摘要：神经网络是基于生物神经网络下的一种简化技术体现，在模拟生物神经网络的同时，基于实际问题来构建人工的神经网络。将人工神经网络模型应用于微带天线设计中,不仅可以提高该天线设计的效率，也可以使设计出的天线具有很高的精度。此课题采用神经网络对E型微带天线进行设计，针对其阻抗带宽的不同，设计出最适应的E型微带天线。通过IE3D软件的仿真计算，共得到144组天线样本。运用MATLAB建立针对E型微带天线的神经网络模型，给出结构参数和天线性能之间的非线性映射关系。其中，将144组数据中的134组数据来训练神经网络，剩下10组数据来测试该网络。结果证明符合设计要求，从而体现基于神经网络设计的E型微带天线具有可靠性。

关键词：E型微带天线；BP神经网络；IE3D天线仿真软件

Abstract：Neural network is based on a simplified technique of biological neural network. In the simulation of biological neural network, artificial neural network is constructed based on practical problems. The artificial neural network model is applied to the design of microstrip antenna, which not only can improve the efficiency of the antenna design, but also can make the design of the antenna with high accuracy. In this paper, the E-type microstrip antenna is designed by neural network, and the most suitable E-type microstrip antenna is designed according to the different impedance bandwidth. Through the simulation of IE3D software, a total of 144 antenna samples are obtained. The neural network model for E - type microstrip antenna is established by MATLAB, and the nonlinear mapping relation between structural parameters and antenna performance is given. 134 sets of data are used to train the neural network and 10 sets of data are used to test the network. The results show that the E-type microstrip antenna based on neural network design is reliable.

Keywords: E - type microstrip antenna; BP neural network; IE3D antenna simulation software

目录

第一章 绪论 1

1.1研究背景 1

1.2国内外研究现状 1

1.3研究方法描述 2

第二章 E型微带天线的设计 3

2.1E型微带天线的介绍 3

2.2IE3D软件设计 4

2.3频率计算 6

第三章 BP神经网络 8

3.1BP神经网络介绍 8

3.2神经元 8

3.3人工神经网络的特征 9

3.4BP神经网络的主要功能 9

3.5BP神经网络的优点和缺点 10

第四章 BP神经网络模型的构建 12

4.1MATLAB工具箱介绍 12

4.2构建神经网络 13

4.2.1BP网络创建函数 13

4.2.2神经元上的传递函数 13

4.2.3BP网络学习函数 14

4.2.4BP网络训练函数 15

4.3神经网络GUI设计 15

第五章 E型微带天线的神经网络设计