Enhanced Spectrum Sensing in Cognitive Radio Networks Using an SVM-RF Hybrid Model with Transfer Learning

Cognitive radio (CR) networks are designed to optimize spectrum utilization in wireless communications by enabling secondary users to detect unused spectrum and avoid interference with primary users. Spectrum sensing is a vital component of this process. Traditional spectrum sensing techniques often rely on feature extraction from a received signal at a single point, but recent advances in artificial intelligence and deep learning have opened new avenues for improving sensing accuracy. This study introduces a hybrid model that combines Support Vector Machines (SVM) and Random Forest (RF), an ensemble learning method that leverages multiple decision trees to enhance classification accuracy. This approach is particularly effective when dealing with many features or complex feature interactions. Additionally, transfer learning is utilized to further boost the accuracy of spectrum sensing, particularly in low Signal-to-Noise Ratio (SNR) environments. Experimental results show that the SVM-RF hybrid model significantly outperforms existing models in terms of sensing accuracy. An analysis of the algorithm’s complexity underscores the performance improvements, demonstrating the effectiveness of the proposed model in dynamic and challenging wireless environments.