Intelligent Spectrum Sensing: When Reinforcement Learning Meets Automatic Repeat Sensing in 5G Communications

Abstract

Spectrum sensing, which helps to resolve the coexistence issue and optimize spectrum efficiency, plays an important role in future wireless communication systems. However, the upcoming 5G communication involves diversified scenarios with different characteristics and diverse requirements. This tendency makes it difficult for spectrum sensing methods to flexibly serve various applications while maintaining satisfactory performance. Motivated by such a challenge, this article combines the reinforcement learning concept with spectrum sensing technique, seeking a feasible way to adaptively deploy spectrum sensing configurations so as to optimize system performance under multifarious scenarios in 5G communications. In this article, we first categorize several major optimization targets for spectrum sensing in future communications. Then we elaborate the full details of the proposed sensing technique. Three dedicated modes with respective optimization objectives are designed thereafter. Numerical results manifest that the proposed sensing technique has the capability of adapting to various scenarios, which is appealing in practice.