基于BP神经网络的高密度5G通信链路负载均衡性测试系统研究

Research on High Density 5G Communication Link Load Balancing Testing System Based on BP Neural Network

摘要:

5G通信网络具有高并发性和低延迟的特点,且网络中涉及大量的设备,包括基站、天线、中继器等。在高密度场景下,大量用户设备同时连接到基站,导致网络容量需求大大增加,会使得某些链路资源不足而其他链路则资源过剩,进而产生负载失衡和链路阻塞问题。为此,设计一种基于BP神经网络的高密度5G通信链路负载均衡性测试系统。搭建通信链路负载均衡性测试框架,设计控制器、流量监测器、网络处理器,完成系统硬件设计。构建通信链路信道传输模型,计算链路负载,应用流量监测器、网络处理器获取通信链路历史负载数值,采用BP神经网络预测下一时刻通信链路负载数值,将适应度函数和链路实际负载能力结合,当负载不均衡程度达到最小化时结束神经网络训练,获取通信链路负载均衡性测试最佳方案,从而实现高密度5G通信链路负载均衡。实验数据显示,在不同实验工况背景下,设计系统应用后获得Jain氏公平性均衡指数最大值为0.95,通信链路阻塞率最小值为5%,网络吞吐量在12Gbps-18Gbps之间。

Abstract:

5G communication networks have the characteristics of high concurrency and low latency, and a large number of devices are involved in the network, including base stations, antennas, Repeaters, etc. In high-density scenarios, a large number of user devices are connected to the base station at the same time, resulting in a large increase in network capacity requirements. As a result, some links have insufficient resources while others have excess resources, resulting in load imbalance and link congestion. Therefore, a load balancing test system of high-density 5G communication link based on BP neural network is designed. Set up the communication link load balancing test framework, design the controller, traffic monitor, network processor, and complete the system hardware design. Build a communication link channel transmission model, calculate the link load, apply traffic monitor and network processor to obtain the historical load value of the communication link, use BP neural network to predict the load value of the communication link at the next moment, combine the fitness function with the actual load capacity of the link, and terminate the neural network training when the load imbalance reaches the minimum. Obtain the best solution of communication link load balancing test, so as to achieve high-density 5G communication link load balancing. The experimental data show that under different experimental conditions, the maximum Jain"s fairness equilibrium index is 0.95, the minimum blocking rate of communication link is 5%, and the network throughput is between 12Gbps-18Gbps.