Identification of systems with cyclostationary input and correlated input/output measurement noise

An alternative approach to Wiener's optimum filtering theory for stationary random time series that is applicable when the system input exhibits cyclostationarity is presented. This alternative approach can, in principle, eliminate performance degradation due to input/output corruption. This method of system identification is called the SPECCORR (spectral correlation ratio) method, since the cross spectra in the ratio are actually spectral correlation functions. An application to interference-tolerant time-difference-of-arrival estimation is described and illustrated by simulation results, showing its effectiveness for system identification with severely corrupted input/output measurements. A comparison of Wiener's method to the SPECCORR method shows the superior performance of the latter.<
>