Automatic Modulation Recognition of Multi-type Communication Signals

The automatic modulation recognition of communication signals plays an important role in military and civilian fields. Most of the existing recognition algorithms focus on a single modulation type, and few involve the mixed types including the digital, analog and spread spectrum modulated signals. Based on the instantaneous features of communication signals, this paper constructs a decision tree classifier and a BP neural network one to solve this problem, respectively. The efficiency and superiority of proposed methods were verified by vast experiments. The results show that because BP does not need to set classification threshold, its recognition accuracy is higher, and is more suitable for the application in the actual non-cooperative communication system.