基于Newton迭代算法的低复杂度信号检测算法  

Low complexity signal detection algorithm based on Newton iterative algorithm

作　　者：刘刚[1] 娄增进 林勤华 郭漪[1] LIU Gang;LOU Zengjin;LIN Qinhua;GUO Yi(State Key Laboratory of Integrated Services Networks,Xidian University,Xi’an 710071,China)

机构地区：[1]西安电子科技大学综合业务网国家重点实验室,陕西西安710071

出　　处：《通信学报》2022年第2期109-117,共9页Journal on Communications

基　　金：国家自然科学基金资助项目(No.62171354,No.61871387,No.61861041,No.61871471)。

摘　　要：为了解决太赫兹通信系统超大规模MIMO检测计算复杂度高、收敛速度慢等问题,提出了基于Newton迭代算法的低复杂度信号检测算法。通过在Newton迭代算法中改进初始矩阵、加入步长因子,降低计算复杂度、提高收敛速度;通过加入调节因子,保证算法的稳定性、可靠性和场景适用性。仿真结果表明,相比传统算法,所提算法具有更低的计算复杂度和更快的收敛速度。当迭代次数为3次时,即可逼近MMSE算法性能。In order to solve the problems of high computational complexity and slow convergence rate of ultra-massive MIMO detection in terahertz communication system, the low complexity signal detection algorithm based on Newton iterative algorithm was proposed. By improving the initial matrix and adding step factor in Newton iteration algorithm,the complexity of the algorithm was reduced, and the convergence speed was increased. By adding adjusting factors, the stability and reliability of the algorithm were guaranteed, and the applicability of the algorithm was also increased. Simulation results show that the proposed algorithm have lower computational complexity and faster convergence speed compared with traditional schemes. When the number of iterations is three, the detection performance is close to MMSE algorithm.

关 键 词：太赫兹通信 超大规模MIMO 信号检测 步长因子 

分 类 号：TN929.5[电子电信—通信与信息系统]