Frequency Hopping Modulation Format Recognition Based on Higher-Order Cyclic Cumulants and Neural Network Classifier

In the field of wireless communications, the recognition of frequency hopping (FH) signal modulation formats is of significant importance for the detection and analysis of anti-drone communication signals. However, existing methods for FH signal modulation have certain limitations in handling complex signal environments, especially in terms of frequency offset resistance, timing error, anti-interference capability, and generalization ability. To address these issues, this study proposes a novel method for recognizing the modulation formats of FH signals. The Tektronix real-time spectrum analyzer is used to capture the FH signals generated by the signal source in real-time, and the signals are then processed to remove the FH. Based on this, higher-order cyclic cumulants are further calculated to deeply analyze the statistical characteristics of the signals. Finally, with the aid of a well-trained neural network classifier, precise recognition of modulation formats such as BPSK, QPSK, 8PSK, and 16QAM is achieved. This study not only improves the accuracy of FH signal recognition but also demonstrates excellent performance in resisting frequency offset and timing errors through the collection and processing of real signals. It provides a novel and efficient method for recognizing the modulation formats of FH signals in communication systems.