Intelligent Beamforming Design in mmWave mMIMO: A Reinforcement Learning Approach

Citations

SCOPUS

3

초록

Bandwidth nowadays becomes increasingly limited due to a huge amount of devices and applications joining the network. Fortunately, along with the development of technology, now we are able to exploit the abundant spectrum available at the millimeter-wave (mmWave) frequency band. This new technique comes with many prospective benefits and one of them is increasing data rates. Nonetheless, one of the major challenge in this is identifying the optimal beamforming (BF) vector in a large antenna array system which is often prohibitively expensive when using an exhaustive search. In this paper, we investigate an intelligent beamforming design in a typical mmWave mMIMO system to overcome the aforementioned problem. Specifically, we propose a solution leveraging the deep q-learning model (DQL) to learn a BF pattern that maximizes the BF rate received at the user equipment. © 2022 IEEE.

키워드

Beamforming (BF); Deep Reinforment Learning (DRL); Massive Multi-input Multi-output (mMIMO); Millimeter Wave (mmWave)
제목
Intelligent Beamforming Design in mmWave mMIMO: A Reinforcement Learning Approach
저자
Hien, Pham T.T.; Nguyen, T.V.; Cho, Sungrae
DOI
10.1109/ICTC55196.2022.9952765
발행일
2022-10
유형
Conference Paper
저널명
International Conference on ICT Convergence
권
2022-October
페이지
1033 ~ 1036