Separation and identification of a transient signal using the characteristics of eigenvectors
James P. Larue, George B. Smith, George E. Ioup, Juliette W. Ioup · The Journal of the Acoustical Society of America · 2004
This is an extension of previous research [Larue et al., J. Acoust. Soc. Am. 113, 2212 (2003)] into the physical characteristics possessed by eigenvectors in a singular value decomposition (SVD) of a covariance matrix formed from a sinusoidal signal corrupted by multipath to which Gaussian noise is added. An animation of the SVD in progress will focus on the creation of the eigenvectors rather than the singular values. In this case, the SVD forms a three-part subspace decomposition corresponding to the three components of the signal. These subspaces are clearly determined from characteristics (Fourier transform and the Kaiser Varimax norm) associated with the eigenvectors and are not so easily determined from the singular values alone. [Research supported by the NRC-AFRL/IFEC and ONR.]