{"title":"基于物理信息神经网络的短偶极子传感器响应线性化","authors":"Alessandro Fasse, Romain Meyer, Esra Neufeld, Maxim Haas, Nicolas Chavannes, Niels Kuster","doi":"10.1002/bem.70010","DOIUrl":null,"url":null,"abstract":"<p>Short-dipole diode sensors loaded with highly resistive lines are commonly used to measure the time-averaged square of the high-frequency electromagnetic field amplitude directly. Their precision, simplicity, broadband, high dynamic range capability, and minimal scattering make them ideal for application in the near-field of sources, particularly for demonstrating compliance with exposure limits. However, the usage of these sensors to cover multiple orders of magnitude of field amplitude requires signal-specific linearization of the sensor response. Traditionally, linearization had been performed for each signal or modulation by measurement and, more recently, by simulations based on a calibrated sensor model. These approaches have become prohibitively expensive with the launch of the fifth generation of mobile communication (5G), which added thousands of diverse and complex modulation schemes. In response to these challenges, we first developed an innovative approach to accelerate sensor model simulations with an enhancement of accuracy, which allows us to subsequently establish a data set comprising a large number of probe parameters and signal characteristic configurations. Subsequently, a physics-informed neural network (PINN) was trained with readily accessible signal characteristics to obtain on-the-fly linearization parameters with acceptable uncertainties across the relevant dynamic range. In contrast to traditional artificial intelligence (AI) models that predominantly rely on pattern recognition from precomputed data, our approach ensures that the model captures the intrinsic relationships and system dynamics inherent to the physical phenomena under study. Our AI-based approach achieves an error below 0.4 dB at peak specific absorption rate (SAR) values of up to <span></span><math>\n <semantics>\n <mrow>\n <mo>></mo>\n <mspace></mspace>\n <mn>200</mn>\n <mspace></mspace>\n <msup>\n <mstyle>\n <mspace></mspace>\n <mtext>W kg</mtext>\n <mspace></mspace>\n </mstyle>\n <mrow>\n <mo>−</mo>\n <mn>1</mn>\n </mrow>\n </msup>\n </mrow>\n <annotation> $\\gt \\,200\\,{\\text{W kg}}^{-1}$</annotation>\n </semantics></math>. In addition, AI accelerates the determination of linearization parameters by a factor <span></span><math>\n <semantics>\n <mrow>\n <mo>></mo>\n </mrow>\n <annotation> $\\gt $</annotation>\n </semantics></math> 34,000<span></span><math>\n <semantics>\n <mrow>\n <mo>×</mo>\n </mrow>\n <annotation> $\\times $</annotation>\n </semantics></math> and reduces storage requirements <span></span><math>\n <semantics>\n <mrow>\n <mo>></mo>\n </mrow>\n <annotation> $\\gt $</annotation>\n </semantics></math> 350,000 times, allowing linearization parameters to be computed on site.</p>","PeriodicalId":8956,"journal":{"name":"Bioelectromagnetics","volume":"46 4","pages":""},"PeriodicalIF":1.8000,"publicationDate":"2025-05-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/bem.70010","citationCount":"0","resultStr":"{\"title\":\"Short-Dipole Sensor Response Linearization Through Physics-Informed Neural Networks\",\"authors\":\"Alessandro Fasse, Romain Meyer, Esra Neufeld, Maxim Haas, Nicolas Chavannes, Niels Kuster\",\"doi\":\"10.1002/bem.70010\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p>Short-dipole diode sensors loaded with highly resistive lines are commonly used to measure the time-averaged square of the high-frequency electromagnetic field amplitude directly. Their precision, simplicity, broadband, high dynamic range capability, and minimal scattering make them ideal for application in the near-field of sources, particularly for demonstrating compliance with exposure limits. However, the usage of these sensors to cover multiple orders of magnitude of field amplitude requires signal-specific linearization of the sensor response. Traditionally, linearization had been performed for each signal or modulation by measurement and, more recently, by simulations based on a calibrated sensor model. These approaches have become prohibitively expensive with the launch of the fifth generation of mobile communication (5G), which added thousands of diverse and complex modulation schemes. In response to these challenges, we first developed an innovative approach to accelerate sensor model simulations with an enhancement of accuracy, which allows us to subsequently establish a data set comprising a large number of probe parameters and signal characteristic configurations. Subsequently, a physics-informed neural network (PINN) was trained with readily accessible signal characteristics to obtain on-the-fly linearization parameters with acceptable uncertainties across the relevant dynamic range. In contrast to traditional artificial intelligence (AI) models that predominantly rely on pattern recognition from precomputed data, our