Performance enhancement in clustering cooperative spectrum sensing for cognitive radio network using metaheuristic...

Area of Science:

• Wireless communication
• Cognitive radio networks
• Signal processing

Background:

• Cognitive radio networks (CRN) aim to improve spectrum utilization by enabling secondary users (SU) to detect and access licensed spectrum dynamically.
• Cooperative spectrum sensing (CSS) enhances detection accuracy but faces challenges with data aggregation and energy consumption.
• Clustering techniques group SUs to streamline data transmission to the fusion center (FC), potentially improving efficiency.

Purpose of the Study:

• To optimize the detection performance of CRN through an advanced clustering-based CSS technique.
• To introduce a novel hybrid machine learning algorithm for efficient spectrum gap identification.
• To address computational complexity issues prevalent in conventional clustering methods for CSS.

Main Methods:

• A hybrid Support Vector Machine (SVM) and Red Deer Algorithm (RDA) named Hybrid SVM-RDA was developed.
• The algorithm was applied to a clustering-based cooperative spectrum sensing framework in CRN.
• Performance was evaluated based on probability of detection (Pd) and probability of error (Pe).

Main Results:

• The Hybrid SVM-RDA algorithm demonstrated superior performance compared to conventional clustering techniques.
• Achieved a high probability of detection (up to 99%) and a low probability of error (up to 1%).
• Outperformed existing methods in terms of computational complexity.

Conclusions:

• The proposed Hybrid SVM-RDA algorithm effectively optimizes CRN detection performance using clustering CSS.
• This approach offers a significant improvement in efficiency and accuracy for dynamic spectrum access.
• The method presents a promising solution for enhancing the reliability of cognitive radio networks.