摘要
针对高速移动场景正交频分复用(orthogonal frequency division multiplexing,OFDM)系统,提出了一种新的基于堆栈式极限学习机(extreme learning machine,ELM)的时变信道预测方法。为了捕获输入数据的深层信息,基于单隐藏层神经网络,首先利用堆栈式ELM方法从历史信道中提取信道的深层特征,并获得网络的初始输出权值。然后,为了适应信道的变化,新方法基于新构造的历史信道样本与初始的输出权值来实时更新网络的输出权值,并基于更新后的输出权值预测得到未来时刻的信道。最后,仿真结果表明,新方法较现有方法具有更高预测精度,适用于高速移动场景。
Aiming at the orthogonal frequency division multiplexing(OFDM)system under high-speed mobile scenario,a novel stacked extreme learning machine(ELM)based time-varying channel prediction method is proposed.Based on the single hidden layer neural network,to capture the deep information of the input data,the ELM method is firstly used to extract the deep features from the historical channel and obtain the initial output weight of the network.Then,to adapt to the channel changes,the proposed method updates the output weights of the network in real time based on the newly constructed historical channel samples and the initial output weights,and obtains the channel at the current moment based on the updated output weights.Finally,the simulation results shav that compared with the existing schemes,the proposed method has high prediction accuracy and is suitable for high-speed mobile scenarios.
作者
张捷
杨丽花
聂倩
ZHANG Jie;YANG Lihua;NIE Qian(College of Communication and Information Engineering, Nanjing University of Posts and Telecommunications, Nanjing 210032, China;Jiangsu Key Laboratory of Wireless Communication, Nanjing 210003, China)
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2022年第2期662-667,共6页
Systems Engineering and Electronics
基金
江苏省科技厅自然科学基金(BK20191378)
江苏省高等学校自然科学研究面上项目(18KJB510034)
第11批中国博士后科学基金(2018T110530)
国家自然科学基金(61771255)资助课题。
关键词
高速移动
正交频分复用
时变信道预测
堆栈式极限学习机
输出权值更新
high-speed mobility
orthogonal frequency division multiplexing(OFDM)
time-varying channel prediction
stacked extreme learning machine(ELM)
output weight update
