基于多项式模型的非平稳信号瞬时特征提取

Polynomial Model Based on Non-stationary Transient Signal Feature Extraction

文章深入分析了非平稳信号瞬时特征的结构特点。应用Gram-Schmidt正交基对非平稳信号的瞬时频率和瞬时幅值进行近似,提出了瞬时频率和瞬时幅值的正交多项式估计模型。应用极大似然法对模型的参数进行估计,应用模拟退火算法求解参数的估计值。估计参数带入多项模型,提取出非平稳信号的瞬时特征。仿真试验证明了该方法的可行性和有效性。

The instantaneous character of non-stationary signal is analyzed. Both instantaneous amplitude （IA） and instantaneous frequency（IF） is approximated by Gram-Schmidt orthonormal base, the model is proposed for estimating IF and IA. The parameters are estimated using maximum likelihood approach and result is worked out by simulated annealing algorithm. The instantaneous characters are extracted through orthonormal polynomial. The validity of the whole method has been verified by computer simulation.