Characterization and parameterization of dynamic wireless channels over long duration using evolutionary channel parameters

The characterization and parameterization of processes that arise in many fields of science and technology are very crucial. Of particular importance are dynamic processes whose statistics are time-varying and are often modeled as stochastic processes. A typical example of such process is the wireless communication channel. Existing methods that are used to characterize and parameterize the dynamic stochastic wireless channel often consider short-term duration over which the channel statistics are invariant. Conversely, this paper presents the characterization of the dynamic wireless communication channel over a long-term duration where time/frequency channel realizations are obtained at sample intervals. To structure such channel realizations over a long duration, the idea of concatenating the 'instantaneous' channel realizations is presented. The resultant concatenated multivariable process is characterized using the concepts of process non-summability and piecewise separability. Based on these concepts, the second-order statistical parameterization of the concatenated stochastic process in both time and frequency domain is presented. The parameterization approach is based on fitting appropriate set of unit step functions that approximate the raw concatenated data using sets of evolutionary stationarity parameters. To illustrate the concepts developed in this paper, measurement-based experiments and analysis are presented and adaptively applied to improve wideband multicarrier system performance.