This paper presents a method for modulation recognition of digital signals using machine learning. This method first extracted seven characteristic parameters according to the instantaneous parameters and high-order cumulant characteristics of the signal, then combined the decision tree classifier, neural network classifier and support vector machine classifier in machine learning with these characteristic parameters, and finally realized the modulation recognition of MASK, MFSK, MPSK and MQAM signals. This method not only has low computational complexity and more recognized signals, but also improves the recognition rate at low SNR.
Title = "Research on Modulation Recognition Technology based on Machine learning",
Journal ="International Journal of Computer Applications Technology and Research(IJCATR)",
Volume = "8",
Pages ="82 - 148",
Year = "2019",
Authors ="Biao Xu, Xiping Wen, Xi Wang"}
The characteristic parameters which can distinguish different signals better are proposed
The joint feature is used to improve the recognition rate at low SNR
Signal modulation recognition is realized by using different machine learning classifiers
The engineering performances of different machine learning algorithms are compared.