Title :
A new recursive partitioned estimation algorithm derived from extended least squares technique
Author :
Amaral, W.C. ; Gimeno, L. ; Mendes, M.
Author_Institution :
UNICAMP, Campinas, SP, Brasil
Abstract :
The generalization of the Extended Least Squares algorithm for parameter estimation in MIMO systems with ARMA for the noise allows a parallel estimation of row parameters in the model. From this result a new algorithm is proposed which estimates process and noise parameters separately resulting in a further reduction in computer time.
Keywords :
Cost function; Difference equations; Least squares approximation; MIMO; Noise reduction; Parallel processing; Parameter estimation; Partitioning algorithms; Polynomials; Recursive estimation;
Conference_Titel :
Decision and Control including the Symposium on Adaptive Processes, 1981 20th IEEE Conference on
Conference_Location :
San Diego, CA, USA
DOI :
10.1109/CDC.1981.269282
