The Delay-Doppler Spectral Analysis of Nonstationary Time Variant Linear Channels.

摘要

The channel characterization and measurement problem for physical time variant channels is complicated by nonstationary statistical behavior. This complication is, in part, analogous to that perceived in the analysis of speech signals, particularly with a view to the characterization of such signals in identifiable classes. The earliest and still popular approach to speech analysis involves a short-term running spectral analysis in order to follow the nonstationary changes and to provide identifiable characteristics (e.g., formants). A similar approach is necessary in the measurement of nonstationary channels and is developed in some detail in the discussion below. In brief, a short-term running two-dimensional delay-Doppler spectral analysis is studied as a means of characterizing the channel on a short-term basis and allowing observance of nonstationary behavior.