Surface defect detection is essential for ensuring the product quality of smartphone screen glass. In this work, a smartphone screen glass defect detection model based on an enhanced YOLOv7 framework with multiscale feature fusion and multiattention, named SSGDD-you only look once (YOLO) is proposed. In the developed SSGDD-YOLO model, the branch fusion block (BFB) is integrated low-level features from multiple scales through parallel processing, to enhance the details in lower level features for minimizing the information loss as less as possible. Furthermore, the SPPCSPC module of the head is improved as the SPPCSPC-I module, by replacing the standard max pooling with local importance-based pooling (LIP) that reflects the importance of features. The developed SPPCSPC-I module allows the network to automatically learn adaptive importance weights of features during downsampling, enhancing the multiscale feature extraction capability with diverse receptive fields. Finally, a contour-mixed attention block (C-MAB) is inserted into the feature fusion section of the network, which enhances spatial and channel information of features to reduce target information loss, improving the representation capability. Experiments are conducted using a challenging real-world defect image dataset gathered from a smartphone screen glass inspection line in an industrial plant. Results show the proposed SSGDD-YOLO model can achieve the highest mAP of 62.46% among all compared methods.
{"title":"SSGDD-YOLO: Multiscale Feature Fusion and Multiattention-Based YOLO for Smartphone Screen Glass Defect Detection","authors":"Ping Wu;Haote Zhou;Yicheng Yu;Zengdi Miao;Qianqian Pan;Xi Zhang;Jinfeng Gao","doi":"10.1109/JSEN.2024.3524584","DOIUrl":"https://doi.org/10.1109/JSEN.2024.3524584","url":null,"abstract":"Surface defect detection is essential for ensuring the product quality of smartphone screen glass. In this work, a smartphone screen glass defect detection model based on an enhanced YOLOv7 framework with multiscale feature fusion and multiattention, named SSGDD-you only look once (YOLO) is proposed. In the developed SSGDD-YOLO model, the branch fusion block (BFB) is integrated low-level features from multiple scales through parallel processing, to enhance the details in lower level features for minimizing the information loss as less as possible. Furthermore, the SPPCSPC module of the head is improved as the SPPCSPC-I module, by replacing the standard max pooling with local importance-based pooling (LIP) that reflects the importance of features. The developed SPPCSPC-I module allows the network to automatically learn adaptive importance weights of features during downsampling, enhancing the multiscale feature extraction capability with diverse receptive fields. Finally, a contour-mixed attention block (C-MAB) is inserted into the feature fusion section of the network, which enhances spatial and channel information of features to reduce target information loss, improving the representation capability. Experiments are conducted using a challenging real-world defect image dataset gathered from a smartphone screen glass inspection line in an industrial plant. Results show the proposed SSGDD-YOLO model can achieve the highest mAP of 62.46% among all compared methods.","PeriodicalId":447,"journal":{"name":"IEEE Sensors Journal","volume":"25 4","pages":"6982-6994"},"PeriodicalIF":4.3,"publicationDate":"2025-01-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143430475","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-09DOI: 10.1109/JPHOTOV.2024.3521121
Mauro Pravettoni;Min Hsian Saw;Muhammad Nabil Bin Abdul Aziz;Stephen En Rong Tay
In Part 1 of our article, we presented a method to quantify the incidence angle modifier (IAM) of photovoltaic (PV) devices, which differs from the methods proposed in IEC 61853-2 through the following: it utilizes a spot-area irradiation, delivered by an optical fiber system, a customized angle probe holder, and a current-to-voltage converter. Part 1 focused on single-cell devices and presented the validation of the new method on two different cell architectures. In Part 2, we generalize that method to commercial-size silicon PV modules, mirroring by the approach already used for module-level spectral responsivity measurements described in IEC 60904-8:2014. The proposed method is motivated by inclusion in the currently ongoing revision of IEC 61853-2, providing research centers and testing laboratories with an additional option to perform IAM measurements indoors. The reproducibility of the proposed method is addressed in this work via interlaboratory comparison with a different measurement method for the same quantity and with a detailed uncertainty analysis.
