Noise robust AM_FM demodulation using least- squares truncated power series approximation

This paper describes a parametric approach for demodulating multicomponent AM-FM sinusoidal signals. The parametric model consist of several truncated power series whose coefficients are estimated using linear least-square minimisation cast within an iterative parameter-substitution framework. It is shown that the proposed technique for AM-FM estimation is relatively robust to background noise compared to other techniques such as the smoothed energy operator separation algorithm and the Hilbert transform method.