APPLICATION OF A HYBRID WAVELET-SPECTRAL METHOD FOR DETECTING DIAGNOSTIC CHARACTERISTICS
The current information and computational aspects are considered in solving problems of diagnosing complex objects and control processes. A hybrid approach based on fast orthogonal wavelet-spectral transform, adapted to the type of characteristics, describing the state of the object being diagnosed, is proposed for the operational formation of the space of diagnostic signs of the minimum dimension. This approach provides high speed of convergence of series for the analyzing class of signals and characterized by high degree of selectivity of transformation in adaptive basis. The advantage of the proposed hybrid approach is the unambiguous calculation of decomposition coefficients in any dimension of the source data and any truncation of the decomposition series. This allows to increase accuracy of the recovery approximated function by adding components from the amputated part of the series.
Authors: А. А. Alekseev, А. М. Spivakovskiy
Direction: Informatics, Computer Technologies And Control
Keywords: Wavelet-spectral transformations, adaptive bases, diagnostics of control objects, diagnostic signs
View full article