CHAPTER 9
Statistical Errors in Advanced Estimates
This chapter continues the development from Chapter 8 on statistical errors in random data analysis. Emphasis is now on frequency domain properties of joint sample records from two different stationary (ergodic) random processes. The advanced parameter estimates discussed in this chapter include magnitude and phase estimates of cross-spectral density functions, followed by various quantities contained in single-input/output problems and multiple-input/output problems, as covered in Chapters 6 and 7. In particular, statistical error formulas are developed for
Frequency response function estimates (gain and phase)
Coherence function estimates
Coherent output spectrum estimates
Multiple coherence function estimates
Partial coherence function estimates
9.1 CROSS-SPECTRAL DENSITY FUNCTION ESTIMATES
Consider the cross-spectral density function between two stationary (ergodic) Gaussian random processes {x(t)} and {y(t)}, as defined in Equation (5.66). Specifically, given a pair of sample records x(t) and y(t) of unlimited length T, the one-sided cross-spectrum is given by
where X(f, T) and Y(f, T) are the finite Fourier transforms of x(t) and y(t), respectively–that is,
It follows that a “raw” estimate (no averages) of the cross-spectrum ...
