The use of digital signal processing for a modulation classification

A modulation classifier, a device that automatically recognizes the modulation type of received radio signals, plays an important role in several communications applications. One solution of a modulation classifier is described in the paper. It is based on the pattern-recognition approach, where some discriminating features are extracted from incoming signal. In the proposed solution, the signal envelope and its instantaneous frequency are selected as features. They are further processed to obtain more appropriable parameters which can indicate modulation type of the signal. The device is intended for the implementation in radio surveillance, supervision and monitoring systems. It allows automatic recognition of AM and FM emissions, usually met in HF band, in the real time. The classifier is designed and realized using DSP technique.