Algorithms for Stable Compensation of Unmeasurable Perturbations in Control Systems for Dynamic Objects

Linearized systems with many inputs and outputs under the action of unmeasured external perturbations are considered. The issues of constructing algorithms for estimating and compensating uncontrolled disturbances in control systems of dynamic objects are analyzed based on fundamental theoretical principles of control by disturbance. It is shown that, under the condition of limited compensating actions, the problem of complete compensation of the movement of a generalized unmeasured perturbation on the state of the system at each moment is unsolvable. The questions of implementation of the regularization method in solving systems of linear algebraic equations are considered. A matrix decomposition using rotation or reflection matrices is proposed as a basic computational scheme. The presented results of the numerical analysis confirmed the efficiency of the algorithms used, which allows them to be used in solving applied problems of synthesis of a control system for dynamic objects under conditions of unmeasured disturbances.