Open AccessDeep Learning Based Energy Efficient Scheme For Massive MIMO
Author(s) -
T. R. V. Anandharajan,
C. Murugalakshmi,
Bimo Adhitya,
K. Swetha
Publication year - 2019
Publication title -
international journal of engineering and advanced technology
Language(s) - English
Resource type - Journals
ISSN - 2249-8958
DOI - 10.35940/ijeat.f1338.0986s319
Subject(s) - mimo , autoencoder , computer science , beamforming , efficient energy use , base station , multipath propagation , energy consumption , deep learning , throughput , energy (signal processing) , computer engineering , scheme (mathematics) , power (physics) , electronic engineering , artificial intelligence , real time computing , telecommunications , wireless , engineering , electrical engineering , mathematics , channel (broadcasting) , mathematical analysis , statistics , physics , quantum mechanics
This paper proposes a Deep Learning Energy Efficient Scheme (DLEE) for a massive multiple input multiple output system (MIMO). Massive MIMO is deployed using large number of antennas for multiple users. The proposed DLEE, learns the relationship between spatial beamforming pattern and the power consumption in a base station. In this work, we design a novel learning method where the spatial correlation across UE antennas are taken as input feature vector and find the output labels which give us the energy efficiency in a BS. Due to multipath propagation, other methods only try to address the energy efficiency problem through the bit rate and the power required for the throughput to be efficient. This paper discusses the unsupervised algorithm DLEE which is similar to an autoencoder by combining the power consumed due to radiation pattern through beamforming and the DL framework to address the energy efficiency to an extent of 12% in a BS.