Conference proceeding
Channel Reciprocity in FDD Multiuser MIMO Systems by Super-resolution
ICC 2021 - IEEE International Conference on Communications, pp.1-6
06/2021
DOI: 10.1109/ICC42927.2021.9500263
Abstract
To fully utilize the degree of freedom of multiuser MIMO channels of wireless communication systems, channel state information at the transmitter (CSIT) is desired. From reverse link training, time division duplex (TDD) systems are able to obtain forward link CSIT because of channel reciprocity. We aim to do the same for frequency division duplex (FDD) multiuser MIMO systems, where the conventional method is to obtain CSIT from the feedback from the receiver. Unfortunately, the delay and the amount of feedback consume too much resource, especially in massive MIMO systems. However, if we could accurately estimate the channel parameters from the pilot signals in one frequency band, then the channel state in another frequency band can be calculated from these parameters. Because of the accuracy requirement, we cannot put the parameters on grids, have to treat them as continuous variables, and shall employ the super-resolution estimation algorithms. In this work, we formulate the FDD reciprocity problem as a super-resolution problem, show how to apply super-resolution algorithms to the problem in different ways, and compare the performances.
Details
- Title: Subtitle
- Channel Reciprocity in FDD Multiuser MIMO Systems by Super-resolution
- Creators
- Wanshan Yang - University of Colorado BoulderZhe Feng - University of Colorado BoulderLijun Chen - University of Colorado BoulderWeiyu Xu - University of IowaYoujian Eugene Liu - University of Colorado Boulder
- Resource Type
- Conference proceeding
- Publication Details
- ICC 2021 - IEEE International Conference on Communications, pp.1-6
- DOI
- 10.1109/ICC42927.2021.9500263
- ISSN
- 1550-3607
- eISSN
- 1938-1883
- Publisher
- IEEE
- Language
- English
- Date published
- 06/2021
- Academic Unit
- Electrical and Computer Engineering
- Record Identifier
- 9984197906402771
Metrics
19 Record Views