Parsimonious correlated non-stationary models for real UWB data

Modelling ultra-wideband (UWB) received signals is an indispensable step to the UWB receiver design and UWB data regeneration. A popular framework for UWB modelling stems from the discrete multipath channel model whose path gains and time delays are random variables and thus, must be specified by their probability density function (pdf). Besides, various partial characterization is used in the literature by virtue of second-order statistics (such as power delay profile), non-parametric characteristics (such as zero-crossing rate), or their combination. So far, little UWB models have the capability to account the correlation structure existing among received UWB data and little work directly addresses the original UWB data. In this paper, we take a different philosophy which believes that the information in the received UWB data itself, as long as fully exploited, plus some simple physical intuition should suffice for the model identification and its parameter estimation. The model so obtained is directly for the original data while having the capability to account for the correlation structure and non-stationarity of UWB data. The application of the new model to data regeneration is illustrated by using the real UWB data provided by the TimeDomain Corporation.