A Communication Signal Recognition Method Based on Improved Entropy Cloud Feature

Abstract

Due to the completing of communication environment, the signal is affected by time-varying noise during transmission, which leads to the characteristics of the signal is unstable. Aiming at this problem, this paper proposes a communication signal modulation recognition method based on improved entropy cloud characteristics. Firstly, the Shannon entropy, index entropy and norm entropy of the signals are extracted. Secondly, using these entropy features and integrated cloud models to get improved entropy cloud features. Finally, extreme learning machine based on particle swarm optimization and principal component analysis (PSO-ELM-PCA) is applied to signal recognition. The results of simulations show that the approach in this paper still has a good classification result at dynamic SNR (Signal to Noise Ratio).