基于改进机器学习的无人机中继通信数据调度控制研究

Research on Data Scheduling Control of UAV Relay Communication Based on Improved Machine Learning

摘要:

为解决无人机通信网络中数据调度行为中断概率过大的问题,实现对通信资源的合理分配,针对基于改进机器学习的无人机中继通信数据调度控制方法展开研究。设计基本网络架构,联合BMRC协议,设置URLLC数据链路单元,联合相关通信数据样本,求解通信中断概率的具体数值,实现对无人机中继通信网络资源的联合优化处理。分别计算时隙分配参量与带宽分配参量,并以此为基础,确定无人机中继位置,实现对中继通信资源的调度。按照机器学习算法标准,定义PCA改进特征,从而完善改进机器学习算法,再联合最优控制器闭环,实现对通信数据调度行为的控制,完成基于改进机器学习的无人机中继通信数据调度控制方法的设计。实验结果表明,改进机器学习算法作用下,随着中继数据累积量的增大,无人机通信网络中数据调度行为中断概率的最大值只能达到7.3%,有效降低了中断概率,符合合理分配通信资源的实际应用需求。

Abstract:

In order to solve the problem of excessive interruption probability of data scheduling behavior in unmanned aerial vehicle communication networks and achieve reasonable allocation of communication resources, research was conducted on data scheduling control methods for unmanned aerial vehicle relay communication based on improved machine learning. Design a basic network architecture, combine the BMRC protocol, set up URLLC data link units, combine relevant communication data samples, and solve specific values of communication interruption probability to achieve joint optimization processing of UAV relay communication network resources. Calculate the slot allocation parameters and bandwidth allocation parameters respectively, and based on this, determine the relay location of the UAV to achieve scheduling of relay communication resources. According to the machine learning algorithm standards, define PCA improvement features to improve the machine learning algorithm, and then combine the optimal controller closed-loop to achieve control of communication data scheduling behavior. Complete the design of a drone relay communication data scheduling control method based on improved machine learning. The experimental results show that under the influence of improved machine learning algorithms, as the accumulation of relay data increases, the maximum probability of interruption in data scheduling behavior in unmanned aerial vehicle communication networks can only reach 7.3%, effectively reducing the probability of interruption and meeting the practical application requirements of reasonable allocation of communication resources.