MCMC-based iterative method for mixed spectrum estimation

From the Bayesian statistical inference theory, a new mixed spectrum estimation method, which is based on Markov chain Monte Carlo (MCMC) approach, is proposed in this paper. The proposed method iteratively extracts the estimates of sinusoid parameters via the traditional methods, and estimates the ones of AR clutter parameters via the MCMC approach. Because the MCMC approach can fully dig the inherent information among the samples, it is more suitable for estimating the AR parameters from the less samples. Consequently, the proposed method is an efficient method for extracting the parameters of mixed spectrum for the applications with a small number of samples. Simulation results show that, comparing with the similar structure method, the method presented in this paper behaves superior estimation performance for the less samples case, and the estimation performance is less influenced by the signal-to-clutter ratio.