{"title":"A Spot-Area Method to Evaluate the Incidence Angle Modifier of Photovoltaic Devices-Part 2: Modules (Differential Method)","authors":"Mauro Pravettoni;Min Hsian Saw;Muhammad Nabil Bin Abdul Aziz;Stephen En Rong Tay","doi":"10.1109/JPHOTOV.2024.3521121","DOIUrl":"https://doi.org/10.1109/JPHOTOV.2024.3521121","url":null,"abstract":"In Part 1 of our article, we presented a method to quantify the incidence angle modifier (IAM) of photovoltaic (PV) devices, which differs from the methods proposed in IEC 61853-2 through the following: it utilizes a spot-area irradiation, delivered by an optical fiber system, a customized angle probe holder, and a current-to-voltage converter. Part 1 focused on single-cell devices and presented the validation of the new method on two different cell architectures. In Part 2, we generalize that method to commercial-size silicon PV modules, mirroring by the approach already used for module-level spectral responsivity measurements described in IEC 60904-8:2014. The proposed method is motivated by inclusion in the currently ongoing revision of IEC 61853-2, providing research centers and testing laboratories with an additional option to perform IAM measurements indoors. The reproducibility of the proposed method is addressed in this work via interlaboratory comparison with a different measurement method for the same quantity and with a detailed uncertainty analysis.","PeriodicalId":445,"journal":{"name":"IEEE Journal of Photovoltaics","volume":"15 2","pages":"280-289"},"PeriodicalIF":2.5,"publicationDate":"2025-01-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143455240","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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 : 2025-01-09DOI: 10.1109/JSEN.2025.3525622
Chunhui Li;Youfu Tang;Na Lei;Xu Wang
Addressing the limitations in feature extraction and model optimization complexity of convolutional neural network (CNN), an intelligent fault diagnosis method based on the Beluga whale optimization (BWO) algorithm optimized parallel CNN (PCNN) is proposed. First, the preprocessed vibration signal of the rolling bearing is converted into a 2-D time-frequency image by continuous wavelet transform (CWT). Second, the PCNN model is constructed, wherein the two branches independently learn distinct image weight values. This approach enhances deep-space feature expression by complementing high-dimensional features. Then, the BWO algorithm is used to optimize the hyperparameters of PCNN, thereby enhancing the model’s feature extraction and classification performance. Finally, multihead self-attention (MSA) is introduced into the PCNN framework to further improve the quality of feature representation and realize fault identification. The effectiveness and superiority of the method are verified by experimental datasets of rolling bearing and field test datasets of reciprocating compressor, the results of which show that the proposed model is significantly superior to the other models, exhibiting higher accuracy and better noise resistance, which can provide reliable technical support for practical industrial applications.
