A Novel Wide-Band Layout-level Synthesis Methodology for CMOS Integrated Spiral Inductors with Artificial Neural Network

In this paper, we develop a fast and accurate synthesis way to effectively generate CMOS spiral inductor´s layout parameter using artificial neural network methodology. Under our synthesis methodology, we can simulate and synthesis the geometric parameters of spiral inductor´s layout up to the frequency of 20 GHz. Experiments and Results based on octagon differential spiral inductors of TSMC 0.13um MM/RF process indicate that this novel methodology can calculate layout parameters fast and accurately and therefore is of great help for RF circuit designers.