An M-QAM Signal Modulation Recognition Algorithm in AWGN Channel

Table 2

Comparison between proposed classifier with other systems in the literature.

Reference	Classifier structure		Modulation type		System performance		Simulation tool
	Features	Complexity	M-QAM	M-PSK	SNR (dB)	Accuracy (%)
[33]		High	16, 64	4	−3	Not covering	MATLAB functions are invoked to implement and evaluate the performance
					0	Not covering
						78.4–100
[34]		High	16, 64	2, 4	−3	Not covering
					0	Not covering
						89.8–100
[11]		Medium	16∼256	No	−3	98.33
					0	100
						100
[35]		High	16, 64, 256	2, 4, 8	−3	Not covering
					0	77.6
						99.96
[23]		High	16, 64, 256	2, 4, 8, and others	−3	Not covering
					0	Not covering
						Unknown
[21]		Medium	16, 64, V.29	4	−3	60
					0	81
						90
[10]		Medium	8, 16, 32, 64	4, 8, 16, 32, 64	−3	26
					0	45
						95
Proposed algorithm	Logarithmic cumulant	Medium	(8∼1024)	4	−3	75–95
					0	80–98
						100

“∼” refers values from M-QAM to M-QAM and “,” refers values M-QAM and M-QAM; note that signal length N = 4096.