Pub Date : 2024-09-18DOI: 10.1109/LSENS.2024.3463977
Yan Siang Yap;Mohd Ridzuan Ahmad
This letter explores the architecture of tiny machine learning (TinyML). Deploying machine learning into embedded devices is challenging due to the limited computation power and memory space. An experimental setup has been designed for the anomaly detection of a USB fan. We collect the normal data from a USB fan, and abnormal data are simulated using a broken fan blade. Two different speeds, namely, speed 1 and speed 2, have been used to collect the normal data and abnormal data. The normal data collected are used to train the standard autoencoder model and our proposed model modified overcomplete asymmetric autoencoder (MOA), respectively. The trained model is then deployed into a microcontroller, i.e., Arduino Nano 33 BLE Sense. The proposed MOA can achieve 99.23% accuracy, recall of 99.70%, precision of 98.77%, F1 score of 99.23%, and false positive rate of 1.222%. Besides that, our MOA model only occupies 17 kB. Therefore, it can be fitted into most microcontrollers for embedded applications.
这封信探讨了微型机器学习（TinyML）的架构。由于计算能力和内存空间有限，在嵌入式设备中部署机器学习具有挑战性。我们为 USB 风扇的异常检测设计了一个实验装置。我们收集了 USB 风扇的正常数据，并使用断裂的风扇叶片模拟异常数据。我们使用两种不同的速度（即速度 1 和速度 2）来收集正常数据和异常数据。收集到的正常数据分别用于训练标准自动编码器模型和我们提出的修正过完整非对称自动编码器（MOA）模型。然后将训练好的模型部署到微控制器中，即 Arduino Nano 33 BLE Sense。所提出的 MOA 准确率为 99.23%，召回率为 99.70%，精确率为 98.77%，F1 分数为 99.23%，误报率为 1.222%。此外，我们的 MOA 模型仅占 17 kB。因此，它可以安装在大多数嵌入式应用的微控制器中。
{"title":"Modified Overcomplete Autoencoder for Anomaly Detection Based on TinyML","authors":"Yan Siang Yap;Mohd Ridzuan Ahmad","doi":"10.1109/LSENS.2024.3463977","DOIUrl":"https://doi.org/10.1109/LSENS.2024.3463977","url":null,"abstract":"This letter explores the architecture of tiny machine learning (TinyML). Deploying machine learning into embedded devices is challenging due to the limited computation power and memory space. An experimental setup has been designed for the anomaly detection of a USB fan. We collect the normal data from a USB fan, and abnormal data are simulated using a broken fan blade. Two different speeds, namely, speed 1 and speed 2, have been used to collect the normal data and abnormal data. The normal data collected are used to train the standard autoencoder model and our proposed model modified overcomplete asymmetric autoencoder (MOA), respectively. The trained model is then deployed into a microcontroller, i.e., Arduino Nano 33 BLE Sense. The proposed MOA can achieve 99.23% accuracy, recall of 99.70%, precision of 98.77%, F1 score of 99.23%, and false positive rate of 1.222%. Besides that, our MOA model only occupies 17 kB. Therefore, it can be fitted into most microcontrollers for embedded applications.","PeriodicalId":13014,"journal":{"name":"IEEE Sensors Letters","volume":null,"pages":null},"PeriodicalIF":2.2,"publicationDate":"2024-09-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142430816","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 : 2024-09-17DOI: 10.1109/LSENS.2024.3462485
Radhika Raina;Kamal Jeet Singh;Suman Kumar
Onions are a valuable cash crop for farmers, providing a reliable source of income; thus, monitoring of the quality of onions kept in storage is an important concern. There are various factors, such as temperature, humidity, and storage period, that are responsible for maintaining the quality of onion. The common factor is, onion emits various gases when it starts rotting. Thus, to address this issue, carbon dioxide (CO 2 ), sulphur dioxide (SO 2 ), hydrogen sulphide (H 2 S), ammonia (NH 3 ), temperature and humidity (SHT40) sensors are used in the proposed onion quality monitoring system. In this letter, we present the approximate ranges of the sensors through repeated experiments on three types of onions: healthy, those beginning to rot and fully rotted onions. In addition, our experiments and the literature both indicate that H 2 S gas is the most effective for early rot detection. Moreover, none of the existing literature works have discussed regarding the power consumption of the onion quality monitoring system. Therefore, a novel battery operated, power efficient onion monitoring device is designed, primarily using H 2 S and SHT40 sensors. This setup has a battery life of approximately 6.03 days with an 11.1 V / 10 Ah battery. When H 2 S levels exceed a threshold indicating the onset of onion rot, all sensors (CO 2 , SO 2 , H 2 S, NH 3 , and SHT40) are activated, reducing battery life to 5.41 days.