{"title":"An Intelligent Fault Diagnosis Method Based on Optimized Parallel Convolutional Neural Network","authors":"Chunhui Li;Youfu Tang;Na Lei;Xu Wang","doi":"10.1109/JSEN.2025.3525622","DOIUrl":"https://doi.org/10.1109/JSEN.2025.3525622","url":null,"abstract":"Addressing the limitations in feature extraction and model optimization complexity of convolutional neural network (CNN), an intelligent fault diagnosis method based on the Beluga whale optimization (BWO) algorithm optimized parallel CNN (PCNN) is proposed. First, the preprocessed vibration signal of the rolling bearing is converted into a 2-D time-frequency image by continuous wavelet transform (CWT). Second, the PCNN model is constructed, wherein the two branches independently learn distinct image weight values. This approach enhances deep-space feature expression by complementing high-dimensional features. Then, the BWO algorithm is used to optimize the hyperparameters of PCNN, thereby enhancing the model’s feature extraction and classification performance. Finally, multihead self-attention (MSA) is introduced into the PCNN framework to further improve the quality of feature representation and realize fault identification. The effectiveness and superiority of the method are verified by experimental datasets of rolling bearing and field test datasets of reciprocating compressor, the results of which show that the proposed model is significantly superior to the other models, exhibiting higher accuracy and better noise resistance, which can provide reliable technical support for practical industrial applications.","PeriodicalId":447,"journal":{"name":"IEEE Sensors Journal","volume":"25 4","pages":"6160-6175"},"PeriodicalIF":4.3,"publicationDate":"2025-01-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143430468","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}
With the development of wearable electronic devices and flexible biosensing technology, organic electrochemical transistors (OECTs) have received more and more attention. Fabric-based OECTs (F-OECTs) have a broad application prospect in wearable electronic devices due to their substrate flexibility, breathability, and other advantages. This article first introduces the structure and operation mechanism of flexible OECTs, second focuses on the relationship between F-OECT materials, fabrication methods, device structures, and so on and transistor performance, and, finally, describes the application of F-OECTs in various fields (e.g., biomolecule monitoring, gas monitoring, health monitoring, and synaptic neuromorphology), which clarifies the importance of F-OECTs for the development of wearable electronic devices. In this article, it is pointed out that F-OECT should ensure the bendability and stability of the transistor without decreasing the electrical performance (certain switching ratio and transconductance) and analyze the role of electrode materials, semiconductor materials, and preparation process in relation to the switching ratio and transconductance of F-OECT, so as to realize the wide application of F-OECT and provide a reference for the design of wearable electronic devices.
{"title":"Research Progress and Application of Fabric-Based Organic Electrochemical Transistors: A Review","authors":"Jingjie Ma;Yin He;Yanyan Bie;Xiaoying Zheng;Hao Liu;Peng Zhou","doi":"10.1109/JSEN.2024.3516773","DOIUrl":"https://doi.org/10.1109/JSEN.2024.3516773","url":null,"abstract":"With the development of wearable electronic devices and flexible biosensing technology, organic electrochemical transistors (OECTs) have received more and more attention. Fabric-based OECTs (F-OECTs) have a broad application prospect in wearable electronic devices due to their substrate flexibility, breathability, and other advantages. This article first introduces the structure and operation mechanism of flexible OECTs, second focuses on the relationship between F-OECT materials, fabrication methods, device structures, and so on and transistor performance, and, finally, describes the application of F-OECTs in various fields (e.g., biomolecule monitoring, gas monitoring, health monitoring, and synaptic neuromorphology), which clarifies the importance of F-OECTs for the development of wearable electronic devices. In this article, it is pointed out that F-OECT should ensure the bendability and stability of the transistor without decreasing the electrical performance (certain switching ratio and transconductance) and analyze the role of electrode materials, semiconductor materials, and preparation process in relation to the switching ratio and transconductance of F-OECT, so as to realize the wide application of F-OECT and provide a reference for the design of wearable electronic devices.","PeriodicalId":447,"journal":{"name":"IEEE Sensors Journal","volume":"25 4","pages":"5903-5915"},"PeriodicalIF":4.3,"publicationDate":"2025-01-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143422905","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}
An enormous challenge for the harmonic state estimation of distribution networks is how to perceive the complex and varied dynamic harmonics in a higher resolution method. To solve this problem, this article proposes an interval dynamic harmonic high-resolution state estimation method for distribution networks based on multisource measurement data fusion. First, to obtain the typical high-resolution harmonic measurement information of distribution networks under the limited measurement devices, a selection method for the measurement sites of high-resolution power quality monitoring devices (PQMDs) is proposed based on the harmonic electrical distance. On this basis, a multisource data fusion method based on the time period inclusion index is proposed to establish hybrid interval measurement datasets. Second, to improve the efficiency of interval dynamic harmonic state estimation, the interval intermediate variables are introduced to construct the three-stage hybrid interval harmonic measurement equations. Finally, an interval dynamic harmonic high-resolution state estimation method is proposed based on the predictor-corrector method, the IGG-III robust interval Kalman filter (IGGIII-RIKF) is used as the predictor stage, and the forward-backward interval constraint propagation (FBICP) algorithm is used as the corrector stage to realize interval dynamic harmonic high-resolution state estimation. The effectiveness and feasibility of the proposed method have been demonstrated on the IEEE 33-bus system and the IEEE 118-bus system.
