Cooperative Spectrum Sensing Based on Multi-Features Combination Network in Cognitive Radio Network

Area of Science:

• Electrical Engineering
• Computer Science
• Signal Processing

Background:

• Spectrum scarcity necessitates efficient radio spectrum utilization.
• Cognitive radio technology and spectrum sensing are crucial for dynamic spectrum access.
• Convolutional Neural Networks (CNNs) and Gate Recurrent Units (GRUs) offer complementary feature extraction capabilities.

Purpose of the Study:

• To develop an enhanced cooperative spectrum sensing scheme for cognitive radio.
• To improve spectrum sensing reliability by fusing multi-node local information.
• To achieve robust signal detection across various modulation types and dynamic signal-to-noise ratios.

Main Methods:

• A hybrid CNN-GRU network was designed to extract spatial and temporal features for single-node spectrum sensing.
• A Multifeatures Combination Network integrated features from CNN-GRU for cooperative sensing.
• The model was trained using eight different signal modulation types.

Main Results:

• The proposed cooperative spectrum sensing scheme demonstrated enhanced reliability through multi-node information fusion.
• The method achieved improved detection performance without requiring prior knowledge of primary user or channel state.
• Competitive performance was observed under conditions of large dynamic signal-to-noise ratio.

Conclusions:

• The CNN-GRU based cooperative spectrum sensing approach effectively enhances spectrum utilization in cognitive radio.
• The fusion of local sensing data from multiple nodes significantly boosts detection accuracy and reliability.
• This method offers a robust solution for spectrum sensing, adaptable to diverse signal types and challenging channel conditions.