Linear and nonlinear prediction techniques for short‐term forecasting of HF fading signals

Abstract

Abstract. There exist two major mechanisms which are responsible for the fading phenomenon at HF frequencies. They are the multiple-mode interference and distortions due to the ionospheric irregularities. Fading time series produced by the first of these mechanisms alone should typically represent a multiple-periodical process. This kind of signal ma•v also be produced by an autonomous dynamical system. The character of the time series produced by the second mechanism depends on the nature of the ionospheric irregularities. Recently, evidence has been accumulated to show that sometimes the ionospheric turbulence on equatorial and middle latitudes represents a low-dimensional deterministic chaotic process. These facts suggest that for both mechanisms the fading time series may have a deterministic nature and therefore is predictable. Accordingly we apply the nonlinear predicting technique proposed by Farmer and Sidorovich [ 1987] to the fading time series obtained by the University of Illinois HF sounder. In its application the prediction technique is modffied to take into account specifics of the HF data. For comparison, the conventional linear autoregression prediction technique is also tested. It is found that, in general, the nonlinear prediction and the linear autoregressive forecasting allow prediction on a few correlation times and work with roughly the same success.

A Fault-Signal-Based Generalizing Remaining Useful Life Prognostics Method for Wheel Hub Bearings

“…x can be prognosticated with the solution of method in Equation (18). The RUL of wheel hub bearings can be obtained in Equation (19).…”

“…The parameters a, b can be acquired with the least square method. The solution of RUL prognostics method in Equations (19) and (21) can be expressed as Equation (22).…”

“…The data driven method for bearing RUL prognostics has shown promising potential to support smart prognostics and to avoid the design of the complicated mathematical expressions. Many data-driven methods were proposed for the RUL estimation of rolling element bearings, including support vector regression [15], sparse representation theory [16], adaptive Wiener process method [17], the time series method with Linear and nonlinear techniques [18,19], nonlinear autoregressive neural network [20,21], localized maintenance method [22], sparse representation method [23], physics-based algorithm [24], the rain flow cycle counting method [25], and grey theory [26,27]. These RUL estimation methods give static estimations of RUL.…”

Prediction the Chaotic Time Series of the Radio Signal Propagates through the Ionosphere

“…The characteristic of the attenuation time series of radio wave have a stable short interval, if we used the conventional methods to predict it, the prediction accuracy will be inevitably affected.so we modified the prediction theories given by Farmer and Sidorovich [4,6].…”