Identification of the parameters of controlled objects in closed systems

Abstract

Matters of applying the correlation approach to solving the problem of identifying controlled objects in closed control systems are examined. It is established that identifiability in a closed system is determined by the type of auto- and cross-correlation functions of stochastic processes at the input and output of the object as well as by the magnitude of the transmission channel delay of the control actions and by the transfer function of the controller. The possibility of using identification criteria in the form of equalities of cross-correlation functions of the output variable and disturbance of the control action is substantiated. A comparison of the correlation method of identifying the parameters of an object in closed systems and least-squares method showed the applicability of the latter approach to a limited class of closed systems.