QML-RANSAC: PPS and FM signals estimation in heavy noise environments.

The QML-RANSAC estimator is proposed. It combines the quasi-maximum likelihood (QML) estimator with the random sample consensus (RANSAC). This technique can with reasonable calculation complexity work for lower the signal-to-noise ratio (SNR) than existing parametric estimators of polynomial phase signals (PPS) and nonparametric estimators of FM signals, i.e., it achieves lower SNR threshold than the state-of-the-art techniques in the field. Obtained results are better for about 3dB with respect to the QML in term of the SNR threshold without increasing the mean squared error (MSE) above the threshold. The proposed estimator is tested on the PPS as a parametric estimator and for general FM signal estimation as a nonparametric estimator. An extension of the algorithm is proposed for multicomponent signals, as well. HighlightsThe QML-RANSAC is proposed for polynomial phase signals parameters estimation.Random sampling is employed in getting parameter estimates in each iteration..In each iteration obtained results are compared with current estimates using maximum likelihood inspired function.Obtained results are excellent surpassing current state-of-the-art techniques in the field.