TY - JOUR
T1 - Partial Perturbation to Alleviate the Performance Degradation of Vector Perturbation with Inaccurate Power Scaling Factors
AU - Hou, Yafei
AU - Denno, Satoshi
N1 - Publisher Copyright:
© 1963-12012 IEEE.
Copyright:
Copyright 2021 Elsevier B.V., All rights reserved.
PY - 2021/3
Y1 - 2021/3
N2 - For multi-user multiple-input-multiple-output (MU-MIMO) system, transmitter utilizes a pre-equalizer based precoding to cancel inter-stream interference for parallel transmission with the aid of accurate feedback of channel state information (CSI) from all receivers. The correct power scaling factor, which normalizes the symbols power of precoding to be a constant one, is required at each receiver. Due to CSI error and limitation of feed-forward link, the received inaccurate power scaling factor will shrink or expand the received constellation points which severely degrades the performance of MIMO system. In this paper, using nonlinear vector perturbation (VP) and linear zero-forcing (ZF) precoding, we analyze the impact of inaccurate power scaling factor on the performance of a MIMO precoding system. The analyzed results show that the mismatched modulo size from inaccurate power scaling factor severely degrades the performance of the VP system, especially for high order M-QAM modulation. In addition, the performance degradation is strongly related with the distribution of representation points for each M-QAM symbol. For linear ZF precoding, the performance loss from the shrinked or expanded constellation points is severe for system using high order M-QAM. In addition, to alleviate the performance degradation, we propose a VP system using partial perturbation points (PPP). By limiting the region of redundant points with vector perturbation for partial M-QAM symbols, the performance degradation due to the inaccurate power scaling factor can be alleviated.
AB - For multi-user multiple-input-multiple-output (MU-MIMO) system, transmitter utilizes a pre-equalizer based precoding to cancel inter-stream interference for parallel transmission with the aid of accurate feedback of channel state information (CSI) from all receivers. The correct power scaling factor, which normalizes the symbols power of precoding to be a constant one, is required at each receiver. Due to CSI error and limitation of feed-forward link, the received inaccurate power scaling factor will shrink or expand the received constellation points which severely degrades the performance of MIMO system. In this paper, using nonlinear vector perturbation (VP) and linear zero-forcing (ZF) precoding, we analyze the impact of inaccurate power scaling factor on the performance of a MIMO precoding system. The analyzed results show that the mismatched modulo size from inaccurate power scaling factor severely degrades the performance of the VP system, especially for high order M-QAM modulation. In addition, the performance degradation is strongly related with the distribution of representation points for each M-QAM symbol. For linear ZF precoding, the performance loss from the shrinked or expanded constellation points is severe for system using high order M-QAM. In addition, to alleviate the performance degradation, we propose a VP system using partial perturbation points (PPP). By limiting the region of redundant points with vector perturbation for partial M-QAM symbols, the performance degradation due to the inaccurate power scaling factor can be alleviated.
KW - MIMO
KW - dirty paper coding
KW - inaccurate power scaling factor
KW - performance analysis
KW - precoding
KW - vector perturbation
UR - http://www.scopus.com/inward/record.url?scp=85081543269&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85081543269&partnerID=8YFLogxK
U2 - 10.1109/TBC.2020.2977514
DO - 10.1109/TBC.2020.2977514
M3 - Article
AN - SCOPUS:85081543269
SN - 0018-9316
VL - 67
SP - 324
EP - 336
JO - IRE Transactions on Broadcasting
JF - IRE Transactions on Broadcasting
IS - 1
M1 - 9033967
ER -