{"title":"Interval Dynamic Harmonic High-Resolution State Estimation for Distribution Networks Based on Multisource Measurement Data Fusion","authors":"Tiechao Zhu;Zhenguo Shao;Junjie Lin;Yan Zhang;Feixiong Chen","doi":"10.1109/JSEN.2024.3517674","DOIUrl":"https://doi.org/10.1109/JSEN.2024.3517674","url":null,"abstract":"An enormous challenge for the harmonic state estimation of distribution networks is how to perceive the complex and varied dynamic harmonics in a higher resolution method. To solve this problem, this article proposes an interval dynamic harmonic high-resolution state estimation method for distribution networks based on multisource measurement data fusion. First, to obtain the typical high-resolution harmonic measurement information of distribution networks under the limited measurement devices, a selection method for the measurement sites of high-resolution power quality monitoring devices (PQMDs) is proposed based on the harmonic electrical distance. On this basis, a multisource data fusion method based on the time period inclusion index is proposed to establish hybrid interval measurement datasets. Second, to improve the efficiency of interval dynamic harmonic state estimation, the interval intermediate variables are introduced to construct the three-stage hybrid interval harmonic measurement equations. Finally, an interval dynamic harmonic high-resolution state estimation method is proposed based on the predictor-corrector method, the IGG-III robust interval Kalman filter (IGGIII-RIKF) is used as the predictor stage, and the forward-backward interval constraint propagation (FBICP) algorithm is used as the corrector stage to realize interval dynamic harmonic high-resolution state estimation. The effectiveness and feasibility of the proposed method have been demonstrated on the IEEE 33-bus system and the IEEE 118-bus system.","PeriodicalId":447,"journal":{"name":"IEEE Sensors Journal","volume":"25 4","pages":"6682-6697"},"PeriodicalIF":4.3,"publicationDate":"2025-01-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143446237","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-08DOI: 10.1109/JSEN.2024.3500136
Guangwu Chen;Xin Zhou;Yongbo Si
At present, integrating the Global Navigation Satellite System (GNSS), microelectromechanical system (MEMS), and odometer (OD) is the most practical and low-cost vehicle multifusion navigation system. However, time-varying noise due to satellite signal rejection can lead to serious degradation or even inaccurate positioning accuracy of the system. To overcome this problem, an adaptive spherical simplex unscented Kalman filter (SSUKF), which optimizes the distribution entropy of the innovation based on the Akaike information criterion, is proposed. Initially, the algorithm optimizes the distribution entropy of the SSUKF innovation sequences by considering the Akaike information criterion. Subsequently, it constructs a dynamic equation of the sliding window using the residual and innovative sequences based on covariance matching. Furthermore, the algorithm estimates and adjusts the statistical characteristics of the systematic process and measurement noise online and improves the adaptive ability of the SSUKF. The algorithm overcomes the problem of degradation and dispersion of the filtration accuracy of the SSUKF when there are unknown, inaccurate, or uncertain noise statistics. Finally, simulation and integrated navigation of actual tests were performed. The test outcomes indicate that the proposed algorithm reduces the errors of the east and north velocities by 67.76% and 70.29%, respectively, with root mean square error (RMSE) values of 0.1449 and 0.1308 m/s, respectively. Additionally, when compared to the SSUKF, the proposed algorithm reduces the errors of the latitude and longitude by 56.55% and 81.78%, respectively, with RMSE values of 3.1072 and 1.6076 m, respectively.
