Simulator guided generation improves em and rf circuit design quality
K-TRAIL: Simulator-Guided Generative Design of EM/RF Circuits
Machine Learning
Summary
Designing electromagnetic and radio frequency circuits is tricky because many different layouts can produce the same electrical behavior, and checking each design with detailed simulations takes a lot of computing time. The authors created K-TRAIL, a method that generates circuit layouts and uses feedback from simulations to improve these designs without needing complex math from the simulators. This approach helps find different designs that meet performance goals and speeds up the design process. Their tests show that this method makes generated layouts match desired circuit behaviors better and can discover new useful designs.
What this means in practice
- •For rf circuit designers: Automatically generate and refine multi-layer RF circuit layouts that meet specific electrical performance targets without manual trial and error.
- •For antenna engineers: Find alternative antenna layouts that satisfy given RF response constraints using automated simulation-guided design tools.
Authors
Piyush Saha, Evan Newell, Hanna O'Leary, Arun Natarajan, Alireza Aghasi
Abstract
Inverse design of RF and electromagnetic (EM) circuits is challenging because the relationship between circuit layout and electrical response is non-unique, and full-wave simulation is computationally expensive. This paper presents K-TRAIL, a simulator-guided generative framework for automated EM/RF circuit synthesis. K-TRAIL combines diffusion-based layout generation with derivative-free ensemble Kalman guidance, allowing feedback from a black-box EM simulator to refine candidate layouts during generation without requiring adjoint sensitivities or differentiable solver models. The framework supports both synthesis from prescribed S-parameter responses and synthesis directly from RF performance constraints. Experiments on multi-layer RFIC structures show that simulator-guided generation improves agreement with target responses and can identify structurally distinct layouts that satisfy circuit-level design requirements. The proposed approach provides a practical path toward generative, verification-aware RF circuit design while retaining the flexibility to explore diverse layout topologies.