A method for time-frequency analysis of
electromagnetic interference signals based on
Hilbert-Huang transform
Yang Liu, Jia Yun-feng
School of Electronic and Information Engineering
Beihang University
Beijing, China
yangliu2116@126.com
Abstract—In the field of electromagnetic compatibility, the
complex electromagnetic data are nonlinear and non-stationary
signals. Traditional signal analysis methods include short-time
Fourier transform, wavelet analysis and so on. Hilbert-Huang
transform is a new kind of method to analyze non-stationary
signals. The complex electromagnetic interference signals
decompose into a number of intrinsic mode functions through
empirical mode decomposition, then do the Hilbert transform of
each function, extracts the instantaneous frequency, finally, time-
frequency spectrum of the original signals is constructed. The
results indicate the Hilbert-Huang transform can be used to
describe the nonlinear and non-stationary electromagnetic
interference signals and is a useful tool of time-frequency analysis.
Keywords—Hilbert-Huang transform, electromagnetic
interference signals, empirical mode decomposition, time-frequency
analysis
I. INTRODUCTION
The complex electromagnetic data are typical nonlinear
and non-stationary signals. Traditional electromagnetic signal
analysis methods include Fourier transform, wavelet analysis
and so on. Fourier transform is based on stationary signals, it
only shows the different frequency components, but it can not
give the time when they occur [1,2]. Therefore, time-
frequency analysis becomes a useful method, it mainly
includes short-time Fourier transform and wavelet analysis.
Short-time Fourier transform is actually windowed Fourier
transform, it usually assumes that the signal is stable within
the valid duration of the window function, but this is hardly
meet, so the short-time Fourier transform can not meet the
accuracy of the analysis. From the past decade, wavelet
transform has become one of the fast-evolving signal
processing tools. Wavelet transform is complete, orthogonal,
local and adaptive. All these are vital for forming a basis to
analyze nonlinear and non-stationary signals. However, the
resolution in frequency at a high-frequency range is poor.
In recent years, a new analysis technique called Hilbert-
Huang transform (HHT) has been proposed for analyzing
nonlinear and non-stationary signals [6]. The main difference
between HHT and all other methods is that the elementary
component is derived from the signal and is adaptive. The
method consists of two parts: the empirical mode
decomposition (EMD) and the Hilbert spectral analysis. The
key part of the method is EMD technique, with which the data
can be decomposed into a small number of intrinsic mode
functions (IMFs). The IMF can determine the instantaneous
frequency through the Hilbert transform. Therefore, the HHT
spectrum is an energy-time-frequency distribution of the
signal. This method overcomes the weakness of the Fourier
transform theory, it shows some unique advantages in practice
[4,5]. HHT is an improved signal processing method based on
Fourier transform. It is rarely used in the field of
electromagnetic compatibility, especially in the complex
electromagnetic data processing. In this paper, the HHT
method is used to analyze the time-frequency properties of
electromagnetic interference signals. According to the
characteristics of electromagnetic interference, we can
effectively avoid electromagnetic interference.
The rest of the paper is organized as follows. The HHT
method is introduced in section 2. In section 3, it shows the
EMD and Hilbert transform of the complex electromagnetic
data. Finally, in section 4, it gives the main conclusions of this
paper.
II. HILBERT-HUANG TRANSFORM
The HHT method is specially developed for analyzing
nonlinear and non-stationary data. The method consists of two
parts: the empirical mode decomposition, and the Hilbert
spectral analysis. Firstly, any data can be decomposed into a
small number of intrinsic mode functions that admit well-
behaved Hilbert transform with the EMD method. Secondly,
with the Hilbert transform, the intrinsic mode functions yield
instantaneous frequencies. The final presentation of the results
is an energy-time-frequency distribution.
A. The intrinsic mode functions and the empirical mode
decomposition
IMF in HHT is a signal that fulfills the following two
conditions:
This paper is sponsored by National Natural Science Foundation of China
NO.61371007.