Signal Waveform Extraction in the Presence of Regular and Random Noise

DOI : http://dx.doi.org/10.13005/bbra/1489

Download this article as: 

The paper focuses on the problem of the signal waveform extraction in the presence of random and regular noise. The principal component analysis has been proposed to extract the waveform. Assuming that the analyzed signal in the recorded sequence is repeated with a certain periodicity, several portions containing the analyzed signal can be extracted using the “caterpillar” method. The obtained matrix is then subjected to singular value decomposition. It is shown that the waveform is defined by the first left singular vector. Mathematical modeling demonstrates the possibility to extract the waveform of the analyzed signal in the presence of random and regular noise. The model calculations prove the possibility to extract the signal waveform in case the level of random noise and the correlation of the extracted signal and regular noise change within a wide range.

Signal analysis; regular and random noise; Waveform recovery; principal component analysis