摘要
结合状态空间描述,利用独立元分析(ICA)方法对状态变量进行估计.解析和算例验证表明,按照状态变量数目计算的ICA分量是状态变量的良好估计,非线性误差得到降低.在含噪声系统中,结合小波去噪可去除ICA无法去除的噪声,获得状态变量的估计值,显著提高信噪比;噪声可以削弱非线性引起的误差,采用先获取状态变量,后小波去噪的方法,能得到更好的状态变量估计值.
According to state space theory, the independent component analysis （ICA） is used to estimat state variables. By analyses and example calculation, independent components （ICs） are fine estimations to state variables while their number is equal, non-linear error is reduced. In system with noise, ICA removes noise in part, other noise which unable to he removed is deducted by wavelet. Signal-noise ratio （SNR） of estimations to state variables is remarkably enhanced. The noise may weaken the non-linearity error, estimations obtained by the proposed method are closer to state variable than that denoised at first.
出处
《控制与决策》
EI
CSCD
北大核心
2006年第1期88-92,96,共6页
Control and Decision
基金
国家自然科学基金项目(20206008)
广西科学基金项目(桂科基0448010)
关键词
独立元分析
小波分析
状态空间
Independent component analysis
Wavelet analysis
State space
