Automatic Modulation Recognition of Digital Signal Based on Auto-encoding Network in MIMO System

Abstract

Automatic Modulation Recognition (AMR) is of great significance in civil and military applications. Cumulant-based recognition is one of the effective methods for AMR. However, different from that in conventional SISO systems., Cumulant-based AMR in space division multiplexing MIMO systems faces the problem of lower recognition rate since the statistical characteristics of signals are different in these systems. In order to solve this problem., we propose to adopt auto-encoding network (AEN) as the data dimension reduction algorithm and artificial neural network (ANN) as the classifier of several kinds of digital modulation signals in MIMO system. In our simulations., the proposed scheme is used to classify five types digital modulation signals., which are 2PSK., 4PSK., 8PSK., 16QAM., 32QAM. Simulation results indicate that the proposed method will realize substantially higher recognition rate compared with direct classification by cumulants.