大尺度电磁散射与逆散射问题的深度学习方法  

Deep learning approach to large-scale electromagnetic scattering and inverse scattering problems

作　　者：王龙刚[1] 钟威 阮恒心 贺凯 李廉林[1] WANG Longgang;ZHONG Wei;RUAN Hengxin;HE Kai;LI Lianlin(School of Electronics Engineering and Computer Sciences,Peking University,Beijing 100871,China)

机构地区：[1]北京大学信息科学技术学院,北京100871

出　　处：《电波科学学报》2018年第5期519-524,共6页Chinese Journal of Radio Science

摘　　要：大尺度电磁散射与逆散射一直是科学研究和工程应用的热点和难点,亟待发展将电磁模型与数据挖掘有机融合的高性能求解方法,针对此,提出了一种针对大尺度电磁散射与逆散射问题的深度学习模型.该模型不仅继承了深度神经网络结构简单、运算速度快等优点,而且还能高精度地解决大尺度电磁散射与逆散射问题.实验结果表明:文中提出的深度学习方法可为解决现有大尺度电磁测算融合和电磁逆散射的计算成本昂贵的难题提供新思路、开辟新方向.Large scale electromagnetic scattering and inverse scattering is a hot and difficult point in scientific research and engineering applications,and it is urgent to develop a high performance solution method to integrate electromagnetic model and data mining.A deep study model for large scale electromagnetic scattering and inverse scattering is proposed.The model not only inherits the advantages of simple structure and fast operation speed,but also can solve the problem of large scale electromagnetic scattering and inverse scattering with high precision.The experimental results show that the advanced learning method proposed in this paper can provide new ideas for solving the difficult problems of the existing large scale electromagnetic calculation fusion and the expensive calculation of electromagnetic inverse scattering.

关 键 词：大尺度 电磁散射 逆散射 深度学习 神经网络 

分 类 号：O451[理学—无线电物理]