approaches. A coherent integration time of 800 ms was achieved, realizing continuous carrier phase measurement and robust centimeter-level positioning. The proposed deeply integrated system, built on the low-cost MEMS IMU and ODO, delivers performance on par with that of a system based on a navigation-grade IMU.
{"title":"MEMS IMU/ODO-Aided GNSS Long Coherent Integration PLL for Urban Vehicle Precise Positioning","authors":"Tisheng Zhang;Huilin Shi;Liqiang Wang;Xin Feng;Yuepei Shi;Xiaoji Niu","doi":"10.1109/TIM.2025.3648069","DOIUrl":"https://doi.org/10.1109/TIM.2025.3648069","url":null,"abstract":"Global navigation satellite system (GNSS) carrier phase measurement is highly vulnerable to signal attenuation, multipath, and blockage in urban environments, which significantly degrades the availability of precise GNSS positioning. Long coherent integration (LCI) serves as an effective approach to suppress thermal noise and mitigate multipath interferences within phase-locked loops (PLLs); however, its performance is constrained by the dynamic stress resulting from satellite–receiver motions. This study proposes a GNSS/inertial navigation system (INS)/odometer (ODO) deeply coupled (GIO-DC) system with LCI PLLs. An (ODO) distance increment measurement model is integrated with a MEMS IMU to estimate and compensate for the PLLs’ dynamic stress with enhanced accuracy and reliability, thereby enabling extended coherent integration time. In addition, a four-quadrant phase discriminator is adopted to expand the PLL pull-in range and reduce the likelihood of cycle slips. Field tests on a wheeled vehicle in typical urban complex environments were conducted to evaluate the performance of the GIO-DC system from multiple perspectives. The results confirmed the superiority of the proposed approaches. A coherent integration time of 800 ms was achieved, realizing continuous carrier phase measurement and robust centimeter-level positioning. The proposed deeply integrated system, built on the low-cost MEMS IMU and ODO, delivers performance on par with that of a system based on a navigation-grade IMU.","PeriodicalId":13341,"journal":{"name":"IEEE Transactions on Instrumentation and Measurement","volume":"75 ","pages":"1-14"},"PeriodicalIF":5.9,"publicationDate":"2025-12-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145904308","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-24DOI: 10.1109/TIM.2025.3647989
Lei Cheng;Lihao Guo;Tianya Zhang;Tam Bang;Austin Harris;Mustafa Hajij;Mina Sartipi;Siyang Cao
Accurate multisensor calibration is essential for deploying robust perception systems in applications such as autonomous driving and intelligent transportation. Existing light detection and ranging (LiDAR)–camera calibration methods often rely on manually placed targets, preliminary parameter estimates, or intensive data preprocessing, limiting their scalability and adaptability in real-world settings. In this work, we propose a fully automatic, targetless, and online calibration framework, CalibRefine, which directly processes raw LiDAR point clouds and camera images. Our approach is divided into four stages: 1) a common feature discriminator (CFD) that leverages relative spatial positions, visual appearance embeddings, and semantic class cues to identify and generate reliable LiDAR–camera correspondences; 2) a coarse homography-based calibration that uses the matched feature correspondences to estimate an initial transformation between the LiDAR and camera frames, serving as the foundation for further refinement; 3) an iterative refinement to incrementally improve alignment as additional data frames become available; and 4) an attention-based refinement that addresses nonplanar distortions by leveraging a vision transformer (ViT) and cross-attention mechanisms. Extensive experiments on two urban traffic datasets demonstrate that CalibRefine achieves high-precision calibration with minimal human input, outperforming state-of-the-art targetless methods and matching or surpassing manually tuned baselines. Our results show that robust object-level feature matching, combined with iterative refinement and self-supervised attention-based refinement, enables reliable sensor alignment in complex real-world conditions without ground-truth matrices or elaborate preprocessing. Code is available at https://github.com/radar-lab/Lidar_Camera_Automatic_Calibration
{"title":"CalibRefine: Deep Learning-Based Online Automatic Targetless LiDAR–Camera Calibration With Iterative and Attention-Driven Post-Refinement","authors":"Lei Cheng;Lihao Guo;Tianya Zhang;Tam Bang;Austin Harris;Mustafa Hajij;Mina Sartipi;Siyang Cao","doi":"10.1109/TIM.2025.3647989","DOIUrl":"https://doi.org/10.1109/TIM.2025.3647989","url":null,"abstract":"Accurate multisensor calibration is essential for deploying robust perception systems in applications such as autonomous driving and intelligent transportation. Existing light detection and ranging (LiDAR)–camera calibration methods often rely on manually placed targets, preliminary parameter estimates, or intensive data preprocessing, limiting their scalability and adaptability in real-world settings. In this work, we propose a fully automatic, targetless, and online calibration framework, <italic>CalibRefine</i>, which directly processes raw LiDAR point clouds and camera images. Our approach is divided into four stages: 1) a common feature discriminator (CFD) that leverages relative spatial positions, visual appearance embeddings, and semantic class cues to identify and generate reliable LiDAR–camera correspondences; 2) a coarse homography-based calibration that uses the matched feature correspondences to estimate an initial transformation between the LiDAR and camera frames, serving as the foundation for further refinement; 3) an iterative refinement to incrementally improve alignment as additional data frames become available; and 4) an attention-based refinement that addresses nonplanar distortions by leveraging a vision transformer (ViT) and cross-attention mechanisms. Extensive experiments on two urban traffic datasets demonstrate that CalibRefine achieves high-precision calibration with minimal human input, outperforming state-of-the-art targetless methods and matching or surpassing manually tuned baselines. Our results show that robust object-level feature matching, combined with iterative refinement and self-supervised attention-based refinement, enables reliable sensor alignment in complex real-world conditions without ground-truth matrices or elaborate preprocessing. Code is available at <uri>https://github.com/radar-lab/Lidar_Camera_Automatic_Calibration</uri>","PeriodicalId":13341,"journal":{"name":"IEEE Transactions on Instrumentation and Measurement","volume":"75 ","pages":"1-18"},"PeriodicalIF":5.9,"publicationDate":"2025-12-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145982170","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","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：