Font Size: a A A

Reseach On Matrix Inversion In Massive MIMO Systems

Posted on:2017-01-14Degree:MasterType:Thesis
Country:ChinaCandidate:S S FengFull Text:PDF
GTID:2308330485988491Subject:Communication and Information System
Abstract/Summary:PDF Full Text Request
As the rapid development of smart cellphones and the commercialize of the fourth generation mobile communication(4G), the way people communicate has been greatly changed, which makes a claim for higher communication channel capacity and transmission rate in the next generation of mobile communication network. As a core technology of LTE, MIMO technology makes it possible to exponentially enhance the channel capacity without any increase of spectrum resources and transmission power. In these systems, more degrees of freedom in terms of data rate and link reliability are available due to the increase in the number of transmit and receive antennas. In consideration of its nice satisfying with the demand of spectral efficiency and energy efficiency, Massive MIMO technology is promised to be a good choice for the next generation of mobile communication network.Those promised benefits of Massive MIMO systems come at the cost of significantly increased computational complexity in the base-station(BS), data detection and precoding are expected to be among the most critical tasks in terms of complexity and power consumption. Zero-Forcing(ZF) has been considered as one of the potential precoding and linear detection method for Massive MIMO systems, resulted by the tens or even hundreds of antennas at the BS side, the dimension of matrices involved in it increases drastically. Being an indispensable part of precoding and linear detection, the matrix inversion suffers a lot from the huge matrix size of massive MIMO, and therefore becomes inefficient for realization. It is of great significance to simplify the design of matrix inversion and reduce the complexity for the development of mobile communication systems.Matrix inversion can be divided into two categories, outright inversion and Matrix Inverse Approximation(MIA). Outright inversion in MIMO systems are commonly decomposed by Cholesky decomposition, LU factorization or QR decomposition at first, and then inverse the triangular matrix or unitary matrix. As the presence of hundreds of antennas in Massive MIMO systems will increase the computational complexity of matrix inversion by orders of magnitude. In this case, by achieving good tradeoff between complexity and performance, Neumann Series has been considered for the Matrix Inverse Approximation(MIA), because of its suitability for Massive MIMO systems and its advantages in hardware implementation.In this paper, we focused on the implementation and complexity analysis of various matrix inversion algorithms. Firstly, we introduced the research status of Massive MIMO systems and matrix inversion, then three conventional outright inversion algorithms were analyzed. In the third chapter we described the effects of the ratio of transmit and receive antennas on the performance of NS-based MIA. We proof the feasibility of NS-based MIA by contrasting the bit error rate(BER) and signal interference ratio(SIR) with outright inversion. Additionally, methods and corresponding complexities for fast matrix inversion updates are studied. Results are evaluated numerically in terms of bit error rate using the Neumann series approximation as the initial inverse matrix. An efficient hardware architecture for approximate matrix inversion is proposed in Chapter 4, this architecture is hardware efficient and suitable for various application with different approximation precisions. Neumann series based approximation inversion turn out to be the best choice of Massive MIMO system.
Keywords/Search Tags:Massive MMO, Matrix Inverse Approximation, Neumann series, complexity
PDF Full Text Request
Related items