Methods and algorithms for presenting energy characteristic of signals in the time-frequency domain
Students Name: Nazar Bohdan Zenonovych
Qualification Level: magister
Speciality: Computerized Control Systems and Automatics
Institute: Institute of Computer Technologies, Automation and Metrology
Mode of Study: full
Academic Year: 2021-2022 н.р.
Language of Defence: ukrainian
Abstract: Many methods exist now for harmonic signals measuring on the basis of the analog-to-digital converting of signals. Such methods, however are being substantially complicated when parameters measurement is considered conformably to the signals with noise interference of different types. Most oftenly the integral measuring methods are now applied in such cases. The utilization of these methods, however, requires some apriority information about a signal (frequency, period, variability character). The most effective of these methods are those based on signals weight integration. Such methods are well described in numerous sources. Although even these methods are in most cases inefficient, if it is necessary to make an estimation of nonperiodical signals with a high noise interference content. It is especially true when their variability form and character are unknown. This work examiner an entirely new measuring method which is based on a signal wavelet transform application. When this method is used, a signal to be measured is preliminary scaled and quantized; afterwards its wavelet transform is performed with one of known methods being used (for example that involving Dauby matrixes) The result of the transform is an output signal in the form of wavelet coefficients. To remove noise interference different methods of truncation are used in the signal. The result of such transforms is an energetic representation of an original signal cleared from noises. Applying an inverse transform a discrete representation of the signal is performed, the representation being cleared from noises and later estimated with one of the known algorithms. Another method is possible of a signal estimation, without an inverse wavelet transform being used. In that case a signal parameters are estimated directly by means of wavelet coefficients, which were received. That estimation, however can be made for signals energetic parameters such as energy or power.