Abstract
This work summarises and suggests new extensions of three recently developed techniques for recovering harmonic signal information in the presence of additive unknown colored noise. This setting is important in many noise and vibration problems wherein it is desirable to distinguish truly periodic components from random narrowband components. The techniques considered include a modified and extended version of Pisarenko's decomposition, a state space approach, and a maximum likelihood spectral family method. After summarising the key elements of each of these and pointing out methods for improving their utility a numerical example is presented to illustrate their harmonic retrieval properties, in contrast to those of the basic discrete Fourier transform and autoregressive methods.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call