Theoretical and Applied Aspects of Use of a Filter with Integral Residuals to Control Dynamic Objects Under Disturbances

In this paper we propose a filtering algorithm, that, in contrast to the classical filtering algorithms, uses in its feedback not only the residuals between the measured and estimated coordinates, but also the integrals of these residuals. In addition to the phase coordinates estimation, the use of integrals from residuals allows to restore the unknown disturbances, acting on the control object. This, in turn, allows to apply the algorithm for real control objects in real operating conditions. External disturbances are treated as union of the external forces from the environment, and disturbances associated with inaccurate information about the model of the control object itself. In the work the analytical dependencies are obtained to evaluate feedback gains, which allows the use of this algorithm in real time with on-board computers. To be able to control the object in the real conditions, a law of control is proposed that includes the obtained estimates of the phase coordinates and unknown disturbances. This control algorithm is based on compensation of the total disturbance vector. The effectiveness of the proposed algorithms was tested on a model of an underwater vehicle that is described by a system of differential equations including the action of wave and spasmodic disturbances with incomplete information about the motion model.