Object output-response predictors (OORP). II: OORP design with incomplete a priori information on an object (Q1287290)

In Part I [Autom. Remote Control 55, No. 12, Pt. 1, 1707-1722 (1994); translation from Avtom. Telemekh. 1994, No. 12, 3-22 (1994)], the wide opportunities of application of the regular and stochastic components of OORP, OORP(r) and OORP(s) respectively, for the successful solution of diverse control problems were shown. However, the problem, which is typical for reality, when the available a priori information about the object is insufficient for a design using the OORP guaranteeing satisfactory accuracy, has remained aside. The purpose of the paper is to fulfill, to some extent, the mentioned blank. A technique of predictor identification on the basis of properly integrated a priori and a posteriori data is set forth and proved. The main attention is paid to the renewal problem of OORP(r).