An OFDM channel prediction method based on adaptive jump learning networks
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
万方数据
维普
信道预测是支撑变电站等电力物联网通信系统自适应传输的重要技术.为了解决过期信道状态信息降低通信系统自适应传输性能的问题,提出了一种基于自适应跳跃学习网络的信道状态信息预测方法.该方法主要包括递归微调算法和混合惩戒网络两部分.其中,前者主要用于微调学习网络的随机输入权重矩阵,后者主要通过两层惩戒网络来解决输出权重矩阵的病态解问题.由于具有oracle属性,自适应跳跃学习网络不仅具有良好的泛化能力,还可以生成稀疏性输出权重矩阵.仿真结果表明,自适应跳跃学习网络在IEEE802.11ah协议的正交频分复用(Orthogonal Frequency Division Multiplexing,OFDM)通信系统中具有良好的单步预测性能和多步预测性能.
Channel prediction is critical to the adaptive transmission of electric Internet of Things(IOT)communication systems,like substations.Since the outdated channel state information(CSI)reduces the adaptive transmission performance of wireless communication systems,this paper proposes a channel state information prediction method based on the adaptive jump learning network.The method includes two parts:the recursive fine-tuning algorithm and the hybrid regularization network.The former is used to fine-tune the random input weight matrix of the learning network,and the latter is to solve the ill-conditioned solution of the output weight matrix by a two-layer punishment network.Due to the oracle property,the adaptive jump learning network has good generalization ability,and can generate a sparse output weight matrix.The simulation results indicate that the adaptive jump learning network has good one-step and multistep prediction performances in the orthogonal frequency division multiplexing(OFDM)communication systems based on the IEEE802.11ah standard.
channel predictionOFDM systemsone-step predictionmultistep predictionIEEE802.11ah standard
万方数据
维普
