TY - JOUR
T1 - Channel Prediction for Millimeter Wave MIMO-OFDM Communications in Rapidly Time-Varying Frequency-Selective Fading Channels
AU - Lv, Changwei
AU - Lin, Jia Chin
AU - Yang, Zhaocheng
N1 - Publisher Copyright:
© 2013 IEEE.
PY - 2019
Y1 - 2019
N2 - The millimeter wave (mmWave) multiple-input multiple-output (MIMO) orthogonal frequency division multiplexing (OFDM) systems communicate at the extremely high-frequency band. In the extremely high band, the channel state information (CSI) from channel estimation will be outdated quickly, and herein, seriously degrading the system performance. In this paper, we focus on the channel prediction to obtain prior CSI in mmWave MIMO-OFDM systems. First, the mmWave MIMO-OFDM channel is categorized and represented in four domains: the array-frequency, array-time, angle-frequency, as well as angle-time. Then, for the above four domains, we investigate the effects of the channel representations on channel prediction, and analyze the mean-squared error performance as well as the computational complexity of the investigated prediction methods. We derive that the angle-time-domain prediction method achieves higher accuracy than the other three prediction techniques. In addition, we propose an enhanced angle-time-domain channel predictor by exploiting the spatial-time sparsity of the MIMO-OFDM channel to further improve the prediction accuracy. Finally, the simulation results confirm the statistical analysis and verify the superiority of the proposed predictors.
AB - The millimeter wave (mmWave) multiple-input multiple-output (MIMO) orthogonal frequency division multiplexing (OFDM) systems communicate at the extremely high-frequency band. In the extremely high band, the channel state information (CSI) from channel estimation will be outdated quickly, and herein, seriously degrading the system performance. In this paper, we focus on the channel prediction to obtain prior CSI in mmWave MIMO-OFDM systems. First, the mmWave MIMO-OFDM channel is categorized and represented in four domains: the array-frequency, array-time, angle-frequency, as well as angle-time. Then, for the above four domains, we investigate the effects of the channel representations on channel prediction, and analyze the mean-squared error performance as well as the computational complexity of the investigated prediction methods. We derive that the angle-time-domain prediction method achieves higher accuracy than the other three prediction techniques. In addition, we propose an enhanced angle-time-domain channel predictor by exploiting the spatial-time sparsity of the MIMO-OFDM channel to further improve the prediction accuracy. Finally, the simulation results confirm the statistical analysis and verify the superiority of the proposed predictors.
KW - Channel prediction
KW - MIMO-OFDM systems
KW - channel representations
KW - millimeter wave
KW - sparse channel
UR - http://www.scopus.com/inward/record.url?scp=85061790731&partnerID=8YFLogxK
U2 - 10.1109/ACCESS.2019.2893619
DO - 10.1109/ACCESS.2019.2893619
M3 - 期刊論文
AN - SCOPUS:85061790731
SN - 2169-3536
VL - 7
SP - 15183
EP - 15195
JO - IEEE Access
JF - IEEE Access
M1 - 8615988
ER -