In this paper, a novel algorithm for computing the minimum eigenvalue and associated eigenvector is presented. This algorithm is a derivative of the inverse iteration method which requires a linear system to be solved in each iteration. Taking advantage of the Toeplitz structure, a pipelined Toeplitz system solver can be applied to solve the Toeplitz system in O(N) time units with O(N) processors. This brings the total computing time for solving the Toeplitz minimum eigenvalue down to O(KN) time units where K is the number of iterations. This algorithm makes it possible to perform real time processing of the Pisarenko's high resolution spectrum estimate.
|Original language||English (US)|
|Number of pages||8|
|Journal||IEEE Transactions on Acoustics, Speech, and Signal Processing|
|State||Published - Jan 1 1985|
All Science Journal Classification (ASJC) codes
- Signal Processing