论文部分内容阅读
鉴于非均匀采样复数据经验模态分解(NSBEMD)相对传统分解方法的优势和噪声的NSBEMD特性,提出了一种基于噪声辅助NSBEMD的混沌信号自适应降噪方法.该方法首先以含噪混沌信号和高斯白噪声分别为实、虚部来构造复数据并进行NSBEMD,然后根据虚部各IMF的能量来估算实部各IMF中包含的噪声能量,最后根据噪声能量的估计值对实部IMF进行奇异值分解(SVD)降噪.噪声估计实验验证了噪声能量估计方法的可行性,而Lorenz信号和太阳黑子月平均数的降噪实验则表明,相对于现有EMD降噪方法,本文方法能够进一步消除噪声,更清晰地恢复出混沌吸引子的拓扑结构.
In view of the advantage of NSBEMD and the NSBEMD of noise, a noise-assisted NSBEMD-based adaptive denoising method for chaotic signals is proposed. This method first uses the noisy chaotic signal And Gaussian white noise are respectively real and imaginary parts to construct complex data and perform NSBEMD, and then estimate the noise energy contained in each IMF of the real part according to the energy of each IMF of the imaginary part. Finally, the real IMF is estimated according to the estimation of the noise energy Singular value decomposition (SVD) noise reduction. The noise estimation experiment verifies the feasibility of the noise energy estimation method. The noise reduction experiment of Lorenz signal and the monthly average of sunspots shows that compared with the existing EMD noise reduction method, Further eliminate the noise and more clearly recover the topological structure of the chaotic attractor.