Adaptive Prediction and Cancellation Digitization Method for Wideband Multistandard Software Radio Base-Station Receivers.

In this study, a new method for digitizing a combination of different analog signals occupying significantly different bandwidths and having a very high dynamic range is proposed and analyzed. Since it is based upon signal-prediction/cancellation principles, it is referred to as adaptive prediction and cancellation digitization (APCD) method and is applied to various families of signals simultaneously received by a multistandard software radio (SWR) base-station receiver. It is shown theoretically and by means of computer simulations that the APCD method can effectively reduce the high dynamic range of the signals before digitization takes place. Hence, the stringent analog-to-digital-converter (ADC) resolution requirements imposed by the operation of such SWR base-station receivers can be significant relaxed. The signal dynamic-range reduction is achieved by applying appropriate signal processing techniques, e.g., autoregressive (AR) and periodic autoregressive (PAR) prediction. Such techniques allow accurate prediction and subsequent cancellation of high-power narrowband signals present among the composite received analog signal. As these signals usually have cyclostationary statistical characteristics, analysis and performance evaluation of AR and PAR predictors, when used to predict cyclostationary signals, were presented. A new adaptive algorithm for implementing the PAR predictor is also proposed, and its validity is justified by theoretical analysis as well as by various performance evaluation results obtained by means of computer simulations. [ABSTRACT FROM AUTHOR]