A differential correlation approach to blind DPSK symbol estimation

This paper presents a new method for the blind symbol estimation of digital communication signals under linearly distortive channels. The algorithm blindly recovers input user symbols by exploiting second order statistics of the channel output signal collected from two data windows of different size. It does not require an intermediate step of channel identification. Moreover, this differential correlation algorithm (DCA) can directly demodulate differentially encoded data without using the more costly Viterbi algorithm. This method is directly applicable to a class of time-varying channels. Simulation examples are presented to demonstrate the performance of this scheme.