Automatic digital modulation classification for ORS satellite relay communication

Abstract

Automatic Modulation Classification (AMC) can be used in automatically identifying and classifying the modulation of communication devices. With the application of digital technique, AMC is developed towards higher frequency, which makes a lower probability of correct classification (PCC) at the conventional method. It is necessary for relay-communication to automatically classify the modulation of satellite. So, AMC plays an important part in heterogeneous satellite networking especially in Operationally Responsive Space (ORS). In order to enhance PCC in low Signal Noise Ratio (SNR) conditions, a novel method based on Radical Basis Function Neural Network (RBFNN) and Gravitational Search Algorithm (GSA) was presented in this paper. This method combined high-order cumulants with low-order statistics features, and supposed additive white Gaussian noise (AWGN) as the channel model. The classification performance of the typical RBFNN was optimized by GSA using information entropy changing to update the “agents” movement velocity, which expand the globe solution sets in exploration phase and escapes the local optimum in exploitation phase. Compared to existed methods, the proposed method does not require any previous knowledge of received signal. Simulation results show that the proposed method is more effective in low SNR conditions and improves the probability of correct classification.