Adaptive Pre-Processing with RLS Algorithm for Automatic M-PSK Modulated Signal Classifier

Abstract

The paper describes problems of multi-level classification of real Multilevel-Phase-Shift Keying modulated signal frames by using methods based on constellation shapes. Modulation classifiers which automatically recognize the type of modulation in high frequency band have been developed. One of the oldest of modulation recognizer uses a bank of demodulators. Each of demodulators is able to recognized only one modulation type. The adaptive pre-processing method based on the Recursive Least-Squares algorithm is used to increase the probability of correct classification of Multilevel-Phase-Shift Keying modulated signals.