Pilot assignment and channel estimation via deep neural network

Seunghwan Lee, Hyungyu Ju, Byonghyo Shim

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

In orthogonal frequency division multiplexing (OFDM) systems, channel estimation is by far the most important operation in the receiver to ensure the accurate detection and decoding. Over the years, pilot-aided channel estimation has been widely used for this purpose. In open-loop systems, since there is no feedback link between the transmitter and receiver, an approach based on the equi-spaced pilot assignment has been widely employed. In this paper, we propose a closed-loop non-uniform pilot allocation strategy based on deep neural network (DNN) technique. From the numerical evaluations, we show that the proposed autoencoder-based pilot allocation technique outperforms conventional approaches by a large margin, demonstrating its ability to learn the statistical characteristics of the wireless channel.

Original languageEnglish
Title of host publication2018 24th Asia-Pacific Conference on Communications, APCC 2018
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages454-458
Number of pages5
ISBN (Electronic)9781538669280
DOIs
Publication statusPublished - 2019 Feb 1
Event24th Asia-Pacific Conference on Communications, APCC 2018 - Ningbo, China
Duration: 2018 Nov 122018 Nov 14

Publication series

Name2018 24th Asia-Pacific Conference on Communications, APCC 2018

Conference

Conference24th Asia-Pacific Conference on Communications, APCC 2018
CountryChina
CityNingbo
Period18/11/1218/11/14

Fingerprint

Channel estimation
neural network
recipient
open system
Orthogonal frequency division multiplexing
Telecommunication links
Decoding
Transmitters
Feedback
ability
evaluation
Deep neural networks

ASJC Scopus subject areas

  • Computer Networks and Communications
  • Signal Processing
  • Communication

Cite this

Lee, S., Ju, H., & Shim, B. (2019). Pilot assignment and channel estimation via deep neural network. In 2018 24th Asia-Pacific Conference on Communications, APCC 2018 (pp. 454-458). [8633453] (2018 24th Asia-Pacific Conference on Communications, APCC 2018). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/APCC.2018.8633453

Pilot assignment and channel estimation via deep neural network. / Lee, Seunghwan; Ju, Hyungyu; Shim, Byonghyo.

2018 24th Asia-Pacific Conference on Communications, APCC 2018. Institute of Electrical and Electronics Engineers Inc., 2019. p. 454-458 8633453 (2018 24th Asia-Pacific Conference on Communications, APCC 2018).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Lee, S, Ju, H & Shim, B 2019, Pilot assignment and channel estimation via deep neural network. in 2018 24th Asia-Pacific Conference on Communications, APCC 2018., 8633453, 2018 24th Asia-Pacific Conference on Communications, APCC 2018, Institute of Electrical and Electronics Engineers Inc., pp. 454-458, 24th Asia-Pacific Conference on Communications, APCC 2018, Ningbo, China, 18/11/12. https://doi.org/10.1109/APCC.2018.8633453
Lee S, Ju H, Shim B. Pilot assignment and channel estimation via deep neural network. In 2018 24th Asia-Pacific Conference on Communications, APCC 2018. Institute of Electrical and Electronics Engineers Inc. 2019. p. 454-458. 8633453. (2018 24th Asia-Pacific Conference on Communications, APCC 2018). https://doi.org/10.1109/APCC.2018.8633453
Lee, Seunghwan ; Ju, Hyungyu ; Shim, Byonghyo. / Pilot assignment and channel estimation via deep neural network. 2018 24th Asia-Pacific Conference on Communications, APCC 2018. Institute of Electrical and Electronics Engineers Inc., 2019. pp. 454-458 (2018 24th Asia-Pacific Conference on Communications, APCC 2018).
@inproceedings{414bf947ca8b44f1bc4ead5a5c615004,
title = "Pilot assignment and channel estimation via deep neural network",
abstract = "In orthogonal frequency division multiplexing (OFDM) systems, channel estimation is by far the most important operation in the receiver to ensure the accurate detection and decoding. Over the years, pilot-aided channel estimation has been widely used for this purpose. In open-loop systems, since there is no feedback link between the transmitter and receiver, an approach based on the equi-spaced pilot assignment has been widely employed. In this paper, we propose a closed-loop non-uniform pilot allocation strategy based on deep neural network (DNN) technique. From the numerical evaluations, we show that the proposed autoencoder-based pilot allocation technique outperforms conventional approaches by a large margin, demonstrating its ability to learn the statistical characteristics of the wireless channel.",
author = "Seunghwan Lee and Hyungyu Ju and Byonghyo Shim",
year = "2019",
month = "2",
day = "1",
doi = "10.1109/APCC.2018.8633453",
language = "English",
series = "2018 24th Asia-Pacific Conference on Communications, APCC 2018",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "454--458",
booktitle = "2018 24th Asia-Pacific Conference on Communications, APCC 2018",

}

TY - GEN

T1 - Pilot assignment and channel estimation via deep neural network

AU - Lee, Seunghwan

AU - Ju, Hyungyu

AU - Shim, Byonghyo

PY - 2019/2/1

Y1 - 2019/2/1

N2 - In orthogonal frequency division multiplexing (OFDM) systems, channel estimation is by far the most important operation in the receiver to ensure the accurate detection and decoding. Over the years, pilot-aided channel estimation has been widely used for this purpose. In open-loop systems, since there is no feedback link between the transmitter and receiver, an approach based on the equi-spaced pilot assignment has been widely employed. In this paper, we propose a closed-loop non-uniform pilot allocation strategy based on deep neural network (DNN) technique. From the numerical evaluations, we show that the proposed autoencoder-based pilot allocation technique outperforms conventional approaches by a large margin, demonstrating its ability to learn the statistical characteristics of the wireless channel.

AB - In orthogonal frequency division multiplexing (OFDM) systems, channel estimation is by far the most important operation in the receiver to ensure the accurate detection and decoding. Over the years, pilot-aided channel estimation has been widely used for this purpose. In open-loop systems, since there is no feedback link between the transmitter and receiver, an approach based on the equi-spaced pilot assignment has been widely employed. In this paper, we propose a closed-loop non-uniform pilot allocation strategy based on deep neural network (DNN) technique. From the numerical evaluations, we show that the proposed autoencoder-based pilot allocation technique outperforms conventional approaches by a large margin, demonstrating its ability to learn the statistical characteristics of the wireless channel.

UR - http://www.scopus.com/inward/record.url?scp=85062872413&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85062872413&partnerID=8YFLogxK

U2 - 10.1109/APCC.2018.8633453

DO - 10.1109/APCC.2018.8633453

M3 - Conference contribution

T3 - 2018 24th Asia-Pacific Conference on Communications, APCC 2018

SP - 454

EP - 458

BT - 2018 24th Asia-Pacific Conference on Communications, APCC 2018

PB - Institute of Electrical and Electronics Engineers Inc.

ER -