Title:
LEARNING CONTROL SYSTEM
Document Type and Number:
Japanese Patent JPH0540504
Kind Code:
A
Abstract:
PURPOSE:To more correctly predict deviation by predicting the deviation in the future by correcting control input so that the weighted sum of squares of the predicted value of the deviation in the future becomes minimum and considering the change of a correction amount after the present time as well. CONSTITUTION:Trial is repeated so as to make the output of a controlled system follow a target command to repeat the same pattern, and the control input uk (i) at time (i) of the k-th trial is given by an expression (I). In the expression (I), sigmak(i) is the correction amount from the control input uk-1(i) of the last time, DELTAsigmak (i) is the incremental value of the correction amount sigmak (i), DELTAsigmap (i) is an incremental correction amount having been inputted till the present time, eL(i) is follow-up deviation at the present time and in the trial of the last time, Wh1 is a constant determined by information related to the dynamic characteristic of the controlled system and a weighted matrix to be multiplied with the predicted value of the follow-up deviation in the future, HF, HP are the constants determined by the information related to the dynamic characteristic of the controlled system, and W is the weighted matrix to be multiplied with the predicted value of the follow-up deviation in the future.
More Like This:
Inventors:
NAKAMURA YUJI
Application Number:
JP22110791A
Publication Date:
February 19, 1993
Filing Date:
August 05, 1991
Export Citation:
Assignee:
YASKAWA ELECTRIC CORP
International Classes:
G05B13/02; G05B13/04; G05B21/02; (IPC1-7): G05B13/02; G05B13/04
Next Patent: FUZZY INFERENCE CONTROL METHOD