Doppler Shift Estimation for High-Speed Railway
Wireless Communication Systems with
Large-Scale Linear Antennas
Dian Fan
†
, Zhangdui Zhong
†
, Gongpu Wang
†
, and Feifei Gao
‡
†
School of Computer and Information Technology, Beijing Jiaotong University, Beijing, China.
‡
Tsinghua National Laboratory for Information Science and Technology, Beijing, China.
Email:
†
fandian@bjtu.edu.cn,
†
zhdzhong@bjtu.edu.cn,
†
gpwang@bjtu.edu.cn,
‡
feifeigao@ieee.org.
Abstract—This paper investigates the Doppler shift estimation
problem for high speed railway (HSR) wireless communication
systems with massive linear receiving antennas. Firstly, a
theoretical channel model is built up with channel coefficient,
direction of arrival (DOA) and Doppler shift taken into account.
Secondly, a new Doppler shift estimation algorithm is proposed
through reconstructing the received signal based on Discrete
Fourier Transform (DFT). Moreover, a phase rotation algorithm
is presented to improve the estimation accuracy. It is found
that the suggested estimation algorithm works well and also
obtains DOA estimate. Finally, simulation results are provided
to corroborate our proposed studies.
Index Termes−High-speed railway, Massive MIMO, Doppler
Shift estimation, DFT.
I. INTRODUCTION
Recently, the development of high-speed railway (HSR) is
rapid in the world and the maximum train moving speed
has reached nearly 575 km/h [1]. High train speed can in-
crease difficulty in reliable high data transmission. Thus the
issue of providing high rate transmission service has attracted
widespread attention.
One key factor to influence the transmission rate is the
accuracy of channel estimation. When a HSR travels at a
speed of around 500 km/h, the wireless channels suffer from
a high Doppler shift and it will increase the difficulty of the
channel estimation. In high-mobility environments, wireless
channels are both fast time-varying and frequency selective,
also known as doubly selective channels [2, 3]. As the quality
of channel estimation has a major impact on the overall
system performance and Doppler shift is a one of the key
parameters of channel estimation, it is necessary to investigate
reliable Doppler shift estimation methods in high mobility
environments.
Doppler shift estimation for fast time-varying channels
has been extensively studied in the literatures, and various
time-varying channel models have been established. There
are four major kinds of Doppler shift estimation methods:
level crossing rate (LCR) [4], maximum likelihood function
[5], correlation function [6] and based on channel estimates
[7]. Authors in [4] proposed an improved level crossing rate
estimation method. This method is composed of a Doppler
adaptive noise suppression procession and a conventional LCR
estimator. The improvement on the accuracy of the estimation
is due to the increase of the estimator complexity. Authors in
[5] presented an estimator based maximum likelihood function
and it has high accuracy of estimation. However, it has high
computation complexity and can not be realized in practical
systems. In [6], the autocorrelation of the cyclic prefix (CP)
of OFDM system is used to estimate the Doppler shift.
The estimated Doppler shift is often overestimated when the
Doppler shift is small, especially at low SNR. The reference
[7] proposed a Doppler shift estimation method based on
channel estimates, which is effective only low Doppler shift.
The another way to achieve high rate transmission in
HSR wireless communication systems is using massive MI-
MO technology. Massive MIMO systems [8] have drawn
considerable interests from both academia and industry. It
is also shown in [9] that each antenna element of a very
large MIMO system will consume low power and the total
power can be made inversely proportional to the number of
antennas. Other advantages, such as high spectral efficiency,
security, robustness and reliable linkage, also play key roles
in promoting massive MIMO systems more appealing for the
next generation wireless systems [10].
However, utilizing massive MIMO technology in HSR
wireless communication systems may result in one challenge:
the high Doppler shift will degrade the system performance.
Therefore, the Doppler shift should be estimated and compen-
sated.
The main contributions of this paper are summarized as
follows. The theoretical channel model is built up with channel
coefficient, direction of arrival (DOA) and Doppler shift taken
into account for HSR wireless communication systems with
massive linear receiving antennas. Then we proposed a novel
method of Doppler shift estimation and it is based on the Dis-
crete Fourier Transformation (DFT) to reconstruct the received
signal. Moreover, a phase rotation algorithm was suggested to
improve the estimation accuracy. Finally, simulation results are
provided to corroborate our proposed studies.
The rest of this paper is organized as follows. First, we
describe a HSR theoretical channel model with massive linear
receiving antennas in Section II. In Section III, we propose