Towards Intelligent Machine-Type Communication

Recent advances in communication networks towards an intelligent communication system have led to new challenges, such as massive connectivity, massive traffic explosion, and extremely low latency. Thus, the next generation, 6G, must meet a wide range of requirements and use cases. This thesis aims to develop and analyze machine learning-based strategies to efficiently meet the stringent requirements of emerging intelligent communication systems. We utilize the concepts of classic machine learning, deep learning, and deep reinforcement learning as potential solutions to such challenges. The proposed techniques in this thesis enhance challenging performance metrics such as real-time data collection, end-to-end latency, energy efficiency, and spectral efficiency. Our novel ideas outperformed the existing state-of-the-art solutions and have the influence to be developed and increasingly used in other future research directions.