Research Article
Low-Complexity MMSE Signal Detection Based on the AOR Iterative Algorithm for Uplink Massive MIMO Systems
@INPROCEEDINGS{10.1007/978-3-319-72823-0_36, author={Zhenyu Zhang and Yuanyuan Dong and Zhongshan Zhang and Xiyuan Wang and Xiaoming Dai and Linglong Dai and Haijun Zhang}, title={Low-Complexity MMSE Signal Detection Based on the AOR Iterative Algorithm for Uplink Massive MIMO Systems}, proceedings={5G for Future Wireless Networks. First International Conference, 5GWN 2017, Beijing, China, April 21-23, 2017, Proceedings}, proceedings_a={5GWN}, year={2018}, month={1}, keywords={Accelerated overrelaxation (AOR) Iterative algorithm Minimum mean square error (MMSE) Convergence Complexity}, doi={10.1007/978-3-319-72823-0_36} }- Zhenyu Zhang
Yuanyuan Dong
Zhongshan Zhang
Xiyuan Wang
Xiaoming Dai
Linglong Dai
Haijun Zhang
Year: 2018
Low-Complexity MMSE Signal Detection Based on the AOR Iterative Algorithm for Uplink Massive MIMO Systems
5GWN
Springer
DOI: 10.1007/978-3-319-72823-0_36
Abstract
Massive multiple-input multiple-output (MIMO) systems can substantially improve the spectral efficiency and system capacity by equipping a large number of antennas at the base station and it is envisaged to be one of the critical technologies in the next generation of wireless communication systems. However, the computational complexity of the signal detection in massive MIMO systems presents a significant challenge for practical hardware implementations. This work proposed a novel minimum mean square error (MMSE) signal detection method based on the accelerated overrelaxation (AOR) iterative algorithm. The proposed AOR-based method can reduce the overall complexity of the classical MMSE signal detection by an order of magnitude from to , where is the number of users. Numerical results illustrate that the proposed AOR-based algorithm can outperform the performance of the recently proposed Neumann series approximation-based algorithm and approach the conventional MMSE signal detection involving exact matrix inversion with significantly reduced complexity.