Subspace-based spectrum estimation by reweighted and regularized nuclear norm minimization in frequency-domain


AKÇAY H., TÜRKAY S.

10th IEEE Conference on Industrial Electronics and Applications, Auckland, Yeni Zelanda, 15 - 17 Haziran 2015, ss.438-443 identifier identifier

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Doi Numarası: 10.1109/iciea.2015.7334153
  • Basıldığı Şehir: Auckland
  • Basıldığı Ülke: Yeni Zelanda
  • Sayfa Sayıları: ss.438-443
  • Anahtar Kelimeler: identification, subspace method, power spectrum, frequency-domain, reweighted nuclear norm, RANK MINIMIZATION, IDENTIFICATION
  • Anadolu Üniversitesi Adresli: Evet

Özet

In this paper, we study model order choice in subspace-based identification algorithms using nonuniformly spaced spectrum measurements. A critical step in these methods is splitting of two invariant subspaces associated with causal and non-causal eigenvalues of some structured matrices built from spectrum measurements via singular-value decomposition in order to determine model error. Mirror image symmetry with respect to the unit circle between the eigenvalue sets of the two invariant spaces required by the subspace algorithms is lost due to noise and insufficient amount of data. Recently, a robust model order selection scheme based on the regularized nuclear norm optimization in combination with a subspacebased spectrum estimation algorithm was proposed. We propose a reweighted version of this scheme. A numerical example shows that the reweighted nuclear norm minimization makes model order selection easier and results in more accurate models compared to unweighted nuclear norm minimization, in particular at high signal-to-noise ratios.