Orthogonal intrinsic mode functions via optimization approach

Xinpeng Wang; Bingo Wing-Kuen Ling; Wei-Chao Kuang; Zhijing Yang

doi:10.3934/jimo.2019098

Article Contents

2021, Volume 17, Issue 1: 51-66. Doi: 10.3934/jimo.2019098

This issue Previous Article Perron vector analysis for irreducible nonnegative tensors and its applications Next Article A chance-constrained stochastic model predictive control problem with disturbance feedback

Orthogonal intrinsic mode functions via optimization approach

Faculty of Information Engineering, Guangdong University of Technology, Guangzhou, 510006, China

^* Corresponding author: Bingo Wing-Kuen Ling
^* Corresponding author: Bingo Wing-Kuen Ling

Received: January 2019

Revised: March 2019

Early access: July 2019

Published: January 2021

The first author is supported by NSF grant xx-xxxx

Abstract Full Text(HTML) Figure(5) / Table(3) Related Papers Cited by

Abstract

This paper proposes an optimization approach to find a set of orthogonal intrinsic mode functions (IMFs). In particular, an optimization problem is formulated in such a way that the total energy of the difference between the original IMFs and the corresponding obtained IMFs is minimized subject to both the orthogonal condition and the IMF conditions. This formulated optimization problem consists of an exclusive or constraint. This exclusive or constraint is further reformulated to an inequality constraint. Using the Lagrange multiplier approach, it is required to solve a linear matrix equation, a quadratic matrix equation and a highly nonlinear matrix equation only dependent on the orthogonal IMFs as well as a nonlinear matrix equation dependent on both the orthogonal IMFs and the Lagrange multipliers. To solve these matrix equations, the first three equations are considered. First, a new optimization problem is formulated in such a way that the error energy of the highly nonlinear matrix equation is minimized subject to the linear matrix equation and the quadratic matrix equation. By finding the nearly global optimal solution of this newly formulated optimization problem and checking whether the objective functional value evaluated at the obtained solution is close to zero or not, the orthogonal IMFs are found. Finally, by substituting the obtained orthogonal IMFs to the last matrix equation, this last matrix equation reduced to a linear matrix equation which is only dependent on the Lagrange multipliers. Therefore, the Lagrange multipliers can be found. Consequently, the solution of the original optimization problem is found. By repeating these procedures with different initial conditions, a nearly global optimal solution is obtained.

Keywords:

Mathematics Subject Classification: Primary: 58F15, 58F17; Secondary: 53C35.

Citation:

Full Text(HTML)

Figure 1. Original signal

Download: Full-size image PowerPoint slide

Figure 2. Signal components

Download: Full-size image PowerPoint slide

Figure 3. Normalized IMFs

Download: Full-size image PowerPoint slide

Figure 4. Components obtained by the Gram Schmidt orthogonalization method

Download: Full-size image PowerPoint slide

Figure 5. Components obtained by our proposed method

Download: Full-size image PowerPoint slide

Table 1. Orthogonal errors of all the IMFs $ err_{orth} $

Signal	Our proposed method	Gram Schmidt orthogonalization method [8]	Normalization method
$x(t)=\sin (2 \pi t) e^{-\frac{t^{2}}{2}}$ $+e^{-t} \cos (2 \pi t)$	0	0	1.3545

| Show Table

DownLoad: CSV

Table 2. Errors between the total number of the extrema and the total number of the zero crossing points of all the IMFs $ err_{ex_-zx} $

Signal	Our proposed method	Gram Schmidt orthogonalization method [8]	Normalization method
$x(t)=\sin (2 \pi t) e^{-\frac{t^{2}}{2}}$ $+e^{-t} \cos (2 \pi t)$	0	25	0

| Show Table

DownLoad: CSV

Table 3. Errors between the total number of the extrema and the total number of the zero crossing points of all the IMFs $ err_{ex_-zx} $

