|
Abstract: Based on the cyclostationary process theory, the cyclic
partial coherence method is proposed to analyze sound field radiated
from multiple cyclostationary sources. The second-order cyclic
statistics is adopted by the proposed method instead of ordinary
second-order statistics. Instead of the power spectral density (PSD)
distribution, the cyclic spectral density (CSD) distribution of partial
sound field radiated from each source respectively can be obtained when
the proposed method is applied in the decomposition of partial sound
field of cyclostationary sound field. Different characteristics of sound
field can be extracted by the CSD distribution in different cyclic
frequencies, which may be mixed up in the frequency domain of the PSD
distribution obtained by the ordinary partial coherence method, and not
be shown clearly. Consequently, the proposed method can extract
information about the cyclostationary sound field more accurately than
the partial coherence method can, which is validated by a simulation
instance.
Key words: Partial coherence analysis Cyclostationary Cyclic
autocorrelation function Cyclic spectrum density
Decomposition of partial sound field
CLC No: TN911
国家自然科学基金资助项目(10474065). Received 20040728,
received in revised form 20050117
|