{"title":"An Adaptive SSUKF Based on Akaike Information Criterion to Optimize the Distribution Entropy of the Innovation","authors":"Guangwu Chen;Xin Zhou;Yongbo Si","doi":"10.1109/JSEN.2024.3500136","DOIUrl":"https://doi.org/10.1109/JSEN.2024.3500136","url":null,"abstract":"At present, integrating the Global Navigation Satellite System (GNSS), microelectromechanical system (MEMS), and odometer (OD) is the most practical and low-cost vehicle multifusion navigation system. However, time-varying noise due to satellite signal rejection can lead to serious degradation or even inaccurate positioning accuracy of the system. To overcome this problem, an adaptive spherical simplex unscented Kalman filter (SSUKF), which optimizes the distribution entropy of the innovation based on the Akaike information criterion, is proposed. Initially, the algorithm optimizes the distribution entropy of the SSUKF innovation sequences by considering the Akaike information criterion. Subsequently, it constructs a dynamic equation of the sliding window using the residual and innovative sequences based on covariance matching. Furthermore, the algorithm estimates and adjusts the statistical characteristics of the systematic process and measurement noise online and improves the adaptive ability of the SSUKF. The algorithm overcomes the problem of degradation and dispersion of the filtration accuracy of the SSUKF when there are unknown, inaccurate, or uncertain noise statistics. Finally, simulation and integrated navigation of actual tests were performed. The test outcomes indicate that the proposed algorithm reduces the errors of the east and north velocities by 67.76% and 70.29%, respectively, with root mean square error (RMSE) values of 0.1449 and 0.1308 m/s, respectively. Additionally, when compared to the SSUKF, the proposed algorithm reduces the errors of the latitude and longitude by 56.55% and 81.78%, respectively, with RMSE values of 3.1072 and 1.6076 m, respectively.","PeriodicalId":447,"journal":{"name":"IEEE Sensors Journal","volume":"25 4","pages":"6055-6066"},"PeriodicalIF":4.3,"publicationDate":"2025-01-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143422907","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-08DOI: 10.1109/JSEN.2024.3524757
Zhaohan Liu;Yunan Han;Bo Zhou;Xianbo Qiu
This article presents an improved cylindrical cavity sensor combined with machine learning techniques for the measurement of moisture and drug content (DC) in capsules. The sensor consists of a cylindrical cavity, two probe pins, and a transparent plastic tube that enables capsule passage. The cylindrical cavity, crafted with copper gilding, features inner dimensions of $phi ~100times 12$ mm, resulting in a minimum resonant frequency of 2.3 GHz. The proposed measurement method demonstrated an average sensitivity of 17 MHz per percentage of relative moisture content (MC). Two machine learning methods, namely, principal component analysis (PCA) and the Naive Bayes (NB) algorithms are applied to separate capsules with different DCs. Performing the ${S} _{{21}}$ amplitude and phase parameters analysis at 13.19–13.21 GHz, the proposed testing method combined with these two machine learning methods achieved 100% classification accuracy of capsules with different DCs in a single measurement. Furthermore, the classification accuracy of capsules with different DCs in five measurements reached 94%. This methodology offers a microwave sensor designed for the concurrent and accurate assessment of moisture and mass content in items such as cigarettes and coffee beans that can traverse the plastic tube, encompassing, but not restricted to capsules.
