English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

In complex electromagnetic environment, the training data contains a lot of noise, which can make classic neural networks overfit easily. In order to solve this problem, a new method based on sampling theory is proposed for constructing models. Through analyzing wavelet networks and radiation fields in the frequency domain, it is proved that the new method can effectively reduce noise on models and help wavelet networks obtain a good generalization. In the simulation, the new method is compared with regularization techniques. The results show that the new method can prevent networks from overfitting and is more robust to noise. Introduction Radiation models are essential in testing and diagnosing antenna performances. For now, there are two methods to build an antenna radiation model: one is a classical physics method; the other is using neural network. The former method, generally, is based on the electromagnetic theory and can obtain a high accurate model[1][2]. However, this method is to great depends on the solutions of electromagnetic equations such as Maxwell equations, and it is still a worldwide problem on obtain the solution of Maxwell equations under complex boundary conditions. The latter method always construct models by the characteristics of training data, so that it can avoid solving the complex equations in former method[3][4]. However, in complex environment, the training data always contain a lot of noise and it will lead the network to overfitting. In order to find a way to solve the problems above, we analyze the frequency band of wavelet network and antenna radiation models, get the result that both of them have a limited frequency band, based on which a modeling method has been proposed. Simultaneously, the filtering algorithm based on sampling theory is applied in this wavelet network, to prevent the interference of the noise. Theoretical analysis proves the feasibility and effectiveness of this method. In the simulation, we compare this method with RBF network, the results show that the proposed method can reduce the influence of noise more effectively. Frequency Band Characteristics of Wavelet Network Wavelet network is a topological expansion of wavelet series. There are many kinds of structures of wavelet network, the structure proposed by literature [5] will be used in this paper, which is shown in Figure 1. The Figure 1has a mathematical expression ( ) (2 ) J ne k k f x c x k     (1) where{ } k c is the output weights, 2 is the input weights, ( ) x  is the scaling function, and k is the translation of the scale function.

Wavelet Network Modeling Method Based on Radiation Characteristics of Array Antenna