火力与指挥控制2018，Vol.43Issue(3)：25-29,5.DOI:10.3969/j.issn.1002-0640.2018.03.006
基于贝叶斯正则化的Elman-NARX神经网络飞行轨迹预测
Prediction of Aircraft Flight Trajectory Using Elman-combined Neural Network Based on Bayesian Regulation
摘要
Abstract
To overcome the problem of large errors and time wasting in prediction of flight trajectory with BP or Elman neural network,a Elman-NARX combined neural network method,based on Bayesian regulation is presented.Firstly,pilot-controlling parameters is analyzed to confirm the input of the network is improved.Then,the structure of NARX neural network,ameliorating its nonlinear and dynamic ability.Meanwhile,the network with Bayesian algorithm,saving the convergence time and improving the generalization ability.Finally,experimental results of flight trajectory prediction revealed that,compared with BP neural network,Elman neural network and Elman-NARX neural network,the method of Elman-NARX combined neural network based on Bayesian regulation has higher forecasting precision and convergence rate.Therefore,it is of great practical engineering application value.关键词
飞行轨迹预测/NARX神经网络/Elman-NARX神经网络/贝叶斯正则化Key words
flight tracking forecast/NARX neural network/Elman-NARX combined neural network/bayesian regulation分类
信息技术与安全科学引用本文复制引用
张振兴,杨任农,房育寰..基于贝叶斯正则化的Elman-NARX神经网络飞行轨迹预测[J].火力与指挥控制,2018,43(3):25-29,5.基金项目
国家自然科学基金青年基金(71501184) （71501184）
航空科学基金资助项目(20155196022) （20155196022）
