A Newton-based Ziskind-Wax alternating projection algorithm applied to spectral estimation

摘要

A Newton-based alternating projection (AP) algorithm for estimating the parameters of exponential signals in noise is presented. Its performance is compared with that of modified forward-backward linear prediction/backward linear prediction (MFBLP/BLP) total least squares (TLS), and iterative quadratic maximum likelihood (IQML) methods using computer simulations. In the case of undamped sinusoids, the AP algorithm yielded a lower threshold SNR than the others. In the case of damped sinusoids, its threshold SNR was the same as that of TLS but was 1 dB lower than that of MFBLP and IQML.