Estimation of electromagnetic far-field from near-field using machine learning

被引：0

作者：

Takizawa, Kohei ^{[1
]}

Watanabe, Yuta ^{[1
]}

Fujiwara, Kohei ^{[1
]}

机构：

[1] Tokyo Metropolitan Ind Technol Res Inst, Tokyo, Japan

来源：

2020 INTERNATIONAL SYMPOSIUM ON ANTENNAS AND PROPAGATION (ISAP) | 2021年

关键词：

EMC; Machine learning; Near-field; Far-field;

D O I：

暂无

中图分类号：

TM [电工技术]; TN [电子技术、通信技术];

学科分类号：

0808 ; 0809 ;

摘要：

In this development, we proposed a method to use machine learning to estimate a far-field from a near-field. A signal source and wires randomly arranged on the substrate are used as a radiation source. The near and far-fields emitted from the substrate are calculated by electromagnetic field simulation to produce the data necessary for machine learning. We used XGBoost for machine learning. The Route Mean Square Percentage Error (RMSPE) of the far-field strength of the estimated and test data was calculated and was less than 7.1 % at 1-6 GHz.

引用

页码：119 / 120

页数：2