Blind equalization and characteristic function based robust modulation recognition

A robust modulation recognition scheme is proposed to identify practical QAM signals, which suffer from various unknown effects including multipath propagation, pulse shaping, time delay, and phase shift. Relying on blind signal processing techniques, the proposed scheme consists of two steps: front-end signal processing followed by modulation classification. At the first step, a fractionally-spaced blind equalizer is applied to accomplish equalization and time synchronization. Phase shift is then estimated and corrected. As required, signal-to-noise ratio is also estimated. At the second step, modulation formats are classified via characteristic function. Simulation results are provided to show the performance of the proposed scheme.