基于BO-LSTM神经网络的地面无人系统网络时延分析  

The Network Time Delay Analysis of Unmanned Ground System Based on BO-LSTM Neural Network

作　　者：王鑫鑫 田卫萍[1] 田野 刘超 党国龙 WANG Xinxin;TIAN Weiping;TIAN Ye;LIU Chao;DANG Guolong(North Automatic Control Technology Institute,Taiyuan 030006,China)

机构地区：[1]北方自动控制技术研究所,太原030006

出　　处：《火力与指挥控制》2024年第3期101-110,118,共11页Fire Control & Command Control

摘　　要：针对通过无线通信网络实现远程控制的地面无人系统,分析了地面无人系统在工作的过程中,网络时延对系统的影响。基于网络时延的分布特性,提出了一种贝叶斯算法(Bayesian algorithm,BO)优化的长短期记忆(long-term and short-term memory,LSTM)神经网络时延预测模型,在Matlab软件中搭建了该模型,并通过网络时延训练集数据对模型进行了训练,在网络时延测试集数据上对训练好的模型进行了测试,最后,就R2、RMSE和MAE评价指标对测试效果和GRU、BO-GRU以及LSTM预测方法进行了对比,结果表明,BO算法优化的LSTM神经网络时延预测模型RMSE和MAE评价结果更低,预测精度更高,预测效果更好,验证了提出的网络时延预测模型的有效性。According to the unmmanded ground system remote controlled by wireless communication network,the impact of network time delay on ground unmanned systems during operation is analyzed.Based on the distribution characteristics of network delay,the Bayesian algorithm(BO)optimized long-term and short-term memory(LSTM)neural network time delay prediction model is proposed,and the model is built on the Matlab software,the model is trained by network time delay training set data.The well trained model is tested on network time delay test set data.Finally,the test effects,GRU,BO-GRU and LSTM predication method are compared by the evaluation indexes of R2,,RMSE and MAE.The results show that the Bayesian algorithm(BO)optimized long-term and short-term memory(LSTM)neural network time delay prediction model RMSE and MAE have lower evaluation results,higher prediction accuracy and better prediction effect,the effectiveness of the proposed network delay prediction model is verified.

关 键 词：无线通信网络 地面无人系统 网络时延 神经网络 时延预测 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]