In this article, through the optimized arrangement of the magnetized plasma, dielectric, and nonlinear Kerr dielectric, a tunable reflection phase retarder with a certain bandwidth is investigated. When the electric field of the incident electromagnetic wave is parallel to the xoy plane and has an angle of 45°with the x-axis, the electric field can be further decomposed into E x and E y along the xand y-axes, indicating TM and TE waves, respectively. The results show that under reasonable parameters, the 180°or −180°phase delay of the reflected wave of the proposed structure is obtained and adjusted. In particular, the nonlinear bistable response at this performance is also demonstrated. Furthermore, to verify the superiority of the obtained bandwidth of the phase retarder, the influences of the incident angle, the plasma frequency, the external magnetic field, and the nonlinear light intensity on the bandwidth of the phase retarder have been studied. It is observed that the total phase retarder bandwidth can reach 2.6 GHz by appropriately optimizing those parameters. Additionally, this work provides constructive suggestions for tunable broadband phase blockers and half-wave plates, and it still holds some potential significance in the radome.

Convolutional neural networks (CNN) have a strong feature extraction ability for images and present a high level of efficiency and accuracy in object detection and image recognition. When CNN is used to model microwave devices, the existing literature generally uses its size parameters as one‐dimensional (1‐D) input, which does not give full play to the image‐processing ability of CNN. In order to make full use of the characteristics of CNN, this letter converts the 1‐D input of microwave devices into the form of an image model, that is, the 1‐D input is transformed into a two‐dimensional (2‐D) matrix composed of 0 and 1 as the input. The image model is combined with CNN, called image‐based CNN (ICNN), which establishes a deep learning surrogate model between the physical parameters and electrical properties of microwave devices and improves the accuracy and generalization ability of the model. Taking the resonant frequency of the microstrip antenna as a simulation example, modelling was carried out by the proposed ICNN and compared with the mainstream machine learning methods. The results show that the proposed method has high convergence and fitting accuracy.

No abstract