Research On Nonbinary LDPC Code And Its Massive MIMO-Concatenated System

Nowadays, people have more and more requirements on the data rate and transmission relia-bility in wireless communication systems, while the limited bandwidth and complicated environ-ments restrict the development. Under this condition, massive multiple-input multiple-output (MI-MO) and excellent channel coding technologies have drawn more and more attention. Compared with traditional MIMO technology, massive MIMO technology can greatly increase the system ca-pacity without increasing the bandwidth and transmit power. Nonbinary low-density-parity-check (LDPC) code, due to its excellent ability in anti random and burst errors, can improve the system reliability effciently. Thus, the joint system of massive MIMO and nonbinary LDPC code, which can improve the data rate and transmission reliability simultaneously, is worth studying.The dissertation focuses on the research of nonbinary LDPC codes and massive MIMO tech-nology. The construction of low rate nonbinary LDPC codes based on multiplicative transforma-tion and the joint iteration system of nonbinary LDPC codes with massive MIMO detection have been researched.Firstly, the fundamentals of MIMO detection technology and nonbinary LDPC codes are in-troduced. For MIMO detection technology, the basic model is introduced, and the linear and nonlinear detection algorithms are presented as the theoretical basis of massive MIMO detection algorithms. For nonbinary LDPC codes, the basic concepts and methods of representations are giv-en, and the encoding and construction algorithms of quasi-cyclic codes with fast encoding structure are introduced. Also, several common decoding algorithms are presented. Besides the individu-al detection and decoding algorithms, the joint iteration system of minimum mean square error (MMSE) detection and nonbinary LDPC code is given, which is used as standard of comparisons in performance and complexity.Then, multiplicative transformation algorithm is studied. The application of multiplicative transformation algorithm in the construction of low rate nonbinary LDPC codes is mainly con-cerned. In order to obtain the best performance, the criterions of choosing the best transformation coefficients in both Galois Field and full permutation of all the symbols are proposed. Simulation results further confirm the improved performance compared with traditional construction algo-rithms. Furthermore, multiplicative transformation algorithm is applied to hybrid automatic repeat request (HARQ) and MIMO transmit diversity systems. In these two systems, the proposed algo-rithm can achieve better performance compared with conventional algorithms without increasing the complexity.Next, the joint iteration system of likelihood ascent search (LAS) detection and nonbinary LDPC code in massive MIMO system is studied. Based on the basic LAS and multistage LAS algorithm, the joint iteration system of M-LAS detection and nonbinary LDPC code in the uplink of massive MIMO system is proposed. In the joint system, the hard decisions of LDPC decoder are fed back to LAS detector as the initial vector, while the extrinsic information and the verification result of the hard decisions are fed back to decide the check candidate set. With the extrinsic infor-mation and the verification result, the size of check candidate set is reduced and the performance of the system can be improved. In the simulation results, the proposed system is compared with the joint iteration system of MMSE detection and nonbinary LDPC code, and is shown to be better suitable for massive MIMO system.Finally, the joint iteration system of belief propagation (BP) detection and nonbinary LDPC code is studied. Based on the basic BP detection algorithm and its Gaussian approximation sim-plification, the symbol domain BP detection algorithm simplified by Gaussian approximation is proposed, and it is applied in both multi-user and single user MIMO system. For the multi-user MIMO system, the joint iteration system of M-LAS detector and nonbinary LDPC decoder is stud-ied. The feedback information of LDPC decoder is improved by verification results of the output hard decisions. For the single user MIMO system, the nonbinary joint iteration system based on factor graph is proposed. In the joint system, the variable nodes in both detector and decoder are combined so that the detection node and check node can be updated simultaneously. Thus, the convergence speed of the system can be fasted and the performance can be improved compared with individually detection and decoding system and traditional joint iteration system.