On a Rational Model Interpolation Technique of Ultra-Wideband Signals
Nicolas H. Younan, C.D. Taylor, Jianping Gu · Kluwer Academic Publishers eBooks · 2005
A regressive rational function interpolation technique for noisy frequency domain data is presented. This technique is based on using a modified ARMA model to fit the transfer function of the underlying noisy spectral data. The coefficients of the model are then calculated based on minimizing a weighted least square criterion iteratively. This procedure incorporates the singular value decomposition technique to accurately estimate the order of the model and at the same time to suppress the noise components, which corrupt the actual data samples. Simulation results are presented to demonstrate the applicability of the interpolation technique.