{"title":"Cylindrical Cavity Resonating Sensor for Testing Moisture and Drug Content in Capsule Based on Machine Learning","authors":"Zhaohan Liu;Yunan Han;Bo Zhou;Xianbo Qiu","doi":"10.1109/JSEN.2024.3524757","DOIUrl":"https://doi.org/10.1109/JSEN.2024.3524757","url":null,"abstract":"This article presents an improved cylindrical cavity sensor combined with machine learning techniques for the measurement of moisture and drug content (DC) in capsules. The sensor consists of a cylindrical cavity, two probe pins, and a transparent plastic tube that enables capsule passage. The cylindrical cavity, crafted with copper gilding, features inner dimensions of <inline-formula> <tex-math>$phi ~100times 12$ </tex-math></inline-formula> mm, resulting in a minimum resonant frequency of 2.3 GHz. The proposed measurement method demonstrated an average sensitivity of 17 MHz per percentage of relative moisture content (MC). Two machine learning methods, namely, principal component analysis (PCA) and the Naive Bayes (NB) algorithms are applied to separate capsules with different DCs. Performing the <inline-formula> <tex-math>${S} _{{21}}$ </tex-math></inline-formula> amplitude and phase parameters analysis at 13.19–13.21 GHz, the proposed testing method combined with these two machine learning methods achieved 100% classification accuracy of capsules with different DCs in a single measurement. Furthermore, the classification accuracy of capsules with different DCs in five measurements reached 94%. This methodology offers a microwave sensor designed for the concurrent and accurate assessment of moisture and mass content in items such as cigarettes and coffee beans that can traverse the plastic tube, encompassing, but not restricted to capsules.","PeriodicalId":447,"journal":{"name":"IEEE Sensors Journal","volume":"25 4","pages":"6290-6300"},"PeriodicalIF":4.3,"publicationDate":"2025-01-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143430489","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-08DOI: 10.1109/JSEN.2024.3521896
Tao Sun;Tian Tan;Dongxuan Li;Bernd Markert;Peter B. Shull;Franz Bamer
Data diversity and quantity are crucial for training deep-learning models. However, the impact of dataset diversity and size on biomechanical variable estimation models has not been explicitly investigated during drop landings. This work investigates the impact of the number of subjects and the number of trials per subject on the performance of wearable inertial measurement unit (IMU)-driven deep-learning models for knee moment and ground reaction force estimation during drop-landing tasks. An investigation dataset with 16 subjects and 25 trials per subject was collected in a biomechanical laboratory. The impact of subject and trial quantification was explored under different model complexity and types, as well as data augmentation methods using the investigation dataset. The deep-learning models were implemented by a feature extractor and an estimator realized by several fully connected (FC) layers. The feature extractor was independently evaluated with FC neural networks, convolutional neural network (CNN), long short-term memory (LSTM) model, and transformer model. Three transformation-based data augmentation methods were proposed and compared with the measured dataset (MD). The results showed that the minimum required number of subjects and trials for the models to achieve an estimation performance of 0.85 of R-squared, 0.4 body weight $times $ body height of root mean square error (RMSE), and 0.1 of relative RMSE (rRMSE) is five subjects and five trials. Intriguingly, adding more subjects to the dataset improved the estimation performance while adding more trials did not. In addition, the proposed data augmentation can alleviate the data scarcity issue when the number of trials is small.
{"title":"Influence of Number of Subjects and Number of Trials on Biomechanical Variable Estimation via Deep-Learning Models and Wearable IMUs During Drop Landings","authors":"Tao Sun;Tian Tan;Dongxuan Li;Bernd Markert;Peter B. Shull;Franz Bamer","doi":"10.1109/JSEN.2024.3521896","DOIUrl":"https://doi.org/10.1109/JSEN.2024.3521896","url":null,"abstract":"Data diversity and quantity are crucial for training deep-learning models. However, the impact of dataset diversity and size on biomechanical variable estimation models has not been explicitly investigated during drop landings. This work investigates the impact of the number of subjects and the number of trials per subject on the performance of wearable inertial measurement unit (IMU)-driven deep-learning models for knee moment and ground reaction force estimation during drop-landing tasks. An investigation dataset with 16 subjects and 25 trials per subject was collected in a biomechanical laboratory. The impact of subject