FACTORIZATION OF RATIONAL MATRICES FOR MULTIPLE DIMENSIONAL STOCHASTIC PROCESSES.
Abstract
This paper presents three algorithms for factoring rational spectral densities along with a summary of the multiple dimensional approach to least squares linear filtering. The first two algorithms are symmetric factorization algorithms based on common factors, the division algorithm, and algebraic number theory. Use of the first two algorithms produces a symmetric factorization in three factors which satisfied the sufficient conditions of optimum least squares linear filtering theory. This factorization is in contrast to the Youla's approach using two factors. The third algorithm is a symmetric simplification algorithm which closely corresponds to that devised by Oono and Yasuura and later by Davis, but which is without the strict requirements on the determinant of the matrix. This third algorithm is particularly useful since it requires no factorization of a polynomial. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Jul 20, 1965
- Accession Number
- AD0478549
Entities
People
- Charles S. Lorens
Organizations
- The Aerospace Corporation