A Novel Hybrid Ensemble Model to Identify Spectrum Anomaly

Abstract

How to precisely detect electromagnetic spectrum anomaly is a major challenge for radio monitoring, especially in the condition of complex electromagnetic environment and lack of pre-knowledge information about frequency use. Based on spectrum data, the paper presents a novel identification method based on back propagation (BP) neural network and empirical mode decomposition (EMD) of typical spectrum anomaly. Principal component analysis (PCA) is proposed to accelerate the model’s operation. EMD-PCA-ANN method’s identification performance superiority is demonstrated comparing with other method. Experimental result shows that the proposed method can effectively detect spectrum anomaly with high detection rate.