Efficient maximum entropy spectral estimation using non-multiplication methods

New efficient autocorrelation and lattice methods of MEM spectral estimation are presented. In these methods, the "covariance property" of Gaussian distributions is exploited such that autocorrelation samples or lattice parameters can be computed with addition operations only. Also the performance of the proposed lattice method is studied through simulations and its application to the problem of spectral analysis of the vibrational modes of damaged and undamaged aircraft wings.