This paper introduces the modern spectral estimation method in to cross-spectral estimation
the cross-correlation-based Yule-Walker equation is proposed. Based on this contribution
the Moment estimation method and Levinsion recursive algorithm are developed. The new algorithms can overcome efficiently the shortcomings of low resolution and high variance which are existing in both FFT and Cross-periodogram methods. Finally
we compare the proposed algorithm with the well-known methods
Pisarenko and SVD methods
by using the parameter estimation results of sinusoids in strong measurement noise (SNR = -30dB) .