AI-Enhanced Cooperative Spectrum Sensing for Non-Orthogonal Multiple Access

Many state-of-the-art techniques are leveraged to improve spectral efficiency, of which cognitive radio and multiple access are the most promising ones. In cognitive radio communications, spectrum sensing is the most fundamental part, whose accuracy has a significant impact on spectrum utilization. Furthermore, due to the complex radio environment, multiple-user CSS has been proposed as a refined solution. NOMA, as an essential technique in 5G, holds great promise in improving spectral efficiency and carrying massive connectivity. In this article, we propose a novel CSS framework for NOMA to further improve the spectral efficiency. Considering the complicated physical layer implementations of NOMA, we introduce an AI based solution to cooperatively sense the spectrum with a nice accuracy rate and acceptable complexity. Numerical results validate the effectiveness of our proposed solution.