Low Complexity Transmission Scheme For Massive MIMO Systems

Massive MIMO (M-MIMO) system is considered one of the key technolo?gies of next-generation SG communication. Equipped with the massive antenna arrays, M-MIMO system can achieve very high efficiency both in spectral and energy. One of the challenges of coherent M-MIMO is the acquisition of chan-nel state information (CSI) and pilot contamination. As the number of antennas at the base station grows large, the system overhead and the complexity asso-ciated with channel estimation will become too high to afford. When consid-ering low-complexity and low-overhead M-MIMO system, some noncoherent transmission schemes may be more favorable.One of the typical noncoherent detectors is the differential detection (DD), each user uses differential PSK mod-ulation and the receiver employs differential detection without explicit channel estimation.However, due to the lack of channel information, the performance of non-coherent detection by a certain constraints, how to design a reasonable and efficient algorithm to improve noncoherent detection performance is the focus of this paper.Based on this situation, the thesis studies the noncoherent detection prob-lem in the range of low complexity in M-MIMO systems. The content and innovation of the thesis are as follows.1 .The multiple symbol differential detection (MSDD) in massive MIMO systems. MSDD jointly detects multiple consecutive symbols, and it has been shown to achieve comparable performance of the coherent counterpart in many interesting scenarios. We extend traditional MSDD with single receive antenna to massive antenna array.2.Soft-input soft-output (SISO) MSDD in M-MIMO. In order to reduce the influence of channel fading?noise and interference, and further improve the system performance, we incorporate the channel codes. The decoding of powerful channel codes, such as LDPC and Turbo code, relies on soft-outputs of detector to realize iterative algorithm within channel decoding. Regarding the M-MIMO system, the existing hard-output decision metric is not suitable for powerful SISO channel decoding. To this end, a soft-output decision metric and a corresponding SISO framework is needed to further improve the performance of the noncoherent system.3.A noncoherent joint MSDD and channel decoding framework is pro-posed for M-MIMO system. Combine the factor graphs of SISO MSDD and channel code, the proposed SISO MSDD can be easily integrated with advanced channel codes. The framework enables more accurate priori information of data symbols is sent into MSDD and more accurate posteriori information of data symbols is sent into decoder, at the cost of extra iterations between SISO MSDD and SISO channel decoder.4. A noncoherent multiuser transmission scheme is proposed for M-MIMO systems without explicit channel estimation. First, a simple user scheduling al-gorithm is proposed by employing the knowledge of power space profile (PSP).We note the impact of PSP on the low-complexity MU-MSDD detector for M-MIMO, and we propose a pattern-fixed random user selection (PF-RUS)scheme to optimize the distributions of power space profile (PSP) of individual users, which alleviates the overlap of PSPs. PF-RUS is almost blind but effec-tively shapes the users' PSP to meet the favor of the proposed detector with PSP matching and SIC. Then, each output stream of the low complexity multi-user separator is fed into a noncoherent successive interference cancellation (N-SIC)processor, and finally into the joint MSDD and channel decoder. The proposed scheme with high performance bears the potential to solve the high channel estimation overhead for conventional coherent M-MIMO systems.