Pub Date : 2018-03-26DOI: 10.1109/COMPEM.2018.8496543
Yong-sheng Li, E. Li, Huan Yu, Hanju Oh, Muhannad S. Bakir, Madhavan Swaminathan
Thermal management for 3-D ICs is not only important but also challenging. While air-cooled heat sink is agreed to become incapable for 3-D ICs, microchannel cooling has provided a better solution. In this paper, a machine learning method, Bayesian Optimization (BO), is applied in 3-D ICs with a time-dependent power map to intelligently control the flow rates of the tier-specific microfluidic heatsink (MFHS) for dynamic thermal management (DTM).
{"title":"Machine Learning for 3D-IC Electric-Thermal Simulation and Management","authors":"Yong-sheng Li, E. Li, Huan Yu, Hanju Oh, Muhannad S. Bakir, Madhavan Swaminathan","doi":"10.1109/COMPEM.2018.8496543","DOIUrl":"https://doi.org/10.1109/COMPEM.2018.8496543","url":null,"abstract":"Thermal management for 3-D ICs is not only important but also challenging. While air-cooled heat sink is agreed to become incapable for 3-D ICs, microchannel cooling has provided a better solution. In this paper, a machine learning method, Bayesian Optimization (BO), is applied in 3-D ICs with a time-dependent power map to intelligently control the flow rates of the tier-specific microfluidic heatsink (MFHS) for dynamic thermal management (DTM).","PeriodicalId":221352,"journal":{"name":"2018 IEEE International Conference on Computational Electromagnetics (ICCEM)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2018-03-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130163233","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 : 2018-03-26DOI: 10.1109/COMPEM.2018.8496558
K. Zhao, L. Petersson
This paper presents a multi-region hybrid solver for solving real-life challenging problems. The method provides a systematic approach based on domain decomposition to integrate various electromagnetic solvers into one powerful hybrid solver. In this work, we have combined finite element-boundary integral method, method of moments for both metallic and dielectric structures, and asymptotic high frequency methods such as physical optics and shooting and bouncing rays.
{"title":"A Robust Hybrid Electromagnetic Solver for Electrically Large Problems","authors":"K. Zhao, L. Petersson","doi":"10.1109/COMPEM.2018.8496558","DOIUrl":"https://doi.org/10.1109/COMPEM.2018.8496558","url":null,"abstract":"This paper presents a multi-region hybrid solver for solving real-life challenging problems. The method provides a systematic approach based on domain decomposition to integrate various electromagnetic solvers into one powerful hybrid solver. In this work, we have combined finite element-boundary integral method, method of moments for both metallic and dielectric structures, and asymptotic high frequency methods such as physical optics and shooting and bouncing rays.","PeriodicalId":221352,"journal":{"name":"2018 IEEE International Conference on Computational Electromagnetics (ICCEM)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2018-03-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122634777","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 : 2018-03-26DOI: 10.1109/COMPEM.2018.8496654
Hao Du, Yuan He, T. Jin
Human activities classification has drawn great attention due to its potential applications in security, surveillance and gesture-based interface. The movements of the human body and limbs result in unique micro-Doppler features which can be exploited for identification of human behavior. In this work, we propose a transfer-learned residual network to classify human activities based on micro-Doppler spectrograms. The residual network (ResNet) is pre-trained on ImageNet and fine-tuned on an empirical non-parametric human model using Motion Capture Database. Compared with typical CNN from scratch, this ResNet-based method requires shorter epochs (within 50 epochs) and achieves higher accuracy (rise nearly 6% on the average classification accuracy) for micro-Doppler spectrograms classification. Apart from statistical evaluation, we implement guided backpropagation method and t-Distributed Stochastic Neighbor Embedding (t-SNE) technique to visualize the transfer learning of residual network using spectrograms.
