Research On Key Technologies In Massive MIMO System

Massive Multiple Input Multiple Output(MIMO) can acquire more energy efficiency, spectral efficiency, higher reliability and system robustness compared with traditional MIMO by using large numbers of antenna elements and simple signal processing approach. Due to these advantages, massive MIMO has great potential to be the key technique of future's Internet such as Internet of Things, cloud network and so on. The optimization of array deployment design, direction of arrival(DOA) estimation, antenna selection are key issues in massive MIMO, which are addressed in this thesis.The main work of the thesis is summarized as follows: 1) The optimization of massive MIMO array deployment design is studied. The correlationcoefficient measuring the spatial correlation and its affecting factor are explored. The spatialfading channel model of MIMO based on spatial correlation is built. The model is used toanalyze the impact of the correlation coefficient between two users and correlation coefficientamong multiple users on channel capacity. The correlation coefficient among multiple users isutilized to allocate the users location properly. The effect of different array deployments on thecapacity of massive MIMO is addressed. 2) The 2D-DOA estimation algorithms of massive MIMO are investigated. First, the classicalpropagator method(PM) is introduced. Because PM needs 2D spectral peak searching, rotationalinvariance based PM is proposed. The proposed algorithm does not require spectral peaksearching and has low complexity and can achieve automatically paired 2D-DOA estimation.Then reduced-dimension propagator method(RD-PM) and successive PM(S-PM) are proposed,which conduct initial estimation via rotational invariant PM and require no eigenvaluedecomposition of covariance matrix and reduce the computational complexity. RD-PM canachieve automatically paired DOA estimation with one-dimensional spectral peak searching andlow computational complexity. S-PM can also achieve automatically paired DOA estimationwith local one-dimensional spectral peak searching, which reduced the complexity. 3) The antenna selection algorithm of massive MIMO is investigated. Traditional antennal selectionalgorithm and its representative improved algorithms are introduced, including decrementalalgorithm, improved decremental algorithm, incremental algorithm, Doolttle-QR decompositionalgorithm and norm-based algorithm. The IST SATURN channel model is adopted, and theoptimization target is to improve the system channel capacity as much as possible. The change of channel capacity is observed after the change of the number of antennas in simulation. The pros and cons of various algorithms are discussed, as well as their impact on the performance of massive MIMO. The performance is measured in terms of complexity analysis, the relationship between channel capacity and SNR, the relationship between channel capacity and correlation coefficient, the relationship between channel capacity and transmit/ receive antenna number, and the cumulative distribution function curve of channel capacity.