State-of-the-Art: AI-Assisted Surrogate Modeling and Optimization for Microwave Filters

Abstract

Microwave filters are indispensable passive devices for modern wireless communication systems. Nowadays, electromagnetic (EM) simulation-based design process is a norm for filter designs. Many EM-based design methodologies for microwave filter design have emerged in recent years to achieve efficiency, automation, and customizability. The majority of EM-based design methods exploit low-cost models (i.e., surrogates) in various forms and artificial intelligence techniques assist the surrogate modeling and optimization processes. Focusing on surrogate assisted microwave filters designs, this paper firstly analyzes the characteristic of filter design based on different design objective functions. Then, the state-of-the-art filter design methodologies are reviewed, including surrogate modeling (machine learning) methods and advanced optimization algorithms. Three essential techniques in filter designs are included: 1) Smart data sampling techniques; 2) Advanced surrogate modeling techniques. 3) Advanced optimization methods and frameworks. To achieve success and stability, they have to be tailored or combined together to achieve the specific characteristics of the microwave filters. Finally, new emerging design applications and future trends in filter design are discussed.

An Early History of Optimization Technology for Automated Design of Microwave Circuits

“…Now with Reykjavik University, Iceland, he is surely the most identifiable researcher in our community with respect to all aspects of surrogate methodologies and their applications. Koziel and his team spearheaded innovative electromagnetic simulation-driven and surrogate-based optimization procedures for microwave circuits and antenna design, including variable-fidelity optimization frameworks [209], [210], [211]; surrogate-assisted tuning [191], [212], [213], yield estimation [214], [215], and multi-objective optimization [216], [217], [218], [219], [220]; methodologies for rapid re-design by inverse surrogates [221], [222]; microwave component miniaturization [223], [224]; dimensionally reduced and domain confined surrogates [225], [226], [227]; response feature-based nominal design [228], [229], [230], [231], [232], yield optimization [233], [234], [235], [236], [237], and robust design by tolerance maximization [238], [239], [240].…”

A Surrogate Modeling Space Definition Method for Efficient Filter Yield Optimization

“…A relatively large surrogate modeling space may work for some EM devices (e.g., antennas [20], [21]), but not for filters. As said above, filter responses often degrade significantly even when a small deviation is added to an optimal design [22]. Therefore, a slightly large surrogate modeling space may cause the MC samples satisfying the design requirements to become very few (e.g., < 1% of the MC samples according to our pilot experiments), leading to insufficient information for surrogate modeling of the design subspaces meeting the requirements.…”