Table 8.
Comparison of classification performances.
| Ref. | Dataset (modulation types) | Channels considered | Method (classifier/features) | Acc. (%) | SNR (dB) |
|---|---|---|---|---|---|
| 19 | OOK, 4ASK, 8ASK, M-PSK(2,4,32), MAPSK(16-128), MQAM(16,64), SSBSC, DSBSC, FM, GMSK, OQPSK | OTA signals | Hierarchical SVM, HOSs | 95 | 5 |
| 31 | BPSK, PAM, QAM, 8PSK | AWGN | Hierarchical classification, cumulant-based | 70 | 5 |
| 32 | 2ASK, 4ASK, 8ASK, BPSK, QPSK, 8PSK | AWGN | Hierarchical, decision tree, higher-order cumulants | 96 | 10 |
| 33 | 4ASK, 8ASK, 2PSK, 4PSK, 8PSK, Star8QAM, V29, 32QAM, 64QAM | AWGN | Hierarchical SVM, higher order moments and cumulants | 96 | 3 |
| 15 | FSK, 2ASK, 4ASK, 8ASK, QPSK, 8PSK, 16QAM, 64QAM | - | Naive SVM, Hierarchical-local density | 97.29 | 16 |
| 34 | M-PSK(2,4,8), M-QAM(16,64,256) | AWGN, slow flat fading | Hierarchical polynomial classifier, cumulants | 86.52 | 5 |
| This work | M-PSK, M-FSK, CPM (M,h,L) | AWGN, SUI-1 to SUI-6 | Hierarchical decision trees, neural network | 98.57 | [-3,15] |
Significant values are in bold.