|
Home
|
Back to issue
|
ISSN 0474-8662. Information Extraction and Processing. 2018. Issue 46 (122)
Covariance LSM-analysis of biperiodic nonstationary vibration signals
Javorskyj I. M.
Karpenko Physico-Mechanical Institute of the NAS of Ukraine, Lviv
Telecommunication Institute of University of Technology and Life Science,
Bydgoszcz, Poland
Dzeryn O. Y.
Karpenko Physico-Mechanical Institute of the NAS of Ukraine, Lviv
Yuzefovych R. M.
Karpenko Physico-Mechanical Institute of the NAS of Ukraine, Lviv
Lviv Polytechnic National University
https://doi.org/10.15407/vidbir2018.46.019
Keywords: biperiodic nonstationary vibration signal, covariance function estimator, least squares method, asymptotical unbiasedness and consistency, leakage
Cite as: Javorskyj I. M., Dzeryn O. Y., Yuzefovych R. M. Covariance LSM-analysis of biperiodic nonstationary vibration signals. Information Extraction and Processing. 2018, 46(122), 19-28. DOI:https://doi.org/10.15407/vidbir2018.46.019
Abstract
The estimators of covariance function of biperiodically correlated random processes – mathematical models of vibration signals with binary stochastic recurrence, obtained with using the least squares method (LSM), are analyzed. It was shown that these estimators are unbiased and consistent under the condition of correlation relationships decaying with the bias rise. The main LSM-estimator advantage over the component estimator is the absence of leakage effects, which can cause significant errors of covariance characteristics estimation when combination frequencies have close values. Formulae obtained in this paper for statistic characteristics of
LSM-estimator give an opportunity to calculate processing errors for specific signal types and also compare them with the errors of component estimation.
References
1. Javorskyj, I. Mathematical models and analysis of stochastic oscillations. Physico-mechanical institute of NAS of Ukraine: Lviv, 2013, p 804.
2. Javorskyj, I.; Yuzefovych, R.; Dzeryn, O. LSM-harmonic analysis of bi-periodic nonstationary vibration signals. Information extraction and processing. 2017; 121(45), 14-25.
3. Javorskyj, I. On period estimate of periodically correlated random processes. Otbor i Peredacha Informatsiyi. 1986; 73, 12-21.
4. Dragan, Y.; Rozhkov, V.; Javorskyj, I. The Methods of Probabilistic Analysis of Oceanological Rhythms. Gidrometeoizdat: Leningrad, 1987; p 319.