{"title":"Gas, Temperature, and Humidity Sensors-Based Onion Quality Monitoring System","authors":"Radhika Raina;Kamal Jeet Singh;Suman Kumar","doi":"10.1109/LSENS.2024.3462485","DOIUrl":"https://doi.org/10.1109/LSENS.2024.3462485","url":null,"abstract":"Onions are a valuable cash crop for farmers, providing a reliable source of income; thus, monitoring of the quality of onions kept in storage is an important concern. There are various factors, such as temperature, humidity, and storage period, that are responsible for maintaining the quality of onion. The common factor is, onion emits various gases when it starts rotting. Thus, to address this issue, carbon dioxide (CO\u0000<sub>2</sub>\u0000), sulphur dioxide (SO\u0000<sub>2</sub>\u0000), hydrogen sulphide (H\u0000<sub>2</sub>\u0000S), ammonia (NH\u0000<sub>3</sub>\u0000), temperature and humidity (SHT40) sensors are used in the proposed onion quality monitoring system. In this letter, we present the approximate ranges of the sensors through repeated experiments on three types of onions: healthy, those beginning to rot and fully rotted onions. In addition, our experiments and the literature both indicate that H\u0000<sub>2</sub>\u0000S gas is the most effective for early rot detection. Moreover, none of the existing literature works have discussed regarding the power consumption of the onion quality monitoring system. Therefore, a novel battery operated, power efficient onion monitoring device is designed, primarily using H\u0000<sub>2</sub>\u0000S and SHT40 sensors. This setup has a battery life of approximately 6.03 days with an 11.1 V / 10 Ah battery. When H\u0000<sub>2</sub>\u0000S levels exceed a threshold indicating the onset of onion rot, all sensors (CO\u0000<sub>2</sub>\u0000, SO\u0000<sub>2</sub>\u0000, H\u0000<sub>2</sub>\u0000S, NH\u0000<sub>3</sub>\u0000, and SHT40) are activated, reducing battery life to 5.41 days.","PeriodicalId":13014,"journal":{"name":"IEEE Sensors Letters","volume":null,"pages":null},"PeriodicalIF":2.2,"publicationDate":"2024-09-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142397407","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 : 2024-09-17DOI: 10.1109/LSENS.2024.3462600
Maissa Taktak;Mohamed Khalil Baazaoui;Ilef Ketata;Salwa Sahnoun;Ahmed Fakhfakh;Faouzi Derbel
Precise distance estimation is essential in various fields, influencing customary aspects from daily activities to advanced research. In wireless sensor networks (WSN) accurate distance estimation is crucial for different applications, such as localization, energy efficiency, dynamic routing, and coverage optimization. In this letter, we strive to assess distance accurate estimation across various technologies, including a sub-GHz low-power, low-data-rate radio chip, and the ultra-wideband (UWB) transceiver. We utilize a combination of Time-of-Flight (ToF), link quality metrics (LQM), and machine learning (ML) techniques to elucidate the strengths and limitations of each technology.
