Rational Approximation of RF Passive Devices for Noisy Data

Abstract

In this article, we present Vector Fitting Algorithm (VFA) to derive the rational function approximation model of Bandpass and Baseband systems in presence of Noise. The Root Mean Square Error (RMSE) and Identification error $(\mathbf{C}\!\!\!\!{\scriptstyle{{}^-}} \mathbf{p})$ is calculated to analyze the modeling accuracy of VFA, for different orders of the system and for different SNR (signal to noise ratio) criteria. VFA is found to be more noise resilient with regard to system identification for higher order systems compared to lower order systems due to sharper frequency cut-off characteristics and corresponding noise suppression for higher order systems. It was also found that, for the same order and SNR, baseband systems have less modeling errors, when compared to bandpass systems.