An Algorithm of Digital Modulation Identification Based on Instantaneous Features

Automatic modulation identification plays an important role in non-cooperative communication systems. An improved approach, based on instantaneous features and binary tree classifier has been proposed for the classification of Medium Frequency (IF) digital signals at low signal-to-noise ratio (SNR) situation. Mean filter was applied to suppress the noise and five new characteristic parameters such as maa, maf, map, raf and maa1, were extracted for the classifying. It is shown that the proposed algorithm has better performance than the traditional approaches. Theoretical arguments are verified via extensive simulations. Simulation results indicate that the correct classification probability (Pcc) with proposed algorithm has been improved compared with the traditional method.