Residual Network Based Direct Synthesis of EM Structures: A Study on One-to-One Transformers

TL;DR

This paper introduces a neural network approach for directly synthesizing on-chip EM passive structures, specifically 1:1 transformers, to facilitate rapid RF/mm-Wave circuit design and optimization.

Contribution

It presents a novel machine learning method for direct EM structure synthesis, demonstrated on 45nm SOI transformers, enabling automated design predictions from s-parameter data.

Findings

Successful neural network prediction of transformer geometries

Demonstrated rapid synthesis process for RF/mm-Wave structures

Potential for automated EM design workflows

Abstract

We propose using machine learning models for the direct synthesis of on-chip electromagnetic (EM) passive structures to enable rapid or even automated designs and optimizations of RF/mm-Wave circuits. As a proof of concept, we demonstrate the direct synthesis of a 1:1 transformer on a 45nm SOI process using our proposed neural network model. Using pre-existing transformer s-parameter files and their geometric design training samples, the model predicts target geometric designs.