MidTerm2005

In addition we have the singular value decomposition

This preview shows page 1. Sign up to view the full content.

This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: full rank. In addition, we have the singular value decomposition of B given as: B = U 11V H 1 a* What is the dimension of the nullspace of B ? 7 8 b* Show that by using singular value decomposition from 8, the cost function kBwk2 2 given in the optimization problem from 7 can be expressed as vH Dv, where v 2 C N is a unit norm vector and D 2 C N N is a diagonal square matrix. Express v and D in terms of U 1, 1 and V 1. Name: Matr.-Nr.: Let us denote in the following the singular values of B as s1 s2    sN . c Show that vHDv  s2 N 8 9 Hint: aH Tc = Pi aTi ci , where T is a diagonal matrix and ai is the ith element of the i vector a, Ti is the ith element along the diagonal of the matrix T , and ci is the ith element of the vector c. Name: Matr.-Nr.: d* For which vector v is the expression given in 9 an equality? e* What is the solution w given in terms of V 1 to the optimization problem given in 7? 9...
View Full Document

• Winter '12
• TUM
• Singular value decomposition, mean square error, optimization problem, Array Signal Processing, Josef A. Nossek, Munich University of Technology

{[ snackBarMessage ]}

Ask a homework question - tutors are online