Material-Centric Design of Advanced Electromagnetic Surfaces

先进的电磁表面（AES）使开发令人兴奋的新天线和无线通信组件成为可能，因为它们可以轻松实现对电磁辐射的非凡控制。它们对于实现定向和自适应天线是不可或缺的，而定向和自适应天线是5G等未来系统的基石。尽管AES衍生天线取得了进展，但在它们同时实现更大的带宽、可重新配置性、效率以及关键的整体设计简易性之前，它们的影响将无法完全实现。 拟议的研究计划提出了一种新的以材料为中心的设计方法，认为他们作为材料，可以使用材料信息技术设计的AES。目前，AES的设计围绕着临时的、基于试错的设计方法，设计时间长达数周或更长。所提出的方法利用计算的进步，使用机器学习（ML）来增强AES的设计。所提出的方法将建立一个大型数据库的AES散射体的几何形状，其特征在于有效的材料参数，允许技术从材料信息学有效地编目这些设计。ML增强设计方法将从这些数据中进行训练，并用于从给定的一组规格中快速准确地推导出所需的AES散射体几何形状，将设计时间缩短至数小时或更短。从长远来看，拟议的研究计划旨在从根本上改变AES的设计方式，并使用这种设计范式来发现新的AES功能。 在短期内，拟议的研究计划旨在设计和实施新的天线架构使用材料为中心的设计方法。三个短期目标是：1）设计在毫米波频率范围内操作的超宽带AES; 2）极大地提高AES衍生天线（诸如多波束卫星天线）的波束效率;以及3）设计使用所提出的方法的可重构AES，以便实现宽带、高效率、自适应天线。 该研究计划将对无线通信系统和AES领域产生重大影响。从以材料为中心的方法衍生的AES将实现新的性能和功能水平，从而将下一代无线通信系统的容量提高到每秒数千兆位的水平。从长远来看，ML增强设计方法将通过促进AES的快速合成和自动发现新AES材料的平台来彻底改变AES的设计方式。通过托管这些材料的公共，不断增长的数据库，我们还寻求通过采用开放的方法来推动设计的变化。这种基本的新设计范式将在未来几年迎来新一代的AES材料、发现和设计师。

Advanced electromagnetic surfaces (AESs) have enabled the development of exciting new antennas and components for wireless communications owing to the ease with which they achieve extraordinary control of electromagnetic radiation. They are indispensable for realizing directive and adaptive antennas, which are cornerstones of future systems such as 5G. Despite advances in AES-derived antennas, their impact will not be fully realized until they simultaneously achieve greater bandwidth, reconfigurability, efficiency, and - critically - overall ease of design. The proposed research program puts forth a new material-centric design methodology for AESs that views them as materials that can be designed using material-informatic techniques. Currently, the design of AESs revolves around ad hoc, trial-and-error based design methods with lengthy design times spanning weeks or longer. The proposed approach harnesses advances in computing to augment the design of AESs using machine learning (ML). The proposed approach will build a large database of AES scatterer geometries characterized by effective material parameters, allowing techniques from material informatics to efficiently catalogue these designs. An ML-augmented design method will be trained from this data and used to quickly and accurately derive required AES scatterer geometries from a given set of specifications, cutting design times to hours or less. In the long term, the proposed research program aims to fundamentally change the way AESs are designed, and use this design paradigm to discover new AES capabilities. In the short term, the proposed research program aims to devise and implement new antenna architectures using the material-centric design approach. Three short term objectives are to: 1) devise ultra-wideband AESs operating in the millimetre-wave frequency range; 2) greatly improve the beam efficiency of AES-derived antennas, such as multi-beam satellite antennas; and 3) devise reconfigurable AESs using the proposed approach in order to achieve wideband, high efficiency, adaptive antennas. The proposed research program will have major impact on wireless communication systems and the field of AESs. AESs derived from the material-centric approach will achieve new levels of performance and functionality, thereby increasing the capacity of next-generation wireless communication systems to multi-gigabit-per-second levels. In the long run, the ML-augmented design approach will revolutionize the way AESs are designed by facilitating both their rapid synthesis and a platform for automated discovery of new AES materials. By hosting a public, ever-growing database of these materials, we also seek to drive change in their design through embracing an open approach. This fundamental new design paradigm will usher in a new generation of AES materials, discoveries, and designers for years to come.