Effects

on GNSS Signals with VMD-MFDFA

AHMED, Wasiu Akande

(LB1617202)
PhD Progress Report Presentation
School of Instrumentation &
Optoelectronic Eng., Beihang University
Supervisor: Prof. Dr. WU Falin
 Outline
Introduction;
Methodology;
Merits of VMD method;
Experimental Results;
Discussion;
Conclusion
 Severe Scintillation degrades the satellite signals intensity;

It thereby resulting to failure in communication, positioning and

navigational services;

Thus, there is the need to develop an ionospheric scintillation

detection and mitigation technique
 Outline
Introduction;
Methodology;
Merits of VMD method;
Experimental Results;
Discussion;
Conclusion
 Variational Mode Decomposition method
Two scenarios are considered for S4 data:
 Synthetic data
S 4  k1    r (1)
where
r  filter  I 2

 jQ 2  is the complex random Gaussian process, k1 is the line of sight

component k1  0 ,  is the phase shift [-π, π], I and Qare the in-phase and quadrature
components of the Gaussian process, respectively.
Real S4 data from Manado station, Indonesia
 Lat. 1.340S and Long. 124.820E
VMD is non-recursive method and extract noise from the signal;
It adaptively decomposes non-stationary and non-linear signals into a
series of sub-signals (IMFs-Intrinsic Mode Function);
The IMFs is given as:
mn t   An t  cos n t  (2)
where mn t   IMF, An  t  signifies non-negative envelope and n t  is the envelope phase
Estimate the bandwidth of the decomposed signal;

min  
2
N  j    jnt 
   t    t    * mn t   e ,
mn , n   n1  t   2

N (3)
s.t. m t   x
n 1
n

where n  1, 2, ..., N, N is the number of IMFs, andmn  : m1 , ..., mN  and

n  : 1 , ...,  N  are the sets of modes and their center frequencies respectively.
 Multifractal Detrended Fluctuation Analysis (MFDFA)
 Criterion designed to detect and classify the IMFs into noisy and
noise free IMFs (sub-signals);
 Classifies signals into correlated (>0.5), uncorrelated (0.5) and
anti-correlated (<0.5);
 The RMS Fluctuation of the signal (F(n)) is given as:
1 N
F n     y   k   yn  k 
2

N k 1 (4)
y(k ) yn (k )
where is the satellite vehicle S4 index data, signifies signal local trend and k=1,2,…N
GNSS Signal Decomposition with VMD
 Model flowchart:
Initialize and then update the mode mˆ , center frequency ˆ 
1 1
n n

ˆ1

and Lagrangian multiplier n for k = 0 using Equation (5), (6) and
(7) k 1 xˆ     i  nm
ˆ ik    ˆ k   / 2
mˆ n   
1  2   n  k 2
(5)

 mˆ   d 
2

k 1
n
 k 1
n  0

 mˆ   d
k 1 2
n
0
where n = 1,2,…N (6)
  N
 (7)
ˆ k 1  ˆ k    xˆ    mˆ nk 1  
 n 1 
where and represent time step and Fourier transforms respectively.
N

 k 1 2
ˆ
m n m
ˆ k
n 2
n 1
N


2
ˆ nk
m
2
(8)
n 1

where  0 is the divergence tolerance.

 Equations (5), (6), (7) and (8) described VMD flowchart as in Fig. 1.
 Start

Compute mode: mˆ 1n ,

Center frequ ency: ˆ n1 ,
Lagrangian multiplier: ˆ 1

For iterations k = 0,
No. of modes N
Penalty term 

ˆ n and n as in
Update m
Equations (5) and (6)
No

If n > N
k=k+1
Yes

Update  based on Equ ation (7)

No If Equation (8) is satisfied

Yes

End

Figure 1: Flow chart of VMD denoising.

 Outline
Introduction;
Methodology;
Merits of VMD method;
Experimental Results;
Discussion;
Conclusion
VMD Reconfiguration of Signals
Synthetic S 4 data Manado: PRN 23, 10 April 2013
2 2

VMD-DFA VMD-DFA

1.8 1.8

1.6 1.6

1.4 1.4

1.2 1.2

DFA Threshold
DFA Threshold

1 1

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0 0
1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8
Number of IMFs Number of IMFs

(a) (b)
Figure 2: (a) Synthetic Scintillation data (b) Real Scintillation data (Manado)
VMD Adaptivity
 Decomposes signals without prior knowledge
 Unlike the wavelet transform
 Outline
Introduction;
Methodology;
Merits of VMD method;
Experimental Results;
Discussion;
Conclusion
Signal Decomposition Analysis
Synthetic S4 data Manado: PRN 23, 10 April 2013
1.8

1.2
1.6

1.4 1

1.2
0.8
1

4
4

S
S

0.8 0.6

0.6
0.4

0.4
0.2
0.2

0 0
0 10 20 30 40 50 60 09:30 10:00 12:00 14:00 16:00 18:00 19:30
Time (s) UTC Time

