Coded antenna radiation pattern prediction network based on DDA algorithm

In the past decades, the interpretability of neural network has gradually become a research hotspot in the fields of deep learning, and the method of combining physical model with neural network is one of its research directions. In this paper, a coded antenna radiation pattern prediction network based on DDA algorithm(ARPN-DA) has been purposed, which replaces part of the approximate calculation in DDA with neural network. Then, it realizes the prediction of antenna pattern under the condition of small samples and few neural network parameters. Compared with pure DNNs, the new system does not rely on massive datasets and has inter-pretabilitiy. We have implemented the network training and testing on the actual coded antenna platform, and the predictions of antenna pattern are very close to the measured results.