Automatic modulation classification for cognitive radio systems: Results for the symbol and waveform domains

MQAM Modulation Recognition Based on AP Clustering Method

“…With the rapid development of modern communication technology, in the civil, modulation recognition of communication signals is mainly applied to software radio [1], spectrum monitoring and management, cognitive radio and so on [2][3]. In the increasingly complicated signal environment, some of the illegal users interfere and utilize the wireless spectrum, affecting the normal communication of legitimate users seriously.…”

Automatic Modulation Classification for adaptive Power Control in cognitive satellite communications

“…These are derived respectively from spectral correlation or coherence functions, coherent processing and Time-Frequency (TF) distributions (short time Fourier transform, Wigner-Ville distribution etc.) [10][11][12], [14][15][16][17][18][19][20][21]. As far as these characteristics are concerned, the FFT algorithm is again the main tool for estimating them.…”

“…The kinds of NNs used are feedforward networks with multiple layers and back propagation networks with multiple layers. A powerful classification tool that previous researchers used are the SVMs [3][4][5], [8], [9], [12], [15], [16]. They have been helpful in many categorization problems and they are the most frequent tool of previous work in this field achieving P DE = 100% at SN R = −12dB [15].…”

QAM signals recognition based on Amplitude distribution

“…Currently there are several approaches to this problem. They are the classical decision algorithm based on the likelihood function [1,2] , pattern classification algorithm [3][4][5][6][7][8][9][10] based on the constellation recovery [4,5] and the character extraction algorithm [6,10] .…”

“…A large portion of the character extraction algorithm needs the threshold setting for identification and the distribution of characters is usually associated with the signal parameters. These will make the algorithm more complicated and the threshold setting more difficult [6,9,10] . Moreover, character extraction algorithm usually depends on a priori knowledge [8] , such as some higher-order statistical features extraction methods [5] , in which the accurate estimation of the carrier or symbol rate is necessary.…”