AMS Circuit Design Optimization Technique Based on ANN Regression Model With VAE Structure

The advanced design of an analog mixed-signal circuit is not simple enough to meet the requirements performance matrix as well robust operations under process-voltage-temperature (PVT) changes. Even commercial products demand stringent specifications while maintaining system’s performance. main objectives this study are increase efficiency optimization process by configuring in multiple regression modeling stages, characterize our target into a model including PVT variations, and enable search for co- optimum points simultaneously checking sensitivity. We used artificial neural network (ANN) develop divided ANN coarse fine simulation steps. In addition, we applied variational autoencoder (VAE) structure reduce training error due insufficient input sample. According proposed algorithm, AMS designer can quickly point, which results best performance, least sensitive operation uses instead launching heavy SPICE simulations. study, voltage-controlled oscillator (VCO) selected prove algorithm. Under various conditions (CMOS 180 nm, 65 45 nm processes), proceed with flow obtain score that be evaluated figure-of-merit (FoM). As result, model-based achieves twice accurate comparison conventional single-step flow.