With the rapid increase in the number of mobile users and the continuous emergence of new wireless multimedia services, 5G mobile communication technology that can achieve spectral efficiency and data rate improvements has become a current research hotspot. This paper analyzes and summarizes the existing Internet of Things user grouping and power allocation algorithms under multi-carrier, and introduces the possible CSI (Channel State Information) error classification and modeling in the channel estimation process. For the dual-user cooperative 5G SWIPT-NOMA (Simultaneous Wireless Information and Power Transfer-Non-Orthogonal Multiple Access) transmission system with FD (Full-Duplex) antennas that exist in the presence of eavesdroppers, the physical layer security energy efficiency is used as the system performance measurement index. In view of the non-convex optimization problem and the high degree of coupling between variables, the joint solution of beamforming and power split ratio usually adopts the concave-convex procedure algorithm. This study proposes an improved particle swarm optimization algorithm, which is simpler than traditional solution methods. Aiming at the shortcomings of higher complexity of optimization algorithm, an improved particle swarm optimization algorithm with higher algorithm efficiency is proposed. The simulation results show that the optimal adaptive time slot optimization algorithm can obtain the optimal solution of the optimization problem, but the algorithm complexity is high. The improved particle swarm optimization algorithm is close to the optimal solution, the error is within the acceptable range, and the algorithm time is complicated.

随着移动用户数量的快速增长和新型无线多媒体业务的不断涌现，能够实现频谱效率和数据速率提升的5G移动通信技术成为当前的研究热点。本文分析总结了现有的多载波下物联网用户分组和功率分配算法，并介绍了信道估计过程中可能的CSI（Channel State Information）错误分类和建模。对于存在窃听者的FD（全双工）天线的双用户协作5G SWIPT-NOMA（同步无线信息和电力传输-非正交多址）传输系统，使用物理层安全能效作为系统性能衡量指标。针对非凸优化问题和变量之间的高度耦合，波束成形和功率分流比的联合求解通常采用凹凸过程算法。本研究提出了一种改进的粒子群优化算法，比传统的求解方法更简单。针对优化算法复杂度较高的缺点，提出了一种算法效率更高的改进粒子群优化算法。仿真结果表明，最优自适应时隙优化算法可以获得优化问题的最优解，但算法复杂度较高。改进后的粒子群优化算法接近最优解，误差在可接受范围内，算法时间复杂。