The theoretical performance of optimal time-dependent filters for co-channel interference rejection

Using the performance criterion of mean-square-error, periodically time-dependent and almost periodically optimal time-dependent filters (OTDF) are shown to be theoretically superior to time-independent optimal filters for suppressing cochannel interference in digital communications. The time-dependent filter is shown to be capable of exploiting the spectral correlation properties of digital communication signals to provide superior performance. The mathematical framework of the OTDF is formulated, and its performance for extracting various modulated signals in noise and for extracting these signals in interference is examined. The performance of the OTDF is described in terms of mean-square error. Examples that demonstrate the theoretical performance advantage of the optimal time-independent filter over the time-independent optimal filter for digitally modulated signals are presented