A Review on Reinforcement Learning enabled Cooperative Spectrum Sensing

As the number of devices joining the network explodes, new radio frequency spectrum bands are in greater demand. It is envisaged that cognitive radio networks would solve this issue by providing secondary users (SUs) with opportunistic access to licensed frequency bands from the main network. In order to overcome multi-path fading and shadowing issues, cooperative spectrum sensing (CSS) had been introduced, which allows SUs to share their sensing results and make decisions in a cooperative manner. Reinforcement learning then enters the scene as a highly potent technology that enables SUs to choose the best possible actions that conserve time and energy while guaranteeing a good performance. This paper presents an overview of existing reinforcement learning-based cooperative spectrum sensing schemes and includes a brief description of several existing challenges as well as possible future directions.