approach ensures that the model captures the intrinsic relationships and system dynamics inherent to the physical phenomena under study. Our AI-based approach achieves an error below 0.4 dB at peak specific absorption rate (SAR) values of up to <span></span><math>\\n <semantics>\\n <mrow>\\n <mo>></mo>\\n <mspace></mspace>\\n <mn>200</mn>\\n <mspace></mspace>\\n <msup>\\n <mstyle>\\n <mspace></mspace>\\n <mtext>W kg</mtext>\\n <mspace></mspace>\\n </mstyle>\\n <mrow>\\n <mo>−</mo>\\n <mn>1</mn>\\n </mrow>\\n </msup>\\n </mrow>\\n <annotation> $\\\\gt \\\\,200\\\\,{\\\\text{W kg}}^{-1}$</annotation>\\n </semantics></math>. In addition, AI accelerates the determination of linearization parameters by a factor <span></span><math>\\n <semantics>\\n <mrow>\\n <mo>></mo>\\n </mrow>\\n <annotation> $\\\\gt $</annotation>\\n </semantics></math> 34,000<span></span><math>\\n <semantics>\\n <mrow>\\n <mo>×</mo>\\n </mrow>\\n <annotation> $\\\\times $</annotation>\\n </semantics></math> and reduces storage requirements <span></span><math>\\n <semantics>\\n <mrow>\\n <mo>></mo>\\n </mrow>\\n <annotation> $\\\\gt $</annotation>\\n </semantics></math> 350,000 times, allowing linearization parameters to be computed on site.</p>\",\"PeriodicalId\":8956,\"journal\":{\"name\":\"Bioelectromagnetics\",\"volume\":\"46 4\",\"pages\":\"\"},\"PeriodicalIF\":1.8000,\"publicationDate\":\"2025-05-22\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"https://onlinelibrary.wiley.com/doi/epdf/10.1002/bem.70010\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Bioelectromagnetics\",\"FirstCategoryId\":\"99\",\"ListUrlMain\":\"https://onlinelibrary.wiley.com/doi/10.1002/bem.70010\",\"RegionNum\":3,\"RegionCategory\":\"生物学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q3\",\"JCRName\":\"BIOLOGY\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Bioelectromagnetics","FirstCategoryId":"99","ListUrlMain":"https://onlinelibrary.wiley.com/doi/10.1002/bem.70010","RegionNum":3,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"BIOLOGY","Score":null,"Total":0}
Short-Dipole Sensor Response Linearization Through Physics-Informed Neural Networks
Short-dipole diode sensors loaded with highly resistive lines are commonly used to measure the time-averaged square of the high-frequency electromagnetic field amplitude directly. Their precision, simplicity, broadband, high dynamic range capability, and minimal scattering make them ideal for application in the near-field of sources, particularly for demonstrating compliance with exposure limits. However, the usage of these sensors to cover multiple orders of magnitude of field amplitude requires signal-specific linearization of the sensor response. Traditionally, linearization had been performed for each signal or modulation by measurement and, more recently, by simulations based on a calibrated sensor model. These approaches have become prohibitively expensive with the launch of the fifth generation of mobile communication (5G), which added thousands of diverse and complex modulation schemes. In response to these challenges, we first developed an innovative approach to accelerate sensor model simulations with an enhancement of accuracy, which allows us to subsequently establish a data set comprising a large number of probe parameters and signal characteristic configurations. Subsequently, a physics-informed neural network (PINN) was trained with readily accessible signal characteristics to obtain on-the-fly linearization parameters with acceptable uncertainties across the relevant dynamic range. In contrast to traditional artificial intelligence (AI) models that predominantly rely on pattern recognition from precomputed data, our approach ensures that the model captures the intrinsic relationships and system dynamics inherent to the physical phenomena under study. Our AI-based approach achieves an error below 0.4 dB at peak specific absorption rate (SAR) values of up to . In addition, AI accelerates the determination of linearization parameters by a factor 34,000 and reduces storage requirements 350,000 times, allowing linearization parameters to be computed on site.
期刊介绍:
Bioelectromagnetics is published by Wiley-Liss, Inc., for the Bioelectromagnetics Society and is the official journal of the Bioelectromagnetics Society and the European Bioelectromagnetics Association. It is a peer-reviewed, internationally circulated scientific journal that specializes in reporting original data on biological effects and applications of electromagnetic fields that range in frequency from zero hertz (static fields) to the terahertz undulations and visible light. Both experimental and clinical data are of interest to the journal''s readers as are theoretical papers or reviews that offer novel insights into or criticism of contemporary concepts and theories of field-body interactions. The Bioelectromagnetics Society, which sponsors the journal, also welcomes experimental or clinical papers on the domains of sonic and ultrasonic radiation.