Robust FM demodulation of discrete-time signals using least squares differential ratio

A technique for the robust demodulation of discrete FM signals in the presence of additive noise is presented. Based upon least squares, the proposed technique is shown to improve the accuracy of instantaneous frequency estimation by 17 to 79% compared with the popular discrete energy separation (DESA), smooth DESA and Hilbert transform-based algorithms, e.g. noisy signals at low SNRs.