{"title":"Transfer Learning for Human Activities Classification Using Micro-Doppler Spectrograms","authors":"Hao Du, Yuan He, T. Jin","doi":"10.1109/COMPEM.2018.8496654","DOIUrl":"https://doi.org/10.1109/COMPEM.2018.8496654","url":null,"abstract":"Human activities classification has drawn great attention due to its potential applications in security, surveillance and gesture-based interface. The movements of the human body and limbs result in unique micro-Doppler features which can be exploited for identification of human behavior. In this work, we propose a transfer-learned residual network to classify human activities based on micro-Doppler spectrograms. The residual network (ResNet) is pre-trained on ImageNet and fine-tuned on an empirical non-parametric human model using Motion Capture Database. Compared with typical CNN from scratch, this ResNet-based method requires shorter epochs (within 50 epochs) and achieves higher accuracy (rise nearly 6% on the average classification accuracy) for micro-Doppler spectrograms classification. Apart from statistical evaluation, we implement guided backpropagation method and t-Distributed Stochastic Neighbor Embedding (t-SNE) technique to visualize the transfer learning of residual network using spectrograms.","PeriodicalId":221352,"journal":{"name":"2018 IEEE International Conference on Computational Electromagnetics (ICCEM)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2018-03-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126623393","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 : 2018-03-26DOI: 10.1109/COMPEM.2018.8496610
Wei Li, S. Qu
A linear dipole array based on wideband electromagnetic bandgap (EBG) is designed in this paper. The linear array is organized along the H plane and the simulated results of an infinite array show that it features a bandwidth of 24.1% (from 5.1 to 6.5GHz) while scanning to 70° in the H plane with $text{VSWR < 2.5}$. Then, a ${1times 20}$ finite linear dipole array based on the above element configuration is designed. The array has the ability to scan up to 75° with a gain reduction less than 3 dB at 5GHz, and scan up to 80° with a gain reduction less than 3 dB at 5.7GHz, and scan up to 88° with a gain reduction about 4 dB at 6.1GHz.
{"title":"Design of Wideband and Wide-Angle Scanning Phased Array Based on Wideband EBG Structure","authors":"Wei Li, S. Qu","doi":"10.1109/COMPEM.2018.8496610","DOIUrl":"https://doi.org/10.1109/COMPEM.2018.8496610","url":null,"abstract":"A linear dipole array based on wideband electromagnetic bandgap (EBG) is designed in this paper. The linear array is organized along the H plane and the simulated results of an infinite array show that it features a bandwidth of 24.1% (from 5.1 to 6.5GHz) while scanning to 70° in the H plane with $text{VSWR < 2.5}$. Then, a ${1times 20}$ finite linear dipole array based on the above element configuration is designed. The array has the ability to scan up to 75° with a gain reduction less than 3 dB at 5GHz, and scan up to 80° with a gain reduction less than 3 dB at 5.7GHz, and scan up to 88° with a gain reduction about 4 dB at 6.1GHz.","PeriodicalId":221352,"journal":{"name":"2018 IEEE International Conference on Computational Electromagnetics (ICCEM)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2018-03-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130096388","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 : 2018-03-26DOI: 10.1109/COMPEM.2018.8496642
A. Patnaik, Wei Zhang, Runbing Hua, J. Fan, D. Pommerenke
A complex system such as a human assisting robot will have many sensors and use parallel processing for achieving the desired action. During a transient disturbance, such as an electrostatic discharge (ESD), one or many of these sensors can be disturbed. In this study, the detect/presence logic is toggled, consequently, the microcontroller reads the sensor status as disabled. This is a soft-failure which can be recovered by a power cycle of the system. Here a case study is investigated where a methodology is developed to help system designers to understand and model the cause of the sensor failure during an ESD event. This methodology will also help system designers to design efficient filters on the critical signal lines to minimize the effect of coupled noise.
{"title":"Root Cause Analysis Methodology of ESD Soft-Failure Applied to a Robot","authors":"A. Patnaik, Wei Zhang, Runbing Hua, J. Fan, D. Pommerenke","doi":"10.1109/COMPEM.2018.8496642","DOIUrl":"https://doi.org/10.1109/COMPEM.2018.8496642","url":null,"abstract":"A complex system such as a human assisting robot will have many sensors and use parallel processing for achieving the desired action. During a transient disturbance, such as an electrostatic discharge (ESD), one or many of these sensors can be disturbed. In this study, the detect/presence logic is toggled, consequently, the microcontroller reads the sensor status as disabled. This is a soft-failure which can be recovered by a power cycle of the system. Here a case study is investigated where a methodology is developed to help system designers to understand and model the cause of the sensor failure during an ESD event. This methodology will also help system designers to design efficient filters on the critical signal lines to minimize the effect of coupled noise.","PeriodicalId":221352,"journal":{"name":"2018 IEEE International Conference on Computational Electromagnetics (ICCEM)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2018-03-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133301115","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 : 2018-03-26DOI: 10.1109/COMPEM.2018.8496707
Wenna Fan, M. Zhang, Jinxing Li
This paper mainly focuses on the simulation analysis of squint synthetic aperture radar (SAR) images from complicated targets based on the electromagnetic (EM) scattering characteristics. For this purpose, the geometrical optics and physical optics (GO-PO) hybrid method is adopted to estimate the EM scattering characteristics from complicated targets in order to get the SAR echo data. When dealing with the obtained squint SAR echo, the back projection (BP) algorithm is utilized, of which the effectiveness is verified. On this basis, a simulation procedure for SAR images of complicate targets under squint mode is proposed. All the simulated results show the effectiveness of the simulation procedure.