{"title":"AI-Enhanced Distance Estimation via Radio Chip Link Quality Metrics and Time-of-Flight Analysis With UWB Technology: A Comparative Evaluation","authors":"Maissa Taktak;Mohamed Khalil Baazaoui;Ilef Ketata;Salwa Sahnoun;Ahmed Fakhfakh;Faouzi Derbel","doi":"10.1109/LSENS.2024.3462600","DOIUrl":"https://doi.org/10.1109/LSENS.2024.3462600","url":null,"abstract":"Precise distance estimation is essential in various fields, influencing customary aspects from daily activities to advanced research. In wireless sensor networks (WSN) accurate distance estimation is crucial for different applications, such as localization, energy efficiency, dynamic routing, and coverage optimization. In this letter, we strive to assess distance accurate estimation across various technologies, including a sub-GHz low-power, low-data-rate radio chip, and the ultra-wideband (UWB) transceiver. We utilize a combination of Time-of-Flight (ToF), link quality metrics (LQM), and machine learning (ML) techniques to elucidate the strengths and limitations of each technology.","PeriodicalId":13014,"journal":{"name":"IEEE Sensors Letters","volume":null,"pages":null},"PeriodicalIF":2.2,"publicationDate":"2024-09-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142377120","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 : 2024-09-16DOI: 10.1109/LSENS.2024.3461811
Xuandong Liang;Yukun Long;Kaiyuan Xiang;Peiyou Shi;Zhuo Wang;Kun Xiao;Santosh Kumar;Xiaoli Li;Rui Min
The crucial role of pulse signals in the field of cardiovascular diseases (CVDs) cannot be overstated, as they provide physiological indicators for cardiovascular health monitoring and disease diagnosis. Pulse signals, as fundamental physiological signals, also offer valuable insights into heart rate variability (HRV) analysis for diagnosing CVDs. The wearable sensors are promising to monitor HRV information. We propose a smartphone-integrated plastic optical fiber (POF) speckle sensor for HRV monitoring, POF with a core diameter of 1000 µm was implemented, which offers the highest signal-to-noise ratio value (11.33 dB) among POFs of this core diameter. HRV tests were conducted, revealing a participant's average heart rate, standard deviation of NN intervals, and Root mean square of successive differences during different motion states. The correlation coefficient between reference R wave to R wave (RR) intervals and measured RR intervals is 0.97018. The results indicate our system holds potential applications in monitoring HRV for prevention and treatment of CVDs.
脉搏信号为心血管健康监测和疾病诊断提供了生理指标，因此在心血管疾病（CVDs）领域发挥着至关重要的作用。脉搏信号作为基本的生理信号，还为心率变异性（HRV）分析提供了诊断心血管疾病的宝贵见解。可穿戴传感器在监测心率变异信息方面大有可为。我们提出了一种用于心率变异监测的智能手机集成塑料光纤（POF）斑点传感器，其芯直径为 1000 微米，在该芯直径的 POF 中具有最高的信噪比值（11.33 dB）。心率变异测试显示了参与者的平均心率、NN 间隔的标准偏差以及不同运动状态下连续差值的均方根。参考 R 波至 R 波 (RR) 间期与测量的 RR 间期之间的相关系数为 0.97018。结果表明，我们的系统在监测心率变异以预防和治疗心血管疾病方面具有潜在的应用价值。
{"title":"Smartphone-Integrated POF Speckle Sensor for Heart Rate Variability Monitoring","authors":"Xuandong Liang;Yukun Long;Kaiyuan Xiang;Peiyou Shi;Zhuo Wang;Kun Xiao;Santosh Kumar;Xiaoli Li;Rui Min","doi":"10.1109/LSENS.2024.3461811","DOIUrl":"https://doi.org/10.1109/LSENS.2024.3461811","url":null,"abstract":"The crucial role of pulse signals in the field of cardiovascular diseases (CVDs) cannot be overstated, as they provide physiological indicators for cardiovascular health monitoring and disease diagnosis. Pulse signals, as fundamental physiological signals, also offer valuable insights into heart rate variability (HRV) analysis for diagnosing CVDs. The wearable sensors are promising to monitor HRV information. We propose a smartphone-integrated plastic optical fiber (POF) speckle sensor for HRV monitoring, POF with a core diameter of 1000 µm was implemented, which offers the highest signal-to-noise ratio value (11.33 dB) among POFs of this core diameter. HRV tests were conducted, revealing a participant's average heart rate, standard deviation of NN intervals, and Root mean square of successive differences during different motion states. The correlation coefficient between reference R wave to R wave (RR) intervals and measured RR intervals is 0.97018. The results indicate our system holds potential applications in monitoring HRV for prevention and treatment of CVDs.","PeriodicalId":13014,"journal":{"name":"IEEE Sensors Letters","volume":null,"pages":null},"PeriodicalIF":2.2,"publicationDate":"2024-09-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142368498","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 : 2024-09-13DOI: 10.1109/LSENS.2024.3460399
Roberto S. Aga;Lemuel Duncan;Laura Davidson;Fahima Ouchen;Rachel Aga;Emily M. Heckman;Carrie M. Bartsch
This letter reports a novel design of a resistance strain sensor and its method of fabrication. The sensor is made of printed silver (Ag), but its sensitivity, which is measured by the gauge factor (GF), surpasses most commercial metal foil strain gauges (COTS). These COTS have a low GF (∼2) because they exhibit a weak piezoresistive effect. As a result, their sensitivity is dominated by the geometric effect. In this design, the GF is not limited by the weak piezoresistive effect in Ag. Its enhanced sensitivity (GF∼55) originates from the junctions that are created when a conductive cross-pattern is laser sintered on a printed Ag pad. The cross-pattern consists of a low-resistivity vertical trace and a high-resistivity horizontal trace. The difference in resistivity is achieved by changing the laser sintering power. The junction that joins the high and the low resistivity traces is a boundary with interfacial resistance. This interfacial resistance exhibits high sensitivity to strain leading to a different design and fabrication of a resistance strain sensor.
