Table 1 Comparison table summarizing different and proposed methods.
References | Method | SNR at the BER of 10–3 | PSD |
|---|---|---|---|
Deep learning method | 6.6 dB | − 900 | |
Autoencoder | 10.3 dB | ||
Machine learning | 12 dB | ||
Hybrid iterative technique | 18 dB | ||
SIC-ML | 6 dB | − 800 | |
Bi-LSTM | 18 dB | ||
Linear detection method | 21 dB | ||
Decoupled signal detection | 15 dB | ||
Message Passing Neural network | 9 dB | ||
Information Geometry Approach | 11 dB | ||
Adversarial Network Approach | 6.8 dB | ||
Proposed method | SIC-RL for 512-QAM, 256-QAM, 64-QAM and 64-QAM (10% error) | 11.2 dB, 6.6 dB, 3.2 dB, , 4.2 dB, 5.8 dB and 7.8 dB | − 500 and − 490 |
