Journal
Scientific and technical journal «Priborostroenie»
UDK681.5.015
Issue:7 (65)
Download PDF478 Kbyte
In the framework of the deterministic approach, an algorithm for identifying nonstationary parameters for the classical linear regression equation is proposed. When synthesizing the algorithm for estimating non-stationary parameters, it is assumed that the dynamic model of their variation is known and is a linear generator with variable coefficients. An additional complication of the problem of estimating parameters for the linear regression equation is the presence of an additive sinusoidal disturbing effect with unknown constant amplitudes, frequencies, and phases. The resulting algorithm provides an accurate estimation of all unknown non-stationary parameters.