Spectrum Sensing for Cognitive Radio in 5G and Beyond - Progressive Approach

As technology evolves in the field of wireless communication, the demand for new applications is also increasing, due to which the need for radio frequency spectrum has also increased significantly. Cognitive radio (CR) technology is one such solution that improves spectral efficiency, which cannot be solved by traditional methods. So far several techniques have been proposed for spectrum sensing. In this study, challenges for cognitive radio have been addressed in 5G, while achieving accuracy, faster performance, low computational complexity, data availability and real time adaption. Compressed Sensing (CS) is used to make spectrum sensing efficient in wideband scenarios. Adaptive decision making can be improved by integrating Reinforcement Learning (RL) into CR. In this paper we propose a hybrid model of RL and CS that optimizes spectrum sensing, in which the RL agent gets rewards by adaptively improving spectrum sensing strategies. It is shown through simulations that this hybrid approach significantly enhances the spectrum sensing performance, and address the identified challenges, this hybrid solution is more useful in dynamic environments like IoT, 5G and beyond.