Signal	Our proposed method	Gram Schmidt orthogonalization method [8]	Normalization method
$x(t)=\sin (2 \pi t) e^{-\frac{t^{2}}{2}}$ $+e^{-t} \cos (2 \pi t)$	$0.9139\times10^{-4}$	$0.9801\times10^{-4}$	$0.6368\times10^{-4}$

| Show Table

DownLoad: CSV

Related Papers

Cited by

References

[1]	Q. Chen, N. Huang, S. Riemenschneider and Y. Xu, A B-spline approach for empirical mode decompositions, Advanced on Comput. Math., 24 (2006), 171-195. doi: 10.1007/s10444-004-7614-3.
[2]	L. E. Cheriyan and P. Janardhanan, Adaptive orthogonal signal decomposition based on empirical mode decomposition and empirical wavelet transform, 1st IRF International Conference, 9 (2014), 67-70.
[3]	P. Flandrin and P. Gonçalvès, Empirical mode decompositions as data-driven wavelet-like expansions, Int. J. of Wavelets, Multires. and Info. Processing, 2 (2004), 477-496. doi: 10.1142/S0219691304000561.
[4]	P. Flandrin, G. Rilling and P. Gonçalvès, Empirical mode decomposition as a filter bank, IEEE Signal Processing Letters, 11 (2004), 112-114.
[5]	C. Y. Ho, B. W. Ling, Y. Q. Liu, P. K. Tam and K. L. Teo, Optimal design of magnitude responses of rational infinite impulse response filters, IEEE Transactions on Signal Processing, 54 (2006), 4039-4046. doi: 10.1109/TSP.2006.880317.
[6]	N. E. Huang, Z. Shen, S. Long, M. Wu, H. Shih, Q. Zheng, N. Yen, C. Tung and H. Liu, The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-stationary time series analysis, Proceedings of the Royal Society London A: Math., Phys. and Engineering Sciences, 454 (1998), 903-955. doi: 10.1098/rspa.1998.0193.
[7]	T. L. Huang, W. X. Ren and M. L. Lou, The orthogonal Hilbert-Huang transform and its application in earthquake motion recordings analysis, The 14th World Conference on Earthquake Engineering, 10 (2008), 17-18.
[8]	Y. Kopsinis and S. McLaughlin, Development of EMD-based denoising methods inspired by wavelet thresholding, IEEE Transactions on Signal Processing, 57 (2009), 1351-1362. doi: 10.1109/TSP.2009.2013885.
[9]	Y. Kopsinis and S. McLaughlin, Investigation and performance enhancement of the empirical mode decomposition method based on a heuristic search optimization approach, IEEE Transactions on Signal Processing, 56 (2008), 1-13. doi: 10.1109/TSP.2007.901155.
[10]	B. W. Ling, N. L. Tian, C. Y. Ho, W. C. Siu, K. L. Teo and Q. Y. Dai, Maximally decimated paraunitary linear phase FIR filter bank design via iterative SVD approach, IEEE Transactions on Signal Processing, 63 (2015), 466-481. doi: 10.1109/TSP.2014.2371779.
[11]	P. Q. Muoi, D. N. Háo, P. Maass and M. Pidcock, Descent gradient methods for nonsmooth minimization problems in ill-posed problems, J. of Comput. and Appl. Math., 298 (2016), 105-122. doi: 10.1016/j.cam.2015.11.039.
[12]	F. Rahpeymaii, M. Kimiaei and A. Bagheri, A limited memory quasi-Newton trust-region method for box constrained optimization, J. of Comput. and Appl. Math., 303 (2016), 105-118. doi: 10.1016/j.cam.2016.02.026.
[13]	Q. S. Ren, Q. Yi and M. Y. Fang, Fast implementation of orthogonal empirical mode decomposition and its application into singular signal detection, IEEE International Conference on Signal Processing and Communications, 6 (2007), 1215-1218. doi: 10.1109/ICSPC.2007.4728544.
[14]	Z. J. Yang, B. W. Ling and C. Bingham, Joint empirical mode decomposition and sparse binary programming for underlying trend extraction, IEEE Transactions on Instrumentation and Measurement, 62 (2013), 2673-2682. doi: 10.1109/TIM.2013.2265451.
[15]	Z. J. Yang, B. W. Ling and C. Bingham, Trend extraction based on separations of consecutive empirical mode decomposition components in Hilbert marginal spectrum, Measurement, 46 (2013), 2481-2491. doi: 10.1016/j.measurement.2013.04.071.