这封信报告了电阻应变传感器的新型设计及其制造方法。该传感器由印刷银（Ag）制成，但其灵敏度（通过量规系数（GF）测量）超过了大多数商用金属箔应变片（COTS）。这些 COTS 的 GF 值（∼2）很低，因为它们的压阻效应很弱。因此，它们的灵敏度主要受几何效应的影响。在这种设计中，GF 不受 Ag 中弱压阻效应的限制。其更高的灵敏度（GF∼55）源于在印刷的银垫上激光烧结导电交叉图案时产生的结点。交叉图案由一条低电阻率垂直轨迹和一条高电阻率水平轨迹组成。电阻率的差异是通过改变激光烧结功率来实现的。连接高电阻率和低电阻率轨迹的交界处是具有界面电阻的边界。这种界面电阻对应变的敏感度很高，因此需要设计和制造不同的电阻应变传感器。
{"title":"Design and Fabrication of a Metal Resistance Strain Sensor With Enhanced Sensitivity","authors":"Roberto S. Aga;Lemuel Duncan;Laura Davidson;Fahima Ouchen;Rachel Aga;Emily M. Heckman;Carrie M. Bartsch","doi":"10.1109/LSENS.2024.3460399","DOIUrl":"https://doi.org/10.1109/LSENS.2024.3460399","url":null,"abstract":"This letter reports a novel design of a resistance strain sensor and its method of fabrication. The sensor is made of printed silver (Ag), but its sensitivity, which is measured by the gauge factor (GF), surpasses most commercial metal foil strain gauges (COTS). These COTS have a low GF (∼2) because they exhibit a weak piezoresistive effect. As a result, their sensitivity is dominated by the geometric effect. In this design, the GF is not limited by the weak piezoresistive effect in Ag. Its enhanced sensitivity (GF∼55) originates from the junctions that are created when a conductive cross-pattern is laser sintered on a printed Ag pad. The cross-pattern consists of a low-resistivity vertical trace and a high-resistivity horizontal trace. The difference in resistivity is achieved by changing the laser sintering power. The junction that joins the high and the low resistivity traces is a boundary with interfacial resistance. This interfacial resistance exhibits high sensitivity to strain leading to a different design and fabrication of a resistance strain sensor.","PeriodicalId":13014,"journal":{"name":"IEEE Sensors Letters","volume":null,"pages":null},"PeriodicalIF":2.2,"publicationDate":"2024-09-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142359739","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 : 2024-09-13DOI: 10.1109/LSENS.2024.3460969
Tina Mitteramskogler;Andreas Fuchsluger;Rafael Ecker;Sebastian Lang;Thomas Wilfinger;Robert Wille;Bernhard Jakoby
The success of semiconductor industry, providing a high- volume, high-accuracy fabrication method of sensor chips, has caused sensors to be omnipresent in everyday consumer products. Typically, these sensors are enclosed into sensor packages and further integrated onto printed circuit boards since, for the connection to the outer world, several length scales have to be bridged. In this work, we show how surface-mount devices (SMDs) can be directly integrated onto poly(methyl methacrylate) (PMMA) chips through the use of open microchannels. To this end, the SMDs are directly placed onto structured PMMA plates with open microchannels connecting them to dedicated liquid reservoirs. When introducing conductive inks to those reservoirs, capillary forces draw the liquid toward the SMDs and ensure the electrical connection between the liquid reservoir and the SMDs themselves. With the addition of crossings and meandering conductive lines, this process can be used for the fabrication of electrical networks out of individual SMD components directly on a PMMA substrate.
