A NOMA-based Q-Learning Random Access Method for Machine Type Communications

Machine Type Communications (MTC) is a main use case of 5G and beyond wireless networks. Moreover, due to the ultra-dense nature of massive MTC networks, Random Access (RA) optimization is very challenging. A promising solution is to use machine learning methods, such as reinforcement learning, to efficiently accommodate the MTC devices in RA slots. In this sense, we propose a distributed method based on Non-Orthogonal Multiple Access (NOMA) and Q-Learning to dynamically allocate RA slots to MTC devices. Numerical results show that the proposed method can significantly improve the network throughput when compared to recent work.