A nonlinear filtering approach to carrier tracking and symbol detection in digital phase modulation

This paper considers the application of nonlinear filtering to carrier estimation and symbol detection in additive white Gaussian noise. Digital phase information is formulated as a sequence of transients of a deterministic dynamical system whose state is randomly initialized at the beginning of each symbol interval, according to an encoding scheme. Random carrier phase is modeled as the first component of a vector Markov process. Symbol-by-symbol detection and phase tracking within symbol intervals are performed by a bank of matched stochastic nonlinear phase estimators. The receiver is a parallel open-loop structure suited to all-digital modems.<>