Dynamic Individual Channel Estimation for One-Way Relay Networks With Time-Multiplexed-Superimposed Training

In this paper, we design a time-multiplexed superimposed training (TMST) scheme to estimate the individual channels in amplify-and-forward one-way relay networks (OWRNs) under a doubly selective channel scenario, where the two-phase zero-prefixed block transmission scheme is adopted. The complex-exp...

Full description

Saved in:
Bibliographic Details
Published in:IEEE transactions on vehicular technology 2014-10, Vol.63 (8), p.3841-3852
Main Authors: Zhang, Shun, Gao, Feifei, Wang, Honggang, Pei, Changxing
Format: Article
Language:English
Subjects:
Algorithms
Amplification
Channel estimation
Channels
Convolution
Estimates
Estimation
Fourier transforms
Least squares approximations
Mathematical models
Mean square errors
Relay networks
Relays
Thermal expansion
Training
Vectors
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In this paper, we design a time-multiplexed superimposed training (TMST) scheme to estimate the individual channels in amplify-and-forward one-way relay networks (OWRNs) under a doubly selective channel scenario, where the two-phase zero-prefixed block transmission scheme is adopted. The complex-exponential basis expansion model (CE-BEM) is utilized to approximate the channel of each individual hop and results in a coefficient vector with much smaller size, called in-BEM-CV. The channel estimation of the individual channel is then converted into the estimation of in-BEM-CVs. We develop an estimation algorithm with three steps: the standard least squares (LS) estimator, the time domain or the fast Fourier transform (FFT)-based decoupler, and the iterative LS-based refiner. We also optimize the training parameters, including the number, the position, and the power allocation of the pilot clusters by minimizing the estimation mean square error (MSE), and derive one performance lower bound for the proposed algorithm. Finally, numerical results are provided to corroborate the proposed studies.
ISSN:0018-9545
1939-9359
DOI:10.1109/TVT.2014.2302435