Blind Particle Filtering Detector over Time-Varying Frequency-Selective and Nonlinear Channels

In digital communications, the channel affects the transmitted sequence with both linear and nonlinear distortions. We address the blind detection over time-varying frequency-selective and nonlinear channels in this paper. A blind detector is derived within the Bayesian framework and implemented via the Particle filtering (PF) method. The proposed PF-based detector incorporates auxiliary PF strategy and a hybrid importance distribution to improve the efficiency and effectiveness. A delayed-weight estimation method is also presented to further improve the performance of the detector. Simulations are provided that demonstrate the performance of the detector under different system settings.