Novel Collaborative Framework for Cloud-Based Cognitive Services

Cognitive Radio (CR) is transforming wireless communications by optimizing spectrum usage. CR users can access the spectrum only when primary users are inactive, requiring reliable spectrum sensing. Cooperative spectrum sensing, where data from multiple nodes is shared with a central hub, has emerged as the most effective method, surpassing local detection in dealing with challenges like noise uncertainty and channel instability. This study explores two fusion algorithms for merging local spectrum data: hard decision fusion (HDF) and soft decision fusion (SDF). While SDF offers better detection capabilities, it demands excessive bandwidth. In spectrum allocation, a complex NP-hard problem, optimization methods, especially swarm intelligence inspired by honey bee behavior, have been researched. To address spectrum assignment issues, a novel hybrid Deep Recurrent Neural Network (DRNN) approach is proposed. This hybrid model, incorporating 2-opt and additional search techniques, enhances spectrum utilization by overcoming slow convergence and local optima challenges. Empirical results demonstrate its superiority in optimizing spectrum allocation.