AR modeling-based features extraction of multiple signals for modulation recognition

Modulation recognition of multiple signals is a new topic in the communication signal processing area. A novel feature-based modulation recognition method for multiple signals is proposed. The main idea of this method is that AR modeling is employed to extract instantaneous frequency and bandwidth features of each signal to be recognized from the mixed waveform intercepted by a receiver, and then a MLP neural network classifier is followed to identify the modulation types. To assess the efficiency of these features, computer simulations are also performed