Signal Modulation Recognition Based on Multi-Feature Extraction

Signal modulation recognition in complex electromagnetic environments is of significant importance for spectrum regulation and electromagnetic security assurance. This paper proposes a method and workflow for electromagnetic signal modulation recognition based on multi-feature fusion. The approach first extracts multi-feature parameters, including higher-order cumulants and instantaneous characteristics from signals, constructs a series of feature parameters, and then designs a classification algorithm to identify modulation types. Compared to single-feature extraction and deep learning methods, this method offers advantages such as low algorithmic complexity, rich feature representation, minimal computational resource requirements, simplicity in engineering implementation, fast recognition speed, and high accuracy. It is particularly suitable for deployment in resource-constrained compact and embedded RF instruments, enabling users to quickly determine signal modulation types and improve efficiency in environmental exploration and spectrum monitoring. The feasibility and effectiveness of the proposed method in engineering application are validated on an ARM-based spectrum analyzer using nine modulation types generated by a vector signal generator.