kumoh national institute of technology
Networked Systems Lab.

Rizki Rivai Ginanjar, Jae-Min Lee, Dong-Seong Kim, "Autoencoder for Low-Complexity CSI Reconstruction in Cellular Network", 2019 KICS Winter Conference,pp.503-504, January 23-25, 2019, Yongpyeong, Korea(N8)
By :
Date : 2018-12-28
Views : 315

In this paper, a novel integration of deep learning approach in the cellular network system is explained. The deep learning technology can be used to solve complex problem of the cellular network system such as channel state information (CSI) reconstruction process. In this paper, a novel deep neural network-based CSI reconstruction system using autoencoder is introduced. By utilizing this approach, the accuracy and processing time of the CSI reconstruction system can be improved to support reliable cellular network system.


Q1: Can we use this system for the Multi user MIMO cellular system?
A1: Yes. It is possible as the capability of the integration of the deep learning technologies which can process huge amount of data.

Q2: is the channel capacity analysis can be performed on this system?
A1: No, because the only focus of this system is on the compression and decompression technique of the channel state information (CSI) without considering any channel properties of the cellular communication system