Short-term reliability prediction for highly noisy radio engineering systems

Evaluating the effectiveness of complex systems is becoming difficult due to the increasing complexity of such systems, as well as the high significance of predictions of their technical condition. Therefore, the experts pay considerable attention to them. This is largely due to the large number of components (blocks, nodes, elements) of such systems and the variety of technical and information links between them. All this causes an avalanche-like increase in the number of studied properties that affect the reliability of the entire system. The paper considers the problem of identifying linear and harmonic components in a highly noisy time series based on information obtained during its operation. It shows how to separate various components (linear, harmonic) from the general time series. The work is aimed at short-term prediction of the functioning reliability of radio engineering devices for solving certain fast operations over a given time interval. After the deterministic components are isolated from the time series, the error of its short-term prediction is estimated based on the remaining random centered component by its variance.