A New Approach to Eliminating the “Flip” Effect of the Approximating Function Under Conditions of a Priori Uncertainty

The paper considers a new approach to compensate for errors that arise when processing measurement results under conditions of a priori uncertainty due to the appearance of the “flip” effect when approximating the measured function over a certain observation interval. The proposed approach is to increase the approximation interval at which a “flip” is detected to a value at which this “flip” effect is not observed. However, the obtained approximation values are used only on the interval of the initial value, i.e. the interval where the “flip” effect was observed. The approach used for processing the measurement results makes it possible to significantly reduce the error in estimating the function of the measured process under conditions of a priori uncertainty. #COMESYSO1120.