3-D frequency estimation of multiple damped sinusoids using subspace approach

In this paper, a computationally efficient method that combines the subspace and principal singular vectors utilization for modal analysis (PUMA) is proposed for 3-D frequency estimation of multiple damped sinusoids. The perspective idea is the construction of 2-D sliced matrices from 3-D sinusoids, and then the 2-D sliced matrices are used to construct a set of covariance matrix to obtain the SVD of each covariance matrix. The available PUMA method is employed to estimate the frequencies and the damped factors in each dimension. Moreover, we can use existing matching algorithms to pair the 3-D frequency parameters. Only the original scale of sampled data is fully utilized, so the proposed method is more computationally attractive than the two existing methods.