Title :
An iterative method for the identification of MA systems
Author_Institution :
Lehrstuhl I fuer Math., RWTH Aachen, West Germany
fDate :
1/1/1988 12:00:00 AM
Abstract :
An iterative method for the identification of an nth-order moving average (MA) system is proposed which is based on a matrix splitting. Because of the special structures of the matrices involved, only O(n) storage is needed. One iteration is accomplished in O(n2) operations, and the algorithm may be executed on a pipeline multiprocessor. Few iterations are required to yield sufficient results, so the algorithm is fast. A simple error estimate can be obtained
Keywords :
computational complexity; error analysis; identification; iterative methods; matrix algebra; pipeline processing; spectral analysis; algorithm; error estimate; iterative method; matrix splitting; nth order moving average system; pipeline multiprocessor; storage complexity; Equations; Finite impulse response filter; Helium; Iterative algorithms; Iterative methods; Least squares methods; Pipelines; Symmetric matrices; Vectors; Velocity measurement;
Journal_Title :
Acoustics, Speech and Signal Processing, IEEE Transactions on