{"title":"SAR Imagery Simulation from Complicated Targets Based on Back Projection Algorithm","authors":"Wenna Fan, M. Zhang, Jinxing Li","doi":"10.1109/COMPEM.2018.8496707","DOIUrl":"https://doi.org/10.1109/COMPEM.2018.8496707","url":null,"abstract":"This paper mainly focuses on the simulation analysis of squint synthetic aperture radar (SAR) images from complicated targets based on the electromagnetic (EM) scattering characteristics. For this purpose, the geometrical optics and physical optics (GO-PO) hybrid method is adopted to estimate the EM scattering characteristics from complicated targets in order to get the SAR echo data. When dealing with the obtained squint SAR echo, the back projection (BP) algorithm is utilized, of which the effectiveness is verified. On this basis, a simulation procedure for SAR images of complicate targets under squint mode is proposed. All the simulated results show the effectiveness of the simulation procedure.","PeriodicalId":221352,"journal":{"name":"2018 IEEE International Conference on Computational Electromagnetics (ICCEM)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2018-03-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131113510","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 : 2018-03-26DOI: 10.1109/COMPEM.2018.8496636
Tianhao Li, L. Guan, Z. Fan, Mengmeng Li, Rushan Chen
This paper introduces a wide conformal quasi printed Log Periodic dipole band-notched antenna. The conformal band-notched antenna is composed of three sections-the low frequency structure, the band-notched structure and the high frequency structure. Installing to the interior of a radome, radiation characteristics of the antenna are very good. The simulation results revealed that the antenna has good VSWR(less than 2) and radiation pattern of realized gain over the design frequnecy range of 3GHz to 6.2GHz and 8GHz to 12GHz.
{"title":"A Novel Band-Notched UWB Conformal Antenna Combined with the Method of Circuitry","authors":"Tianhao Li, L. Guan, Z. Fan, Mengmeng Li, Rushan Chen","doi":"10.1109/COMPEM.2018.8496636","DOIUrl":"https://doi.org/10.1109/COMPEM.2018.8496636","url":null,"abstract":"This paper introduces a wide conformal quasi printed Log Periodic dipole band-notched antenna. The conformal band-notched antenna is composed of three sections-the low frequency structure, the band-notched structure and the high frequency structure. Installing to the interior of a radome, radiation characteristics of the antenna are very good. The simulation results revealed that the antenna has good VSWR(less than 2) and radiation pattern of realized gain over the design frequnecy range of 3GHz to 6.2GHz and 8GHz to 12GHz.","PeriodicalId":221352,"journal":{"name":"2018 IEEE International Conference on Computational Electromagnetics (ICCEM)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2018-03-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133916489","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 : 2018-03-01DOI: 10.1109/COMPEM.2018.8496506
Jianbing Li, V. Chandrasekar, XueSong Wang
For a flow involved with a number of particles of different sizes, it is difficult to retrieve the flow's characteristic parameters from the particles' Doppler velocity alone because the correspondences between particle sizes and Doppler components are difficult to discriminate. A feasible retrieval approach is to find out the dominant-scattering particles (making the most prominent contribution to the maximum Doppler component) and solve the motion equation of these particles with their Doppler velocity. In this approach, the first key issue is to determine the Dominant-Scattering particles' size. Taking rainfall as the physical medium, this paper proposes the relationships between the Dominant-Scattering raindrops' size $(D_{s})$ and the polarimetric variables (circular depolarization ration $C_{dr}$ or differential reflectivity $Z_{dr}$) through a series of scattering simulations. Results show that $C_{dr}$ is a good polarimetric variable to discriminate the Dominant-Scattering raindrop size, and the relationship $D_{s}-C_{dr}$ for different radar observation angles can be fitted to very simple polynomials.
