We can continue this process on the N 1 N 1 submatrix on the lower right We

# We can continue this process on the n 1 n 1 submatrix

• Notes
• 31

This preview shows page 24 - 27 out of 31 pages.

We can continue this process on the ( N - 1) × ( N - 1) submatrix on the lower right. We take x 2 = 0 0 ˜ a 32 ˜ a 42 . . . ˜ a N 2 , y 2 = 0 0 r 2 0 . . . 0 , with r 2 = k x 2 k 2 , and H 2 = I - 2 u 2 u T 2 , u 2 = x 2 - y 2 k x 2 - y 2 k 2 . Then A 3 = H 2 A 2 H 2 = a 11 r 1 0 0 · · · 0 r 1 ˜ a 22 r 2 0 · · · 0 0 r 2 b a 33 b a 34 · · · b a 3 N 0 0 b a 43 b a 44 · · · b a 4 N . . . . . . . . . . . . . . . . . . 0 0 b a N 3 b a N 4 · · · b a NN Repeating this procedure N - 1 times gives us a tri-diagonal matrix T = A N = H N H N - 1 · · · H 1 AH 1 H 2 · · · H N = Q T AQ . where Q = H 1 H 2 · · · H N . Note that since all of the H n are orthonormal, so is Q . 23 Georgia Tech ECE 6250 Fall 2019; Notes by J. Romberg and M. Davenport. Last updated 13:22, November 13, 2019

Subscribe to view the full document.

Technical Details: Solving Toeplitz systems and the Levinson-Durbin algorithm We start by looking at how to solve Hv = y for a very particular right-hand side. Consider h 0 h 1 · · · h N - 1 h 1 h 0 h 1 · · · h N - 2 h 2 h 1 h 0 · · · h N - 3 . . . h N - 1 · · · · · · h 0 v [1] v [2] . . . . . . v [ N ] = h 1 h 2 . . . h N - 1 h N . (7) Here the first N - 1 entries of the “observation” vector y match the last N - 1 entries in the first column of H . This system is special- ized, but not at all contrived — ( 7 ) are called the Yule-Walker equations , and appear in many different places in statistical signal processing. Moreover, solving systems with general right-hand sides solve systems of the form ( 7 ) as an intermediate step. The crux of the Levinson-Durbin algorithm relies upon a seemingly innocuous fact. Let J be the N × N exchange matrix (also called the counter identity ): J = 0 0 · · · 0 0 1 0 0 · · · 0 1 0 0 0 · · · 1 0 0 . . . . . . 1 0 · · · 0 . Applying J to a vector x reverses the entries in x . For example, 0 0 1 0 1 0 1 0 0 5 - 1 7 = 7 - 1 5 . 24 Georgia Tech ECE 6250 Fall 2019; Notes by J. Romberg and M. Davenport. Last updated 13:22, November 13, 2019
It should be clear that J T = J and J 2 = JJ = I . Applying J to the left of a matrix reverses all of its columns, while applying J to the right reverses the rows. In particular, if H is a symmetric Toeplitz matrix, then J H = 0 0 · · · 0 0 1 0 0 · · · 0 1 0 0 0 · · · 1 0 0 . . . . . . 1 0 · · · 0 h 0 h 1 · · · h N - 1 h 1 h 0 h 1 · · · h N - 2 h 2 h 1 h 0 · · · h N - 3 . . . h N - 1 · · · · · · h 0 = h N - 1 h N - 2 · · · h 0 h N - 2 h N - 3 h N - 4 · · · h 1 h N - 3 h N - 4 h N - 5 · · · h 2 .

Subscribe to view the full document.

• Fall '08
• Staff

### What students are saying

• As a current student on this bumpy collegiate pathway, I stumbled upon Course Hero, where I can find study resources for nearly all my courses, get online help from tutors 24/7, and even share my old projects, papers, and lecture notes with other students.

Kiran Temple University Fox School of Business ‘17, Course Hero Intern

• I cannot even describe how much Course Hero helped me this summer. It’s truly become something I can always rely on and help me. In the end, I was not only able to survive summer classes, but I was able to thrive thanks to Course Hero.

Dana University of Pennsylvania ‘17, Course Hero Intern

• The ability to access any university’s resources through Course Hero proved invaluable in my case. I was behind on Tulane coursework and actually used UCLA’s materials to help me move forward and get everything together on time.

Jill Tulane University ‘16, Course Hero Intern