...
首页> 外文期刊>International Journal of Innovative Computing Information and Control >DETERMINATION OF MODEL STRUCTURE VIA CYCLO-STATIONARITY BASED NEURAL NETWORK
【24h】

DETERMINATION OF MODEL STRUCTURE VIA CYCLO-STATIONARITY BASED NEURAL NETWORK

机译:基于循环平稳性的神经网络确定模型结构

获取原文
获取原文并翻译 | 示例
           

摘要

Determination of model structure commonly influences the performance of system identification and model applications. It has to be performed by the data-driven methods if the priori structure information is not available, whereas the data are sometimes collected under severe experiment conditions. In this paper, a cyclo-stationarity based neural network is applied to determining the model structure through compounding information indices obtained by the output over-sampling scheme. It is illustrated that several distinct information indices on the cyclo-stationarity are detected from the experimental data. Then, different indices are compounded by a neural network to improve the determination performance. The effectiveness of the proposed approach is demonstrated through an identification experiment on a magnetic levitation system, while the performance of conventional methods degrades largely due to the severe numerical conditions.
机译:确定模型结构通常会影响系统识别和模型应用程序的性能。如果先验结构信息不可用,则必须通过数据驱动的方法来执行,而有时数据是在严格的实验条件下收集的。在本文中,基于循环平稳性的神经网络被用于通过复合由输出过采样方案获得的信息指标来确定模型结构。可以看出,从实验数据中可以检测到几种关于循环平稳性的信息索引。然后,通过神经网络复合不同的指标,以提高测定性能。通过在磁悬浮系统上的识别实验证明了该方法的有效性,而常规方法的性能由于严苛的数值条件而大大降低。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号