CNN-LSTM Hybrid Architecture for Over-the-Air Automatic Modulation Classification Using SDR
The episode discusses a paper detailing a CNN-LSTM hybrid architecture used for automatic modulation classification via SDR. The authors achieved 93.48% accuracy while maintaining high reliability across extreme noise levels, specifically from zero to twenty dB SNR. This hybrid approach surpasses older methods by combining spatial and temporal analysis, making it suitable for robust, real-world deployment.