and trial quantification was explored under different model complexity and types, as well as data augmentation methods using the investigation dataset. The deep-learning models were implemented by a feature extractor and an estimator realized by several fully connected (FC) layers. The feature extractor was independently evaluated with FC neural networks, convolutional neural network (CNN), long short-term memory (LSTM) model, and transformer model. Three transformation-based data augmentation methods were proposed and compared with the measured dataset (MD). The results showed that the minimum required number of subjects and trials for the models to achieve an estimation performance of 0.85 of R-squared, 0.4 body weight <inline-formula> <tex-math>$times $ </tex-math></inline-formula> body height of root mean square error (RMSE), and 0.1 of relative RMSE (rRMSE) is five subjects and five trials. Intriguingly, adding more subjects to the dataset improved the estimation performance while adding more trials did not. In addition, the proposed data augmentation can alleviate the data scarcity issue when the number of trials is small.","PeriodicalId":447,"journal":{"name":"IEEE Sensors Journal","volume":"25 4","pages":"7532-7543"},"PeriodicalIF":4.3,"publicationDate":"2025-01-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143446173","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-08DOI: 10.1109/JSEN.2024.3523849
John A. Berkebile;Asim H. Gazi;Michael Chan;Tyler D. Albarran;Christopher J. Rozell;Omer T. Inan;Paul A. Beach
In neurodegenerative conditions like Parkinson’s disease (PD) and multiple system atrophy (MSA), cardiovascular autonomic dysfunction (CVAD) is associated with several poor long-term health outcomes. CVAD commonly manifests as orthostatic hypotension (OH), a sustained drop in blood pressure (BP) upon standing that can cause syncope and falls. Conventional screening methods for OH are suboptimal and formal autonomic testing is limited to specialized centers. This study explores a multimodal wearable sensing patch for remote monitoring of CVAD. We collected waveform data during clinical autonomic testing and a 24-h period at home from 20 participants with synucleinopathies (12 with OH) and six healthy controls. We developed an automated posture detection pipeline that identified 103 at-home orthostatic events. Then, physiomarkers related to heart rate variability (HRV), cardiac mechanics, and vasomotor function were derived during the supine and standing periods associated with clinical and at-home orthostatic transitions. Comparisons of baroreflex-related supine physiomarkers revealed significant differences between those with and without OH. We characterized cardiovascular autonomic dynamics while standing, leveraging low-dimensional representations, and found marked differences in the aggregate responses between groups. We also observed significantly higher within-subject similarity between the at-home responses of the OH group. Finally, we examined the discriminative power of the patch’s physiomarkers and demonstrated accurate classification of persons with OH during the clinical stand testing (${F} 1=0.83$ ). This study is the first to couple orthostatic event detection with machine learning (ML) analysis of wearable-derived physiomarkers, illustrating that wearable sensing can accurately classify OH and provide novel insights into CVAD outside the clinic.
{"title":"Remote Monitoring of Cardiovascular Autonomic Dysfunction in Synucleinopathies With a Wearable Chest Patch","authors":"John A. Berkebile;Asim H. Gazi;Michael Chan;Tyler D. Albarran;Christopher J. Rozell;Omer T. Inan;Paul A. Beach","doi":"10.1109/JSEN.2024.3523849","DOIUrl":"https://doi.org/10.1109/JSEN.2024.3523849","url":null,"abstract":"In neurodegenerative conditions like Parkinson’s disease (PD) and multiple system atrophy (MSA), cardiovascular autonomic dysfunction (CVAD) is associated with several poor long-term health outcomes. CVAD commonly manifests as orthostatic hypotension (OH), a sustained drop in blood pressure (BP) upon standing that can cause syncope and falls. Conventional screening methods for OH are suboptimal and formal autonomic testing is limited to specialized centers. This study explores a multimodal wearable sensing patch for remote monitoring of CVAD. We collected waveform data during clinical autonomic testing and a 24-h period at home from 20 participants with synucleinopathies (12 with OH) and six healthy controls. We developed an automated posture detection pipeline that identified 103 at-home orthostatic events. Then, physiomarkers related to heart rate variability (HRV), cardiac mechanics, and vasomotor function were derived during the supine and standing periods associated with clinical and at-home orthostatic