Research On High-order QAM Signal Detection In Massive MIMO System

As one of the key technologies of 5G and next-generation communication systems,massive multiple-input multiple-output(MIMO)technology has extremely high spectral efficiency,large system capacity,and stronger reliability,meeting people’s growing communication needs.However,with the increase of modulation order and the antenna scale of base station,the traditional MIMO detection algorithm will result in poor detection performance.Therefore,it is necessary to study high-order QAM signal detection algorithms with excellent detection performance in massive MIMO systems.In this thesis,a new MIMO detection model is proposed to transform high-order QAM signal detection into binary detection.Based on this model,the Alternating Direction Method Of Multipliers(ADMM)method is used to solve the problem,and two high-order QAM signal detection algorithms are proposed.The simulation results show that the algorithm proposed has better detection performance.The main contents and innovations of this thesis are as follows:1.This thesis introduces the development background,advantages and challenges of massive MIMO system,and expounds the research significance of massive MIMO detection algorithm.By investigating the current research status of MIMO detection algorithms,several traditional massive MIMO detection algorithms are summarized,and the theory of ADMM method is briefly introduced.2.For the massive MIMO detector of high-order QAM signals,a new MIMO detection model is proposed,in which the variables of model are transformed from the transmitted signal vector into the selection matrix.Under high-order modulation,the constellation points will be mapped into binary values {0,1},and the model has fewer optimization parameters,which is convenient to adjust the optimization parameters.Based on this detection model and using the "relaxation-tightening" technique,we propose a massive MIMO detection algorithm based on quadratic penalty and ADMM,referred to as penalty-ADMM algorithm.Then we discuss the performance of the algorithm,including the computational complexity and the convergence characteristics.Finally,the performance of the proposed algorithm is verified by simulation.The simulation results show that the penalty ADMM detection algorithm has better detection performance,compared with other MIMO detectors.3.By introducing the L1 norm which is more suitable for high-dimensional data processing,sparse selection and cluster classification,the L1-ADMM algorithm is proposed to solve the problem that the penalty ADMM algorithm has a large number of iterations under high-order modulation.Firstly,we construct a new model based on the L1 norm.Then it is solved with L1-ADMM algorithm,and its computational complexity is analyzed.In addition,the detailed theory is given to prove the convergence of algorithm: 1)Stepwise differences of augmented Lagrange multipliers are bounded;2)The augmented Lagrange function decreases and has a lower bound in the iterative process;3)The proposed algorithm converges to the stationary point of original problem.Finally,simulation results show that L1-ADMM algorithm has faster iteration speed and better detection performance than penalty-ADMM algorithm under high-order modulation,and has better detection performance than other advanced detectors.