On Higher-Order Spectral Moments and the Normality of Random Processes: Application to Linear and Nonlinear Transformations,
PRINCETON UNIV N J DEPT OF ELECTRICAL ENGINEERING
Pagination or Media Count:
The authors apply the theory of higher-order spectral moments to the problem of narrow-band filtering of random processes. The authors prove that the output converges to a Gaussian process. The authors pay particular attention to the rate of convergence and to methods to increase this rate. It is shown that band-pass filters are better behaved in this respect than low-pass filters of equal bandwidth. For a large class of processes it is shown that memoryless nonlinear transformations spread the output spectrum. In particular, this is true for Gaussian processes.
- Statistics and Probability