{"title":"Integration of Surface-Mount Devices in Microsystems Using Tracks Consisting of Nanoparticles","authors":"Tina Mitteramskogler;Andreas Fuchsluger;Rafael Ecker;Sebastian Lang;Thomas Wilfinger;Robert Wille;Bernhard Jakoby","doi":"10.1109/LSENS.2024.3460969","DOIUrl":"https://doi.org/10.1109/LSENS.2024.3460969","url":null,"abstract":"The success of semiconductor industry, providing a high- volume, high-accuracy fabrication method of sensor chips, has caused sensors to be omnipresent in everyday consumer products. Typically, these sensors are enclosed into sensor packages and further integrated onto printed circuit boards since, for the connection to the outer world, several length scales have to be bridged. In this work, we show how surface-mount devices (SMDs) can be directly integrated onto poly(methyl methacrylate) (PMMA) chips through the use of open microchannels. To this end, the SMDs are directly placed onto structured PMMA plates with open microchannels connecting them to dedicated liquid reservoirs. When introducing conductive inks to those reservoirs, capillary forces draw the liquid toward the SMDs and ensure the electrical connection between the liquid reservoir and the SMDs themselves. With the addition of crossings and meandering conductive lines, this process can be used for the fabrication of electrical networks out of individual SMD components directly on a PMMA substrate.","PeriodicalId":13014,"journal":{"name":"IEEE Sensors Letters","volume":null,"pages":null},"PeriodicalIF":2.2,"publicationDate":"2024-09-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10680343","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142313146","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}
Pub Date : 2024-09-13DOI: 10.1109/LSENS.2024.3460975
Arslan Majal;Aamir Hussain Chughtai
In this letter, we propose a version of the unscented Rauch–Tung–Striebel (RTS) smoother robust to outliers in the observations. We consider a common case where data are collected from independent sensors with additive white Gaussian noise. Our method is primarily motivated by recent arguments and results presented in favor of learning-based outlier-robust state estimators, which assume adaptive residual cost functions in their formulation. We resort to the variational Bayesian (VB) theory to design an algorithm that selectively discards the corrupted measurements unlike other learning-based methods. Moreover, with the assumption that data are obtained from independent sensors, we are able to leverage computational results from advances in the unscented filtering theory that exploit the sparsity in the measurement covariance. For performance bench-marking, we present the Bayesian Cramér–Rao bound for a smoother with perfect outliers detection and rejection capabilities. Numerical experiments under different scenarios showcase performance gains in comparison with similar learning-based smoothers derived with the VB approach.
{"title":"Outlier-Robust Unscented RTS Smoothing for Independent Sensing Data","authors":"Arslan Majal;Aamir Hussain Chughtai","doi":"10.1109/LSENS.2024.3460975","DOIUrl":"https://doi.org/10.1109/LSENS.2024.3460975","url":null,"abstract":"In this letter, we propose a version of the unscented Rauch–Tung–Striebel (RTS) smoother robust to outliers in the observations. We consider a common case where data are collected from independent sensors with additive white Gaussian noise. Our method is primarily motivated by recent arguments and results presented in favor of learning-based outlier-robust state estimators, which assume adaptive residual cost functions in their formulation. We resort to the variational Bayesian (VB) theory to design an algorithm that selectively discards the corrupted measurements unlike other learning-based methods. Moreover, with the assumption that data are obtained from independent sensors, we are able to leverage computational results from advances in the unscented filtering theory that exploit the sparsity in the measurement covariance. For performance bench-marking, we present the Bayesian Cramér–Rao bound for a smoother with perfect outliers detection and rejection capabilities. Numerical experiments under different scenarios showcase performance gains in comparison with similar learning-based smoothers derived with the VB approach.","PeriodicalId":13014,"journal":{"name":"IEEE Sensors Letters","volume":null,"pages":null},"PeriodicalIF":2.2,"publicationDate":"2024-09-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142328426","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 : 2024-09-12DOI: 10.1109/LSENS.2024.3458931
Kaveti Pavan;Ankit Singh;Tomohiko Igasaki;Digvijay S. Pawar;Nagarajan Ganapathy
Textile sensors enable noninvasive health monitoring, crucial for ensuring road safety by conducting mental well-being checks for drivers. Assessing driver stress with multimodal data from textile electrodes requires effectively integrating and interpreting diverse physiological signals and kinematic data. In this study, we evaluated textile electrodes and cross-attention mechanisms for assessing driver stress using multimodal data. Electrocardiography and respiration data were collected from 15 healthy volunteers wearing smart shirts in two driving scenarios. Signals were sampled at 256 and 128 Hz, respectively, with vehicle data also recorded. Segmented physiological and vehicle data enter separate networks, 1-D convolutional layers for signals and fully connected layers for vehicle data. Cross-attention fuses physiological data; these features are combined with vehicle data for stress classification using sigmoid. The proposed approach is able to classify driver stress states using multimodal data, achieving an average accuracy of 79 ${%}$ and an average F-score of 75 ${%}$ . The integration of a cross-attention mechanism facilitates the capture of intermodality information.
