一种利用多阶小波分解来提高修正协方差法的频率分辨率和适应噪声能力的方法

A Method of Improving Frequency Resolution of Modified Covariance Spectrum Estimation and its Abilities to Adapt Noise Using Multistage Wavelet Decomposition

与经典谱估计方法相比,修正协方差法具有更高的频率分辨率。但随着信噪比的降低,分辨率也跟随降低,甚至还会低于经典谱估计方法。本文利用Daubechies小波变换对修正协方差法进行改进,以较大幅度地提高其频率分辨率和适应噪声的能力。仿真结果表明,该方法能适用于信噪比很低的场合。

The method of modified covariance spectrum estimation （MSE） has higher frequency resolution than those of classic spectrum estimation （CSE）. But the frequency resolution drops if signal -noise -ratio （SNR） lowers and it is even much lower than CSE＇s when SNR is rather low. A new method based on Daubechies Wavelet that it can form. Keyw to improve MSE＇s frequency resolution and abilities to adapt noise. Simulation results show apply to those occasions with very low SNR.