基于射线跟踪与离开角空间聚类的5G-R网络优化技术研究  

Research on 5G-R Network Optimization Technology Based on Ray Tracing and Departure Angle Spatial Clustering

作　　者：弓子悦 官科[1] 周敏 杨琪 葛伟涛 李珉璇 GONG Ziyue;GUAN Ke;ZHOU Min;YAGN Qi;GE Weitao;LI Minxuan

机构地区：[1]北京交通大学先进轨道交通自主运行全国重点实验室宽带移动信息通信铁路行业重点实验室,北京100044 [2]中国铁路设计集团有限公司电化电信工程设计研究院,天津300308

出　　处：《铁道通信信号》2024年第1期1-13,共13页Railway Signalling & Communication

基　　金：北京市自然科学基金海淀原始创新联合基金(L212029);中国铁路设计集团有限公司科技开发课题(2021A240802);教育部基金项目(8091B032123)。

摘　　要：当前反复“路测-调整”的传统无线网络优化方式难以满足铁路5G专用移动通信系统(5G-R)的网络优化需求。面向京沈铁路干线场景,在确定了射线跟踪传播机理模型后,进行了5G-R无线信道建模仿真,提出一种基于射线跟踪与离开角空间聚类的网络优化算法。该算法以全向天线仿真结果为基础,使用K-means++算法对射线跟踪仿真的角度-能量域数据进行聚类,将水平离开角的空间聚类中心作为扇区方位角;结合水平离开角的聚类中心与高铁行车路径的空间位置关系,计算相应扇区的下倾角;以上述基于射线跟踪与离开角空间聚类的结果为初值,基于粒子群算法进行优化迭代,高效地完成铁路干线场景下的5G-R网络优化。结果表明,在相同的计算资源和仿真条件下,基于射线跟踪与离开角空间聚类的5G-R网络优化算法对比直接使用粒子群算法,在收敛速度方面提升了约10%,在优化效果方面提升了约30%。该方法针对铁路干线场景能够实现在迭代计算次数更少的情况下,给出更好的网络优化方案,为未来建设高质量5G-R通信系统提供技术积累和参考。The current traditional wireless network optimization method of repeated“road tests and manual adjustment”is difficult to meet the network optimization requirements of railway 5G dedicated mobile communication systems(5G-R).For the scene of Beijing-Shenyang railway trunk line,after determining the ray tracing propagation mechanism model,the 5G-R wireless channel modeling and simulation are carried out,and a network optimization algorithm based on ray tracing and departure angle space clustering is proposed.Based on the omnidirectional antenna simulation results,the algorithm uses K-means++to cluster the angle-energy domain data through ray tracing simulation.The spatial cluster center of azimuth angle of departure will be used as the sector azimuth.According to the relationship between the cluster center of the horizontal departure angle and the spatial position of the high-speed train operation path,the downward inclination angle of the corresponding sector is calculated.Using the above-mentioned results based on ray tracing and departure angle space clustering as initial values,the optimization iteration is carried out based on particle swarm optimization algorithm,and 5G-R network optimization in railway trunk line scenario is efficiently completed.The results show that,under the same computing resources and simulation conditions,the 5G-R network optimization algorithm based on ray tracing and departure angle spatial clustering has improved convergence speed by about 10%and optimization effect by about 30%compared to directly using particle swarm optimization algorithm.This method can provide better network optimization solutions for railway trunk line scenarios with fewer iterations,providing technical accumulation and reference for the construction of high-quality 5G-R communication systems in the future.

关 键 词：5G-R 无线网络优化 射线跟踪 K-means++聚类算法 离开角 粒子群算法 

分 类 号：TN92[电子电信—通信与信息系统]