For Full-Text PDF, please login, if you are a member of IEICE,|
or go to Pay Per View on menu list, if you are a nonmember of IEICE.
Accelerated Widely-Linear Signal Detection by Polynomials for Over-Loaded Large-Scale MIMO Systems
Qian DENG Li GUO Chao DONG Jiaru LIN Xueyan CHEN
IEICE TRANSACTIONS on Communications
Publication Date: 2018/01/01
Online ISSN: 1745-1345
Type of Manuscript: PAPER
Category: Antennas and Propagation
over-loaded large-scale MIMO, low-complexity, Chebyshev acceleration, widely-linear detection,
Full Text: PDF(2MB)>>
In this paper, we propose a low-complexity widely-linear minimum mean square error (WL-MMSE) signal detection based on the Chebyshev polynomials accelerated symmetric successive over relaxation (SSORcheb) algorithm for uplink (UL) over-loaded large-scale multiple-input multiple-output (MIMO) systems. The technique of utilizing Chebyshev acceleration not only speeds up the convergence rate significantly, and maximizes the data throughput, but also reduces the cost. By utilizing the random matrix theory, we present good estimates for the Chebyshev acceleration parameters of the proposed signal detection in real large-scale MIMO systems. Simulation results demonstrate that the new WL-SSORcheb-MMSE detection not only outperforms the recently proposed linear iterative detection, and the optimal polynomial expansion (PE) WL-MMSE detection, but also achieves a performance close to the exact WL-MMSE detection. Additionally, the proposed detection offers superior sum rate and bit error rate (BER) performance compared to the precision MMSE detection with substantially fewer arithmetic operations in a short coherence time. Therefore, the proposed detection can satisfy the high-density and high-mobility requirements of some of the emerging wireless networks, such as, the high-mobility Internet of Things (IoT) networks.