Finding antenna designs that satisfy frequency requirements and are also optimal with respect to multiple physical criteria is a critical component in designing next generation hardware. However, such a process is non-trivial because the objective function is typically highly nonlinear and sensitive to subtle design change. Moreover, the objective to be optimized often involves electromagnetic (EM) simulations, which is slow and expensive with commercial simulation software. In this work, we propose a sample-efficient and accurate surrogate model, named CZP (Constant Zeros Poles), to directly estimate the scattering coefficients in the frequency domain of a given 2D planar antenna design, without using a simulator. CZP achieves this by predicting the complex zeros and poles for the frequency response of scattering coefficients, which we have theoretically justified for any linear PDE, including Maxwell’s equations. Moreover, instead of using low-dimensional representations, CZP leverages a novel image-based representation for antenna topology inspired by the existing mesh-based EM simulation techniques, and attention-based neural network architectures. We demonstrate experimentally that CZP not only outperforms baselines in terms of test loss, but also is able to find 2D antenna designs verifiable by commercial software with only 40k training samples, when coupling with advanced sequential search techniques like reinforcement learning.

Designing modern antenna structures is a challenging endeavor. It is laborious and heavily reliant on engineering insight and experience, especially at the initial stages oriented towards the development of a suitable antenna architecture. Due to its interactive nature and hands-on procedures (mainly parametric studies) for validating the suitability of particular geometric setups, typical antenna development requires many weeks and significant involvement of a human expert. The same reasons only allow the designer to try out a very limited number of options in terms of antenna geometry arrangements. Automated topology development and dimension sizing is therefore of high interest, especially from an industry perspective where time-to-market and expert-related expenses are of paramount importance. This paper discusses a novel approach to unsupervised specification-driven design of planar antennas. The presented methodology capitalizes on a flexible and scalable antenna parameterization, which enables the realization of complex geometries while maintaining reasonably small parameter space dimensionality. A customized nature-inspired algorithm is employed to carry out space exploration and identification of a quasi-optimum antenna topology in a global sense. A fast gradient-based procedure is then incorporated to fine-tune antenna dimensions. The design framework works entirely in a black-box fashion with the only input being design specifications, and optional constraints, e.g., concerning the structure size. Numerous illustration case studies demonstrate the capability of the presented technique to generate unconventional antenna topologies of satisfactory performance using reasonable computational budgets, and with no human expert interaction necessary whatsoever.

To extend antenna designs on printed circuit boards for more engineers of interest, we propose a new simple method that models PCB antennas by a few basic components. By taking two separate steps to decide their geometrical properties, antenna prototypes can be facilitated with no experience required. Random sampling techniques are used in selecting high quality candidates. A novel image-based classifier is introduced to convolutional neural networks to further filtering out the premium topological structures. Examples from wearable products have been chosen to examine the entire workflow. Their final designs are realistic and their performance metrics are not inferior to the ones designed by experienced engineers.

Linear Partial Differential Equations (PDEs) govern the spatial-temporal dynamics of physical systems that are essential to building modern technology. When working with linear PDEs, designing a physical system for a specific outcome is difficult and costly due to slow and expensive explicit simulation of PDEs and the highly nonlinear relationship between a system's configuration and its behavior. In this work, we prove a parametric form that certain physical quantities in the Fourier domain must obey in linear PDEs, named the CZP (Constant-Zeros-Poles) framework. Applying CZP to antenna design, an industrial application using linear PDEs (i.e., Maxwell's equations), we derive a sample-efficient parametric surrogate model that directly predicts its scattering coefficients without explicit numerical PDE simulation. Combined with a novel image-based antenna representation and an attention-based neural network architecture, CZP outperforms baselines by 10% to 25% in terms of test loss and also is able to find 2D antenna designs verifiable by commercial software with $33\%$ greater success than baselines, when coupled with sequential search techniques like reinforcement learning.

The key to rapid antenna design is to improve the optimization of design parameters by simplifying the modeling of antenna geometric structures. We propose a flexible geometric modeling scheme that can be formed into any arbitrary shape by simply connecting nodes. We also propose a machine learning based generative method for high dimensional design parameter problems. It consists of discriminators and generators. The discriminators predict the performance of geometric models and the generators create new candidates that will pass the discrim-inators. Further, an evolutionary criterion approach is applied to update the discriminators and generators. The efficiency of the new design methodology is illustrated by a dual resonance antenna design example.

To facilitate the antenna design with the aid of computer, one of the practices in consumer electronic industry is to model and optimize antenna performances with a simplified antenna geometric scheme. The ease of handling multi-dimensional optimization problems and the less dependence on the engineers' knowledge and experience are the key to achieve the popularity of simulation-driven antenna design and optimization for the industry. In this paper, we introduce a flexible geometric scheme with the concept of mesh network that can form any arbitrary shape by connecting different nodes. For such problems with high dimensional parameters, we propose a machine learning based generative method to assist the searching of optimal solutions. It consists of discriminators and generators. The discriminators are used to predict the performance of geometric models, and the generators to create new candidates that will pass the discriminators. Moreover, an evolutionary criterion approach is proposed for further improving the efficiency of our method. Finally, not only optimal solutions can be found, but also the well trained generators can be used to automate future antenna design and optimization. For a dual resonance antenna design, our proposed method is better than the other mature algorithms.