Controller Identification Scheme
The following scheme is used for identification of a reduced order controller based on closed-loop output error structure.
where R/S is the discrete-time model of the controller and B/A is the model of the plant supposed to be known. The model of the reduced order controller to be identified is Rr/Sr, where Rr and Sr are the polynomial of q-1as follows:
Rr(q-1) = r0+r1q-1+...+rnrq-nr
Sr(q-1) = 1+s1q-1+...+snsq-ns
r(t) is the reference signal (usually a Pseudo Random Binary Sequence) added to the input of the controller and u(t) is the controller output. The closed loop prediction error e(t)=u(t)-û(t) is used in the Parameteric Adaptation Algorithm (P.A.A.) as an adaptation error in order to estimate recursively the parameters of the reduced order controller.
This scheme is used for the controller identification methods : conid, conidf and conidaf as well as for the controller order reduction functions cor and cortr.
See: Controller order reduction module .
(file cidschem.htm)