Preliminary Decomposition into Modes in Information-Measuring and Control Systems

In the measurement and control technique, one of the most much-needed tasks is the determination of the parameters and characteristics of complex signals. Well- proven parametric methods based on autoregressive models are very labor intensive. On the other hand, a typical signal processing task is to decompose a signal into components. In the classical version, the representation of a signal as a sum of components (harmonics) is obtained by Fourier transformation. But at present the most promising is decomposition into such oscillations that reflect physical processes that determine the nature of the analyzed signal. The article considers the approach to signal processing based on preliminary decomposition into alternating components using decomposition into empirical modes and extreme filtering, as well as the performance capabilities and applications of this method. Components and their parameters obtained through decomposition methods allow us to analyze the physical nature of the process, obtain spectral estimates, define free and forced oscillations, perform filtering, and significantly reduce the complexity of parametric analysis by applying it not directly to the signal, but to the obtained components. The expediency of these approaches and the prospects for using each of them are shown.