About

     AI-assisted High Throughput Exploration of Novel Dielectrics for Beyond 5G and 6G Wireless Telecommunication and Automotive Radar  

 

The three-year proposal is comprised of Subproject I (Prediction and Reverse Technology of 6G and 5G+ Materials and Devices by Machine Learning; PI: Profs. Ting and Su), Subproject II (High Throughput Fabrication and Characterization of Porous Inorganic-organic Composite Dielectric Films; PI: Profs. Ruan and Chang), and Subproject III (Development of Three-dimensional Metamaterial-based Antenna Design for 6G Applications; PI: Prof. Chen). Our team also includes 智易科技 (HsinChu City, Taiwan) and Drs. Chikyow and Nagata at National Institute for Materials and Science (NIMS) in Japan. The five PIs will dedicate themselves to studying AI-assisted modelling and calculations, high throughput processing of porous inorganic-organic composite films, and maneuvering of metamaterials structures. A leading company智易科技in telecommunication technology will provide measurement facilities up to 90 GHz and their device under test layout for our resulting films. Drs. Chikyow and Nagata will employ high throughput related characterizations and dielectric measurement up to 45 GHz.

Our idea is constructed as follows. Machine learning through an artificial neural networks with genetic algorithms will be employed for the prediction and reverse technology of 5G+ and 6G materials in Subproject I. Subprojects II and III will first provide characteristics of potential Mg₂SiO₄-based materials and basic information of geometrical and structural designs for metamaterials, respectively, for the machine learning and training. The calculated results will be used for tailoring manufacturing parameters of high throughput experiments and optimizing metamaterials-based architecture design in Subprojects II and III, respectively. The resulting experimental data will then be fed back to Subproject I for model training. The machine learning loop will enable the algorithm to make scientific prediction of novel materials properties. Rapid-screening characterizations will then be employed to justify the performance in Subproject I. The materials developed in Subprojects I and II will be integrated into desirable structural patterns (metamaterials) obtained in Subproject III to enhance dielectric properties for beyond 5G and 6G wireless telecommunication and automotive radar applications. Moreover, Subproject I will also provide possible solutions through the reverse technology of the algorithm to accelerate the exploration process. Furthermore, various physical properties of high throughput resulting samples will be rapidly characterized/mapped at NIMS.

Our goal is to establish a multidisciplinary and competent team to effectively integrate resources from academia, local companies, and international research institutes and maximize our research capability through intimate interorganizational collaboration to pursue novel and reliable low dielectric constant and loss dielectrics for the realization of beyond 5G and 6G wireless telecommunication devices and automotive radar. Resulting achievements will pave a way to the pursuit of international intellectual property.