{"title":"Assessment of Driver's Stress State Using Smart T-Shirt Textile Electrodes and Multimodal Cross-Attention Networks","authors":"Kaveti Pavan;Ankit Singh;Tomohiko Igasaki;Digvijay S. Pawar;Nagarajan Ganapathy","doi":"10.1109/LSENS.2024.3458931","DOIUrl":"https://doi.org/10.1109/LSENS.2024.3458931","url":null,"abstract":"Textile sensors enable noninvasive health monitoring, crucial for ensuring road safety by conducting mental well-being checks for drivers. Assessing driver stress with multimodal data from textile electrodes requires effectively integrating and interpreting diverse physiological signals and kinematic data. In this study, we evaluated textile electrodes and cross-attention mechanisms for assessing driver stress using multimodal data. Electrocardiography and respiration data were collected from 15 healthy volunteers wearing smart shirts in two driving scenarios. Signals were sampled at 256 and 128 Hz, respectively, with vehicle data also recorded. Segmented physiological and vehicle data enter separate networks, 1-D convolutional layers for signals and fully connected layers for vehicle data. Cross-attention fuses physiological data; these features are combined with vehicle data for stress classification using sigmoid. The proposed approach is able to classify driver stress states using multimodal data, achieving an average accuracy of 79\u0000<inline-formula><tex-math>${%}$</tex-math></inline-formula>\u0000 and an average F-score of 75\u0000<inline-formula><tex-math>${%}$</tex-math></inline-formula>\u0000. The integration of a cross-attention mechanism facilitates the capture of intermodality information.","PeriodicalId":13014,"journal":{"name":"IEEE Sensors Letters","volume":null,"pages":null},"PeriodicalIF":2.2,"publicationDate":"2024-09-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142397408","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 : 2024-09-12DOI: 10.1109/LSENS.2024.3459939
Weiyi Li;Jiangang Liu;Shisheng Guo;Yong Jia
The human body displays typical properties of multiple sca-ttering when targeted by radar, and the strong echo signal scattered from the torso often masks the weak echo signal scattered from the other body parts such as limbs and the head, limiting the performance of activity recognition. To address this issue, a human activity recognition method based on scattering separation using multifrequency radar data is proposed. First, the multifrequency echo data of human activity collected from stepped-frequency continuous wave radar is stacked, followed by principal component analysis to separate the trunk signal and the branch signal into the first two components, which helps in avoiding the interference caused by masking effects. Subsequently, the two time-frequency spectrograms that express the characteristics of human activities jointly are put into two parallel convolutional neural networks to complete feature extraction and classification. Datasets encompassing six activities were gathered using the stepped-frequency radar. Test results demonstrate that this method can enhance the average recognition effectually compared to the approach without scattering separation.
