A Direct Blind Demodulation Method for FSK Signals Based on Feature Extraction
Xun Han, Jia Zheng, Xin Feng, Sujun Wang, Wen Li Wei, Yin Kuang · 2024
Blind demodulation of frequency-shift keying (FSK) signals serves as the foundation for subsequent information extraction. However, the existing two-stage architecture blind demodulation method is constrained by front stage parameter estimation errors, resulting in significant symbol rate estimation inaccuracies and demodulation losses. To address this limitation, this paper proposes a novel direct blind demodulation method for FSK signals based on feature extraction. The proposed method establishes a theoretical model for blind FSK signal demodulation and quantitatively describes the demodulation results using features such as frequency distribution entropy and centroid position. Furthermore, a precise synchronization process is designed based on these features to integrate symbol rate estimation and non-cooperative demodulation. Simulation experimental results demonstrate that compared to existing algorithms, the proposed method significantly improves both symbol rate estimation accuracy and reduces demodulation losses.