{"title":"Determination of Dominant-Scattering Raindrops' Size Using Dual-Polarization Radar Observations","authors":"Jianbing Li, V. Chandrasekar, XueSong Wang","doi":"10.1109/COMPEM.2018.8496506","DOIUrl":"https://doi.org/10.1109/COMPEM.2018.8496506","url":null,"abstract":"For a flow involved with a number of particles of different sizes, it is difficult to retrieve the flow's characteristic parameters from the particles' Doppler velocity alone because the correspondences between particle sizes and Doppler components are difficult to discriminate. A feasible retrieval approach is to find out the dominant-scattering particles (making the most prominent contribution to the maximum Doppler component) and solve the motion equation of these particles with their Doppler velocity. In this approach, the first key issue is to determine the Dominant-Scattering particles' size. Taking rainfall as the physical medium, this paper proposes the relationships between the Dominant-Scattering raindrops' size $(D_{s})$ and the polarimetric variables (circular depolarization ration $C_{dr}$ or differential reflectivity $Z_{dr}$) through a series of scattering simulations. Results show that $C_{dr}$ is a good polarimetric variable to discriminate the Dominant-Scattering raindrop size, and the relationship $D_{s}-C_{dr}$ for different radar observation angles can be fitted to very simple polynomials.","PeriodicalId":221352,"journal":{"name":"2018 IEEE International Conference on Computational Electromagnetics (ICCEM)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2018-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115444001","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 : 2018-03-01DOI: 10.1109/COMPEM.2018.8496489
Q. Du, Xuelei Fu, Sun-A Song, Mo Li
A nanohole array based bio-sensor has been achieved numerically. The nanohole array is arranged in a hexagonal lattice and the material of the metal film is aluminum. With optimized parameters, the figure of merit of the presented nanohole array based sensor is around 169 which has surpassed the reported figure of merit of nanohole array based sensors.
{"title":"Hybrid Mode Based Ultrasensitive Nanohole Array Bio-Sensor","authors":"Q. Du, Xuelei Fu, Sun-A Song, Mo Li","doi":"10.1109/COMPEM.2018.8496489","DOIUrl":"https://doi.org/10.1109/COMPEM.2018.8496489","url":null,"abstract":"A nanohole array based bio-sensor has been achieved numerically. The nanohole array is arranged in a hexagonal lattice and the material of the metal film is aluminum. With optimized parameters, the figure of merit of the presented nanohole array based sensor is around 169 which has surpassed the reported figure of merit of nanohole array based sensors.","PeriodicalId":221352,"journal":{"name":"2018 IEEE International Conference on Computational Electromagnetics (ICCEM)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2018-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115676519","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 : 2018-03-01DOI: 10.1109/COMPEM.2018.8496555
Xuanying Hou, Na Liu, Q. Liu
To accelerate the simulation of geological fractures by numerical method, a mixed finite element method (FEM) with surface current boundary condition (SCBC) is proposed. It is the combination of the transverse components of vectorial Helmholtz equation and the Gauss' law in mixed FEM that makes the system matrix good conditioned so as to restrain the low-frequency breakdown phenomenon. In addition, the fracture, through which the electric field distribution is modeled by SCBC, is represented geometrically by a one-dimensional line instead of a thin layer. The method effectively reduces the degrees of freedom and save CPU time, avoiding the local mesh refinement. We demonstrate this method by computing the secondary electric field scattered by a thin layer of hydrocarbon.
{"title":"The Mixed Finite Element Method with Surface Current Boundary Condition for Modeling Geological Fracture","authors":"Xuanying Hou, Na Liu, Q. Liu","doi":"10.1109/COMPEM.2018.8496555","DOIUrl":"https://doi.org/10.1109/COMPEM.2018.8496555","url":null,"abstract":"To accelerate the simulation of geological fractures by numerical method, a mixed finite element method (FEM) with surface current boundary condition (SCBC) is proposed. It is the combination of the transverse components of vectorial Helmholtz equation and the Gauss' law in mixed FEM that makes the system matrix good conditioned so as to restrain the low-frequency breakdown phenomenon. In addition, the fracture, through which the electric field distribution is modeled by SCBC, is represented geometrically by a one-dimensional line instead of a thin layer. The method effectively reduces the degrees of freedom and save CPU time, avoiding the local mesh refinement. We demonstrate this method by computing the secondary electric field scattered by a thin layer of hydrocarbon.","PeriodicalId":221352,"journal":{"name":"2018 IEEE International Conference on Computational Electromagnetics (ICCEM)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2018-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122134040","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：