面向天地融合网络的无线资源智能分配方法

Intelligent Allocation Method of Radio Resource for Integrated Space-ground Network

为满足天地融合网络全时、全域通信需求,采用认知无线电技术可实现有限频谱资源的感知与高效利用,有效缓解同频干扰问题。文章提出了一种用于天地一体认知网络的信道选择和功率调整的无线资源智能分配方法,在保证主用户服务质量的前提下最大化系统数据速率。首先,将天地融合网络建模为异质图结构,通过用户距离估计信道状态信息,并且利用图卷积网络提取和分析关键环境特征。其次,采用深度强化学习探索底层拓扑环境信息,通过试错与奖惩机制不断优化资源分配策略。仿真结果验证了所提方法的收敛性,并且证明系统数据速率能够得到显著提升。

To satisfy the full-time and full-domain communications requirements for integrated space-ground network,cognitive radio technology can be applied to achieve the perception and efficient utilization based on limited spectrum resources,and can effectively mitigate the problem of co-channel interference.In this paper,a channel selection and power adaptation method is proposed for the integrated space-ground cognitive network,which maximizes the system data rate while ensuring the quality of service for the primary user.Firstly,the communications network is modeled as a heterogeneous graph structure,estimating the channel state information based on user distances,and extracting and analyzing the critical environmental features by the graph convolutional network.Secondly,deep reinforcement learning is adopted to explore the underlying topological information and optimize resource allocation strategies through interacting with the environment.The simulation results validate strate the convergence of the proposed method and demonstrate that the system data rate can be significantly improved.