New approach of automatic modulation classification based on in phase-quadrature diagram combined with artificial neural network

Abstract

Automatic Modulation Classification (AMC) with intelligent system is an attracting area of research due to the development of SDR (Software Defined Radio). This paper proposes a new algorithm based on a combination of k-means clustering and Artificial Neural Network (ANN). We use constellation diagram I-Q (In phase, Quadrature) as basic information. K-means algorithm is used to normalize data transmitted and pollute by the Additive White Gaussian Noise (AWGN), then the new diagram obtained is considered as an image and coded in pixel before entering in MLP (Multi-Layer Perceptron) Neural Network. Simulation results show an improvement of recognition rate under low SNR (Signal Noise Rate) compare to some results obtained in the literature.