Modulation classification of MQAM signals using particle swarm optimization and subtractive clustering

This paper proposes a novel algorithm for modulation recognition of MQAM signals which uses the combination of subtractive clustering (SC) and particle swarm optimization (PSO) (PSO-SC) to extract the discriminating features. The method uses PSO to search for the best clustering radius of SC in order to enable reconstructed constellation optimal. Then, the best clustering radius (CR) is used as classification feature. Compared with the classification methods available using subtractive clustering, the algorithm proposed by this paper has higher correct classification rate in modulation classification for MQAM signals. In addition, simulation results show that the modulation classification method performs robust in the low signal-noise ratio.