基于贝叶斯算法的天线端到端优化  

End-to-end antenna optimization based on Bayesian optimization algorithm

作　　者：田春明[1] 杨安 叶乐[2] 李建星 贺雨晨 TIAN Chunming;YANG An;YE Le;LI Jianxing;HE Yuchen(School of Information and Communications Engineering,Xi’an Jiaotong University,Xi’an 710049,China;Institute of Microelectronics,Peking University,Beijing 100871,China;School of Electronic Science and Engineering,Xi’an Jiaotong University,Xi’an 710049,China)

机构地区：[1]西安交通大学信息与通信工程学院,陕西西安710049 [2]北京大学微纳电子学系,北京100871 [3]西安交通大学电子科学与工程学院,陕西西安710049

出　　处：《系统工程与电子技术》2021年第12期3413-3419,共7页Systems Engineering and Electronics

基　　金：国家重点研发计划(2019YFB1803205);陕西省重点研发计划(2019GY-007);强脉冲辐射环境模拟与效应国家重点实验室专项经费(SKLIPR1815)资助课题。

摘　　要：在确定天线的拓扑结构以后,通常需要对天线的结构参数开展反复的优化才能达到设计目标,快速有效的优化算法有利于缩短天线的设计周期。在建立综合目标函数的基础上,同时考虑天线的多个优化目标和限制条件,使用贝叶斯优化算法对天线进行端到端优化。基于在线更新的数据集,高斯过程估计出目标函数的后验分布,进而使用获得函数进行迭代。通过两种天线模型对提出的优化算法进行仿真验证,结果表明,由于建立了天线参数到综合目标函数的映射关系,整个优化过程以端到端的方式实现,与传统的优化方法相比,所提算法的优化结果和优化速度都具有明显的优势。Once the topology of an antenna is determined,optimization of various antenna parameters is conducted to achieve the desired objectives.A fast and effective optimization algorithm is significant to shorten the antenna design period.With the use of a comprehensive objective function which fuses multiple targets and constraints into one single function,the antenna is optimized end-to-end by employing Bayesian optimization algorithm.Based on the online updated dataset,the Gaussian process estimates the posterior distribution of the comprehensive objective function,and afterwards acquisition function is utilized for the iteration.To demonstrate the proposed method,two antenna models are simulated and optimized.The results show that because the proposed method is implemented in an end-to-end way through a mapping from the antenna parameters to a comprehensive objective function,it can achieve better optimization results and higher optimization efficiency compared with the conventional methods.

关 键 词：天线优化 多目标优化 端到端 贝叶斯优化算法 

分 类 号：TN82[电子电信—信息与通信工程]