EM-Based Recursive Estimation of Spatiotemporal Correlation Statistics for Non-stationary MIMO Channel

Yousuke NARUSE; Jun-ichi TAKADA

doi:10.1587/transcom.E98.B.324

EM-Based Recursive Estimation of Spatiotemporal Correlation Statistics for Non-stationary MIMO Channel

Yousuke NARUSE, Jun-ichi TAKADA

Full Text Views

0

Share
Cite this

Summary :

We introduce a MIMO channel estimation method that exploits the channel's spatiotemporal correlation without the aid of a priori channel statistical information. A simplified Gauss-Markov model that has fewer parameters to be estimated is presented for the Kalman filter. In order to obtain statistical parameters on the time evolution of the channel, considering that the time evolution is a latent statistical variable, the expectation-maximization (EM) algorithm is applied for accurate estimation. Numerical simulations reveal that the proposed method is able to enhance estimation capability by exploiting spatiotemporal correlations, and the method works well even if the forgetting factor is small.

Publication: IEICE TRANSACTIONS on Communications Vol.E98-B No.2 pp.324-334

Publication Date: 2015/02/01

Publicized

Online ISSN: 1745-1345

DOI: 10.1587/transcom.E98.B.324

Type of Manuscript: PAPER

Category: Wireless Communication Technologies

Authors

Yousuke NARUSE
Tokyo Institute of Technology
Jun-ichi TAKADA
Tokyo Institute of Technology

Keyword

MIMO, channel estimation, Gauss-Markov model, Kalman filter, EM algorithm, posterior probability

Cite this

Copy

Yousuke NARUSE, Jun-ichi TAKADA, "EM-Based Recursive Estimation of Spatiotemporal Correlation Statistics for Non-stationary MIMO Channel" in IEICE TRANSACTIONS on Communications, vol. E98-B, no. 2, pp. 324-334, February 2015, doi: 10.1587/transcom.E98.B.324.
Abstract: We introduce a MIMO channel estimation method that exploits the channel's spatiotemporal correlation without the aid of a priori channel statistical information. A simplified Gauss-Markov model that has fewer parameters to be estimated is presented for the Kalman filter. In order to obtain statistical parameters on the time evolution of the channel, considering that the time evolution is a latent statistical variable, the expectation-maximization (EM) algorithm is applied for accurate estimation. Numerical simulations reveal that the proposed method is able to enhance estimation capability by exploiting spatiotemporal correlations, and the method works well even if the forgetting factor is small.
URL: https://globals.ieice.org/en_transactions/communications/10.1587/transcom.E98.B.324/_p

Copy

@ARTICLE{e98-b_2_324,
author={Yousuke NARUSE, Jun-ichi TAKADA, },
journal={IEICE TRANSACTIONS on Communications},
title={EM-Based Recursive Estimation of Spatiotemporal Correlation Statistics for Non-stationary MIMO Channel},
year={2015},
volume={E98-B},
number={2},
pages={324-334},
abstract={We introduce a MIMO channel estimation method that exploits the channel's spatiotemporal correlation without the aid of a priori channel statistical information. A simplified Gauss-Markov model that has fewer parameters to be estimated is presented for the Kalman filter. In order to obtain statistical parameters on the time evolution of the channel, considering that the time evolution is a latent statistical variable, the expectation-maximization (EM) algorithm is applied for accurate estimation. Numerical simulations reveal that the proposed method is able to enhance estimation capability by exploiting spatiotemporal correlations, and the method works well even if the forgetting factor is small.},
keywords={},
doi={10.1587/transcom.E98.B.324},
ISSN={1745-1345},
month={February},}

Copy

TY - JOUR
TI - EM-Based Recursive Estimation of Spatiotemporal Correlation Statistics for Non-stationary MIMO Channel
T2 - IEICE TRANSACTIONS on Communications
SP - 324
EP - 334
AU - Yousuke NARUSE
AU - Jun-ichi TAKADA
PY - 2015
DO - 10.1587/transcom.E98.B.324
JO - IEICE TRANSACTIONS on Communications
SN - 1745-1345
VL - E98-B
IS - 2
JA - IEICE TRANSACTIONS on Communications
Y1 - February 2015
AB - We introduce a MIMO channel estimation method that exploits the channel's spatiotemporal correlation without the aid of a priori channel statistical information. A simplified Gauss-Markov model that has fewer parameters to be estimated is presented for the Kalman filter. In order to obtain statistical parameters on the time evolution of the channel, considering that the time evolution is a latent statistical variable, the expectation-maximization (EM) algorithm is applied for accurate estimation. Numerical simulations reveal that the proposed method is able to enhance estimation capability by exploiting spatiotemporal correlations, and the method works well even if the forgetting factor is small.
ER -