Find in Library

Empirical mode decomposition (EMD) and its variants are adaptive algorithms that decompose a time series into a few oscillation components called intrinsic mode functions (IMFs). They are powerful signal processing tools and have been successfully applied in many applications. Previous research shows that EMD is an efficient algorithm with computational complexity <inline-formula> <tex-math notation="LaTeX">$O\left ({n }\right)$ </tex-math></inline-formula> for a given number of IMFs, where <inline-formula> <tex-math notation="LaTeX">$n$ </tex-math></inline-formula> is the signal length, but its memory is as large as <inline-formula> <tex-math notation="LaTeX">$\left ({13+m_{imf} }\right)n$ </tex-math></inline-formula>, where <inline-formula> <tex-math notation="LaTeX">$m_{imf}$ </tex-math></inline-formula> is the number of IMFs. This huge memory requirement hinders many applications of EMD. A physical or physiological oscillation (PO) mode often consists of a single IMF or the sum of several adjacent IMFs. Let <inline-formula> <tex-math notation="LaTeX">$m_{out}$ </tex-math></inline-formula> denote the number of PO modes and, by definition, <inline-formula> <tex-math notation="LaTeX">$m_{Out}\le m_{imf}$ </tex-math></inline-formula>. In this paper, we will propose a low memory cost implementation of EMD and prove that the memory can be optimized to <inline-formula> <tex-math notation="LaTeX">$\left ({2+m_{out} }\right)n$ </tex-math></inline-formula> without aggravating the computational complexity, while gives the same results. Finally, we discuss the optimized memory requirements for different noise-assisted EMD algorithms.