transitions. Comparisons of baroreflex-related supine physiomarkers revealed significant differences between those with and without OH. We characterized cardiovascular autonomic dynamics while standing, leveraging low-dimensional representations, and found marked differences in the aggregate responses between groups. We also observed significantly higher within-subject similarity between the at-home responses of the OH group. Finally, we examined the discriminative power of the patch’s physiomarkers and demonstrated accurate classification of persons with OH during the clinical stand testing (<inline-formula> <tex-math>${F} 1=0.83$ </tex-math></inline-formula>). This study is the first to couple orthostatic event detection with machine learning (ML) analysis of wearable-derived physiomarkers, illustrating that wearable sensing can accurately classify OH and provide novel insights into CVAD outside the clinic.","PeriodicalId":447,"journal":{"name":"IEEE Sensors Journal","volume":"25 4","pages":"7250-7262"},"PeriodicalIF":4.3,"publicationDate":"2025-01-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143430585","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-08DOI: 10.1109/JSEN.2024.3524909
S. López-Soriano;P. Brunet;Mohammed A. Alsultan;Joan Melià-Seguí
Remote characterization of liquids can be beneficial in various industry sectors such as food and oil industries, medical diagnostics, agriculture, or waste management. However, current wireless solutions are often expensive and labor-intensive. Antenna-based sensors (ABSs) can potentially decrease the complexity and cost of current solutions. Ultrahigh-frequency (UHF) radio frequency identification (RFID) sensors for liquid characterization have the potential to provide remote monitoring while fulfilling the previous requirements. This work demonstrates the combined effects of the dielectric properties on the operation of RFID-based sensors and it presents an innovative approach for estimating the dielectric properties of a liquid under test (LUT) from the read range peak frequency and magnitude variations of a UHF RFID tag. The tag antenna consists of a patch-like antenna with an absorbent embedded into its substrate. Filling the absorbent with different LUTs modifies the dielectric properties of the substrate which has a measurable effect on the tag read range. Measurements show that the proposed method together with the specific sensor design enables the dielectric characterization of liquids using an energy-efficient and low-cost solution achieving an accuracy over 93.5% and 84% in the estimation of the LUT relative permittivity and the loss tangent, respectively, compared to the transmission line (TL) method.
{"title":"Remote Identification of Liquids Using Absorbent Materials: A Passive UHF RFID-Based Method","authors":"S. López-Soriano;P. Brunet;Mohammed A. Alsultan;Joan Melià-Seguí","doi":"10.1109/JSEN.2024.3524909","DOIUrl":"https://doi.org/10.1109/JSEN.2024.3524909","url":null,"abstract":"Remote characterization of liquids can be beneficial in various industry sectors such as food and oil industries, medical diagnostics, agriculture, or waste management. However, current wireless solutions are often expensive and labor-intensive. Antenna-based sensors (ABSs) can potentially decrease the complexity and cost of current solutions. Ultrahigh-frequency (UHF) radio frequency identification (RFID) sensors for liquid characterization have the potential to provide remote monitoring while fulfilling the previous requirements. This work demonstrates the combined effects of the dielectric properties on the operation of RFID-based sensors and it presents an innovative approach for estimating the dielectric properties of a liquid under test (LUT) from the read range peak frequency and magnitude variations of a UHF RFID tag. The tag antenna consists of a patch-like antenna with an absorbent embedded into its substrate. Filling the absorbent with different LUTs modifies the dielectric properties of the substrate which has a measurable effect on the tag read range. Measurements show that the proposed method together with the specific sensor design enables the dielectric characterization of liquids using an energy-efficient and low-cost solution achieving an accuracy over 93.5% and 84% in the estimation of the LUT relative permittivity and the loss tangent, respectively, compared to the transmission line (TL) method.","PeriodicalId":447,"journal":{"name":"IEEE Sensors Journal","volume":"25 4","pages":"7301-7309"},"PeriodicalIF":4.3,"publicationDate":"2025-01-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10834518","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143446286","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}
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：