Title :
Off- and online identification of discrete time LTI systems using separable least squares
Author :
El Edrissi, H. ; Verhaegen, M.H.G. ; Haverkamp, B.R.J. ; Chou, C.T.
Author_Institution :
Dept. of Electr. Eng., Delft Univ. of Technol., Netherlands
Abstract :
We demonstrate how the principle of the separable least squares technique of Golub and Perera (1973), can be exploited in off-line and online identification of LTI state space models. Though the approach outlined in this paper is applicable to both output-error and innovation state space models, for the sake of brevity the exposure is restricted to output-error models only. The chief advantage of using the principle of SLS lies in the fact that the separable optimization problem has a better conditioning compared with its non-separable counterpart. Furthermore, the speed of convergence may increase
Keywords :
discrete time systems; identification; linear systems; state-space methods; conditioning; convergence speed; discrete time LTI systems; innovation state space models; off-line identification; online identification; output-error state space models; separable least squares; state space models; Control systems; Convergence; Cost function; Laser sintering; Least squares methods; Parameter estimation; State-space methods; System identification; Technological innovation; USA Councils;
Conference_Titel :
Decision and Control, 1998. Proceedings of the 37th IEEE Conference on
Conference_Location :
Tampa, FL
Print_ISBN :
0-7803-4394-8
DOI :
10.1109/CDC.1998.760770
