基于支持向量机集成的圆极化微带天线设计  

Design of circularly-polarized microstrip antenna by SVM ensemble

作　　者：田雨波[1] 孙菲艳 任作琳 

机构地区：[1]江苏科技大学电子信息学院,镇江212003

出　　处：《电波科学学报》2015年第6期1086-1092,共7页Chinese Journal of Radio Science

基　　金：国家自然科学基金(61401179)

摘　　要：人工神经网络(Artificial Neural Network,ANN)设计圆极化微带天线(Circularly-polarized Microstrip Antenna,CPMSA)时需要进行大量数据样本的准备,网络结构一般都比较复杂.为了解决这个问题,利用支持向量机(Support Vector Machine,SVM)在解决小样本数据处理问题时具有拟合精度高、泛化能力强、结构简单等优点,结合二进制粒子群(Binary Particle Swarm Optimization,BiPSO)算法选择出合适的SVM个体参与集成,形成一种基于BiPSO算法的选择性SVM集成(SVM Ensemble,SVME)方法,并将该方法用于单馈切角方形CPMSA的综合设计.仿真结果表明:这种SVME方法提高了算法的鲁棒性和有效性,有更好的预测精度,通过与ANN、SVM以及现有文献的预测结果对比可以看出,由该模型得出的结果优于此问题的已有结论.A large deal of calculating data and complex structure is needed when design circularly-polarized microstrip antenna (CPMSA) with artificial neural network. To solve this problem, a synthesis model based on SVM ensemble (SVME) is proposed for the design of single-feed CPSMA with truncated corners. This method was based on the features of high fitting precision, simple structure, and the strong generalization ability of the support vector machine (SVM). The basic idea of the method was to optimally select differential SVMs to construct SVME with the aid of binary particle swarm optimization (BiPSO) algorithm. The model is validated by comparing its results with artificial neural network and a single SVM. Experiments show that the method is effective. It may improve the generalization ability of SVM and reduce the prediction error, and this model is superior to the problem of existing conclusions. Copyright © 2015 by Editorial Department of Chinese Journal of Radio Science

关 键 词：圆极化 微带天线 切角 粒子群优化 支持向量机集成 

分 类 号：TN822[电子电信—信息与通信工程] TP181[自动化与计算机技术—控制理论与控制工程]