首页> 外文期刊>Mechanical systems and signal processing >Component isolation for multi-component signal analysis using a non-parametric gaussian latent feature model
【24h】

Component isolation for multi-component signal analysis using a non-parametric gaussian latent feature model

机译:使用非参数高斯潜在特征模型进行多分量信号分析的分量隔离

获取原文
获取原文并翻译 | 示例
           

摘要

A challenge in analysing non-stationary multi-component signals is to isolate nonlinearly time-varying signals especially when they are overlapped in time and frequency plane. In this paper, a framework integrating time-frequency analysis-based demodulation and a non-parametric Gaussian latent feature model is proposed to isolate and recover components of such signals. The former aims to remove high-order frequency modulation (FM) such that the latter is able to infer demodulated components while simultaneously discovering the number of the target components. The proposed method is effective in isolating multiple components that have the same FM behavior. In addition, the results show that the proposed method is superior to generalised demodulation with singular-value decomposition-based method, parametric time-frequency analysis with filter-based method and empirical model decomposition base method, in recovering the amplitude and phase of superimposed components.
机译:分析非平稳多分量信号的一个挑战是隔离非线性时变信号,尤其是当它们在时间和频率平面上重叠时。本文提出了一种基于时频分析的解调和非参数高斯潜在特征模型相结合的框架,以隔离和恢复此类信号的成分。前者旨在消除高阶频率调制(FM),以便后者能够推断已解调的分量,同时发现目标分量的数量。所提出的方法可有效地隔离具有相同FM行为的多个组件。此外,结果表明,该方法在恢复叠加分量的幅度和相位方面优于基于奇异值分解的广义解调,基于滤波器的参数时频分析和基于经验模型分解的基础方法。 。

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