Chicken swarm optimization modelling for cognitive radio networks using deep belief network-enabled spectrum sensing...

Area of Science:

• Wireless Communication
• Signal Processing
• Artificial Intelligence

Background:

• Cognitive radio networks (CRN) are crucial for efficient spectrum utilization.
• Traditional spectrum sensing techniques face challenges with low Signal to Noise Ratio (SNR) and limited signal samples.
• Existing models struggle to adapt to dynamic radio environments, impacting performance.

Purpose of the Study:

• To propose a novel spectral sensing technique for cognitive radio networks (SST-CRN).
• To address the limitations of traditional energy detection models in CRNs.
• To improve spectrum sensing accuracy and efficiency through advanced machine learning.

Main Methods:

• A deep belief network (DBN) is employed for spectral feature extraction and pattern recognition.
• The chicken swarm algorithm (CSA) is utilized to establish a nonlinear threshold for sensing.
• The SST-CRN technique operates in two phases: offline training and online adaptation.

Main Results:

• The proposed DBN-enabled SST-CRN achieved a higher probability of detection (Pd) of 0.810 at -24dB SNR.
• This surpasses existing methods, which recorded lower Pds (0.577 and 0.736).
• The technique demonstrates improved spectrum occupancy assessments and resilience.

Conclusions:

• The DBN and CSA synergistic framework significantly enhances CRN spectrum efficiency and reduces interference.
• Deep learning, specifically convolutional neural networks, enables automatic adaptation to complex radio environments.
• The proposed SST-CRN offers superior accuracy and flexibility compared to classic spectrum sensing approaches.