{"title":"Human Activity Recognition Method Based on Scattering Separation Using Multifrequency Radar Data","authors":"Weiyi Li;Jiangang Liu;Shisheng Guo;Yong Jia","doi":"10.1109/LSENS.2024.3459939","DOIUrl":"https://doi.org/10.1109/LSENS.2024.3459939","url":null,"abstract":"The human body displays typical properties of multiple sca-ttering when targeted by radar, and the strong echo signal scattered from the torso often masks the weak echo signal scattered from the other body parts such as limbs and the head, limiting the performance of activity recognition. To address this issue, a human activity recognition method based on scattering separation using multifrequency radar data is proposed. First, the multifrequency echo data of human activity collected from stepped-frequency continuous wave radar is stacked, followed by principal component analysis to separate the trunk signal and the branch signal into the first two components, which helps in avoiding the interference caused by masking effects. Subsequently, the two time-frequency spectrograms that express the characteristics of human activities jointly are put into two parallel convolutional neural networks to complete feature extraction and classification. Datasets encompassing six activities were gathered using the stepped-frequency radar. Test results demonstrate that this method can enhance the average recognition effectually compared to the approach without scattering separation.","PeriodicalId":13014,"journal":{"name":"IEEE Sensors Letters","volume":null,"pages":null},"PeriodicalIF":2.2,"publicationDate":"2024-09-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142328403","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}
Advances in textile materials research and the health sector have contributed significantly to the invention of various flexible, smart, conductive materials that serve as electrodes for continuous monitoring of human physiological signals, such as electrocardiogram (ECG), EEG, electromyography (EMG), etc. EMG is used to record and evaluate the electrophysiological activity of the human muscle. The surface EMG signals are generally measured with an Ag/AgCl gel electrode. Although a circle-shaped electrode is significantly used, there will be a chance of evidence that another shape may be preferred. Therefore, this study investigated the performance of different electrode shapes, such as circles, triangles, rectangles, and squares, from their sensitivity, stability, and range of the electrode developed from the coated conductive textile material. This acts as an alternative to conventional gel electrodes. The dry electrodes were developed using conductive carbon paint coated with cotton fabric. The area of the electrode is 3.75 cm 2 , where the area is maintained the same for all the shapes. The morphological structure of the material is analyzed using a scanning electron microscope (SEM). The dry electrode resistance for each shape is calculated from IV measurement. The resistance of triangle- and rectangle-shape dry electrode is slightly higher than that of the circle- and square-shaped dry electrode. Even though the triangle-shaped electrode has higher sensitivity than other shapes, the rectangle-shaped electrode shows more deviation in signal range, which is 0.7 to −0.3 mV, while the range of commercial electrodes is 0.1 to −0.2 mV. The circular and square-shaped dry electrode signals are quite similar and within the range of commercial electrodes with less sensitivity value. Further, the circular and square-shaped electrodes are analyzed with different sizes, and it is observed that the larger the surface area, the lower the resistance with reliable signal. The EMG signal acquisition without interconnection is also observed in this research.
{"title":"Durability Design and Construction Enhancement of Textile Electrode Shape on Analysis of Its Surface Electromyography Signal","authors":"Subashini J M;L Gethsia Judin;Amirdha S;Sharath Chandra Vislavath;Rohini Palanisamy;Suhasini Sathiyamoorthy;Pandiyarasan Veluswamy","doi":"10.1109/LSENS.2024.3458906","DOIUrl":"https://doi.org/10.1109/LSENS.2024.3458906","url":null,"abstract":"Advances in textile materials research and the health sector have contributed significantly to the invention of various flexible, smart, conductive materials that serve as electrodes for continuous monitoring of human physiological signals, such as electrocardiogram (ECG), EEG, electromyography (EMG), etc. EMG is used to record and evaluate the electrophysiological activity of the human muscle. The surface EMG signals are generally measured with an Ag/AgCl gel electrode. Although a circle-shaped electrode is significantly used, there will be a chance of evidence that another shape may be preferred. Therefore, this study investigated the performance of different electrode shapes, such as circles, triangles, rectangles, and squares, from their sensitivity, stability, and range of the electrode developed from the coated conductive textile material. This acts as an alternative to conventional gel electrodes. The dry electrodes were developed using conductive carbon paint coated with cotton fabric. The area of the electrode is 3.75 cm\u0000<sup>2</sup>\u0000, where the area is maintained the same for all the shapes. The morphological structure of the material is analyzed using a scanning electron microscope (SEM). The dry electrode resistance for each shape is calculated from IV measurement. The resistance of triangle- and rectangle-shape dry electrode is slightly higher than that of the circle- and square-shaped dry electrode. Even though the triangle-shaped electrode has higher sensitivity than other shapes, the rectangle-shaped electrode shows more deviation in signal range, which is 0.7 to −0.3 mV, while the range of commercial electrodes is 0.1 to −0.2 mV. The circular and square-shaped dry electrode signals are quite similar and within the range of commercial electrodes with less sensitivity value. Further, the circular and square-shaped electrodes are analyzed with different sizes, and it is observed that the larger the surface area, the lower the resistance with reliable signal. The EMG signal acquisition without interconnection is also observed in this research.","PeriodicalId":13014,"journal":{"name":"IEEE Sensors Letters","volume":null,"pages":null},"PeriodicalIF":2.2,"publicationDate":"2024-09-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142397405","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}
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：