This paper presents a new method for developing computer-aided design (CAD) models, using spectral domain (SD) formulation. Using the artificial neural network (ANN) technique, a combination of continuous function and delta functions constitutes the spectral domain Green's functions. We obtain closed-form formulas for integration involving these equations. Another neural network relates different antenna parameters. Utilizing the reverse modeling for patch dimension determination, it becomes useful as a CAD model for patch antenna design. Designs, of simple rectangular patch antennas, serve as illustration of this method.

Designing rectangular patch antenna using the neurospectral method

Active antenna which offers much potential as useful circuits for future applications are a newly developing area. Several methods for analysing active antennas are presented. The first method relies on the FDTD technique. In this approach the time domain model of active devices is introduced into the FDTD frame. The large signal characteristics of active devices and the behaviour of antennas are both taken into account. This is considered to be the most rigorous method. This paper presents a simple method for analysing active antennas. The method is based on the piecewise harmonic balance technique. A neural network is used to speed up the optimization by replacing repeated circuit simulation. Therefore, the new approach is much faster than the traditional optimization technique. Compared with the data obtained from FDTD, the results also agree well.

Xiangdong Wang

Papers

A combination of piecewise harmonic balance technique and neural network for analyzing active antenna