基于强化学习的多用户聚合帧长度优化方案  

Multi-User Aggregation Frame Length Optimization Scheme Based on Reinforcement Learning

作　　者：邓邱 方旭明[1] DENG Qiu;FANG Xuming(Southwest Jiaotong University,Chengdu Sichuan 611756,China)

机构地区：[1]西南交通大学,四川成都611756

出　　处：《通信技术》2023年第5期590-597,共8页Communications Technology

基　　金：国家自然科学基金(62071393);四川省应用基础研发项目(2020YJ0218)。

摘　　要：IEEE 802.11ax标准引入了正交频分多址(Orthogonal Frequency Division Multiple Access,OFDMA)技术来支持多用户帧聚合传输。在传输时,所有用户的传输时间必须对齐,对于传输时间较短的用户帧需进行比特填充。然而过多的填充比特会降低传输效率,影响系统吞吐量。为解决此问题,提出了一种基于强化学习的多用户聚合帧长度优化方案。无线接入点(Access Point,AP)通过学习和训练,根据站点的缓存状态决策出最佳的多用户聚合帧长度。仿真结果表明,所提方案能够在多用户帧聚合传输过程中,有效减少填充比特,增加传输负载,提升系统的吞吐量。The IEEE 802.11ax standard introduces OFDMA(Orthogonal Frequency Division Multiple Access)technology to support multi-user frame aggregation transmission.During transmission,the transmission times of all users must be aligned,and the user frames with shorter transmission times need to be padded with bits.However,too many padding bits will reduce transmission efficiency and affect system throughput.To address this problem,a multi-user aggregation frame length optimization scheme based on reinforcement learning is proposed.Through learning and training,the wireless AP(Access Point)decides the optimal multi-user aggregation frame length according to the buffer state of the station.Simulation results indicate that the proposed scheme can effectively reduce the padding bits,increase transmission load,and improve system throughput in the process of multi-user frame aggregation transmission.

关 键 词：帧聚合传输 802.11 ax 填充比特 强化学习 吞吐量 

分 类 号：TN929.5[电子电信—通信与信息系统]