Modulation Recognition of 5G Signals Based on AlexNet Convolutional Neural Network
Qing Zhang, Guobing Hu, Pinjiao Zhao, Yang Li · Journal of Physics Conference Series · 2020
Abstract Aiming at the problem that signal modulation recognition under a non-cooperative condition requires substantial a priori information of the signal and a complex artificial selection of the features, this paper proposes a modulation recognition method for the 5th-generation (5G) signal modulation based on the AlexNet convolutional neural network. For the five commonly used 5G signals (3GPP R15 protocol recommendations) of π/2-BPSK, QPSK, 16QAM, 64QAM, and 256QAM, the constellation is selected as input feature of the AlexNet network to construct the recognition classification algorithm. The simulation results show that the average recognition accuracy of the five commonly used 5G signals is up to 90% under a 15 dB signal-to-noise ratio (SNR), an improved performance compared with that of the existing recognition algorithms based on signal scatter plots.