کنفرانس مهندسی برق ایران

صفحه اصلی / سی و سومین کنفرانس بین المللی مهندسی برق

Energy Allocation Methods in NOMA Modulation Using Machine Learning Algorithms in the Presence of Jamming

نویسندگان :

Khashayar Saremi¹ Bahareh Akhbari²

1- دانشگاه صنعتی خواجه نصیرالدین طوسی 2- دانشگاه صنعتی خواجه نصیرالدین طوسی

کلمات کلیدی :

NOMA،SWIPT،Reinforcement learning،Cooperation،Jamming

چکیده :

This research aims to investigate the capability of Non-Orthogonal Multiple Access (NOMA) systems to encounter jammed channels, focusing on both intelligent and non-intelligent jammers. Our case study involves utilizing message relay along the optimal path, leveraging energy harvesting and artificial intelligence (AI). Specifically, we employ reinforcement learning (RL) methods, including Q-learning and Dyna-Q learning. The latter is a combination of the Q-learning algorithm with Dyna architecture and additional modifications, resulting in our customized Dyna-Q algorithm. This algorithm is integrated with NOMA cooperative and NOMA Simultaneous Wireless Information and Power Transfer (SWIPT) to effectively navigate jammed channels. The system’s performance is evaluated against various jammers, and our approach is benchmarked against classical methods. The research mainly emphasizes the NOMA system’s ability to simultaneously transmit energy and information while utilizing Dyna-Q-based learning.