Communication Emitter Identification Based on Kernel Semi-supervised Discriminant Analysis

In this paper, an efficient algorithm based on kernel semi-supervised discriminant analysis for communication emitter identification is proposed. This algorithm uses square integral bispectra to extract the bispectra features from the communication emitter signal as its fingerprint. Simultaneously, to improve the communication emitter identification results, the kernel semi-supervised discriminant analysis algorithm maps high-dimensional bispectra feature data to a low-dimensional subspace and then identifies this data in that subspace by its nonlinear manifold information and partial label information. Sample data of ten FM radio stations with the same manufacturer, batch and type are used in an identification experiment. The experiment result proves that, despite there being fewer labeled training samples, the proposed method still has high recognition performance.