A novel multi-user collaborative cognitive radio spectrum sensing model: Based on a CNN-LSTM model

Area of Science:

• Wireless Communications
• Artificial Intelligence
• Signal Processing

Background:

• Cognitive Radio (CR) technology enhances spectrum utilization by enabling devices to sense their environment.
• Efficient spectrum sensing is crucial for avoiding interference with licensed users.
• Existing methods require improvement in accuracy and efficiency for multi-user systems.

Purpose of the Study:

• To develop a novel hybrid deep learning model for multi-user cooperative spectrum sensing in CR systems.
• To enhance spectrum sensing accuracy and efficiency by combining CNN and LSTM capabilities.
• To improve the model's adaptability and robustness in dynamic wireless environments.

Main Methods:

• A hybrid model integrating Convolutional Neural Networks (CNN) for feature extraction and Long Short-Term Memory (LSTM) for sequential data handling.
• Incorporation of a multi-head self-attention mechanism to optimize information flow.
• Quantitative performance evaluation through simulation experiments.

Main Results:

• The proposed CNN-LSTM model achieved low sensing error rates across various user configurations (16-48 users).
• A sensing error of 9.9658% was recorded under a 32-user configuration.
• The model demonstrated a 12% lower sensing error compared to other deep learning models under low-power conditions (100 mW).

Conclusions:

• The developed CNN-LSTM model significantly enhances spectrum sensing accuracy and efficiency in multi-user CR systems.
• The integration of CNN and LSTM effectively handles long-term dependencies in time-series data.
• The multi-head self-attention mechanism improves adaptability to complex environments, showing practical application potential.