(b)
(a)
Figure 3. Ionospheric Amplitude Scintillation of Synthetic and real data: (a) Synthetic scintillation data
(b) Manado station scintillation data.
 Synthetic S 4 data
Synthetic S 4 data 2
2
0

4
S
0

4
S
-2
-2
IMF 1
IMF 1 0.2
2 0
0 -0.2
-2
IMF 2
IMF 2 0.2
0.5 0
0 -0.2
-0.5
IMF 3 IMF 3
0.5 0.2
0 0
-0.5 -0.2
IMF 4 IMF 4
0.2 0.2
0

S4
0
4

-0.2
S

-0.2
IMF 5 IMF 5
0.1 0.2
0 0
-0.1 -0.2
IMF 6 IMF 6
0.2 0.2
0 0
-0.2 -0.2
IMF 7 IMF 7
0.1 0.5
0 0
-0.1 -0.5
IMF 8 IMF 8
0.2 1
0 0
-0.2 -1
0 6 12 18 24 30 36 42 48 54 60 0 6 12 18 24 30 36 42 48 54 60
Time (s) Time (s)

(a) (b)
Figure 4. Decomposition of synthetic data into IMFs using VMD and CEEMD methods: (a) VMD method and
(b) CEEMD method.
 Manado: PRN 23, 10 April 2013 Manado: PRN 23, 10 April 2013
2 2
4 1 1

4
S

S
0 0
IMF 1 IMF 1
1 0.2
0.5 0
0 -0.2
IMF 2 IMF 2
0.2 0.2
0 0
-0.2 -0.2
IMF 3 IMF 3
0.2 0.2
0 0
-0.2 -0.2
IMF 4 IMF 4
0.1 0.2
0 0
4

4
S

S
-0.1 -0.2
IMF 5 IMF 5
0.1 0.2
0 0
-0.1 -0.2
IMF 6 IMF 6
0.1 0.4
0 0
-0.1 -0.4
IMF 7 IMF 7
0.1 0.5
0 0
-0.1 -0.5
IMF 8 IMF 8
0.1 0.4
0 0.2
-0.1 0
09:30 10:00 12:00 14:00 16:00 18:00 19:30 09:30 10:00 12:00 14:00 16:00 18:00 19:30
UTC Time UTC Time
(a) (b)
Figure 5. Decomposition of Manado real data into IMFs using VMD and CEEMD methods: (a) VMD method and
(b) CEEMD method.
 From Fig. 4 and 5
 Panel 1 represents the severe scintillation time series of the original
signal;
 Panels 2 to 8 show the generated IMFs (sub-signal) by VMD and
CEEMD;
 IMFs with high-frequency oscillations signify the scintillation noise
while those with low-frequency oscillations are the real signal;
 Synthetic S4 index signal values at 6s, 11s, and 54s were 1.64, 1.62,
and 1.76
 While at 13:47, 14:04, and 14:35 UTC S4 index values were 1.3, 1.2,
and 1.1, respectively;

 Indicating the presence of severe ionospheric scintillation

 GNSS signal is greatly affected by introducing error, thereby reducing

carrier to noise ratio) and integrity (accuracy);
Signal Reconfiguration Analysis
MFDFA has a decision criteria to detect ionospheric scintillation
features;

MFDFDA through its robust threshold classifies signals;

MFDFA presents the various IMFs obtained from the VMD and
CEEMD methods as shown in Fig. 6 (a) and (b).
 Synthetic S 4 data Manado: PRN 23, 10 April 2013
2.5 2.5
CEEMD-MFDFA CEEMD-MFDFA
VMD-MFDFA VMD-MFDFA

2 2

1.5 1.5

MFDFA Threshold
MFDFA Threshold

1 1

0.5
0.5

0
0

-0.5
-0.5 1 2 3 4 5 6 7 8
1 2 3 4 5 6 7 8 Number of IMFs
Number of IMFs

(a) (b)
Figure 6. Multifractal detrended fluctuation analysis (MFDFA) threshold for the decomposed IMFs of synthetic
and real scintillation data using CEEMD and VMD methods: (a) synthetic scintillation data and
(b) Manado station scintillation data
 The frequency of each IMF in CEEMD decreases as the scaling
exponent increases whereas it is the opposite in VMD;

 IMFs 1 and 2 of VMD and IMFs 3, 4, 5, 6, 7, and 8 of CEEMD had their

MFDFA threshold above 0.5 (representing the main signal);

 IMFs 1 and 2 of CEEMD and IMFs 3, 4, 5, 6, 7, and 8 of VMD were

rejected because of their high-frequency components.
 Signal Denoising
Synthetic S 4 data Manado: PRN 23, 10 April 2013
1.8
Original data
Original data
VMD-DFA
CEEMD-MFDFA 1.2 VMD-DFA
1.6 VMD-MFDFA CEEMD-MFDFA
VMD-MFDFA

1.4
1

1.2
0.8
1

S4
4
S

0.8 0.6

0.6
0.4

0.4

0.2
0.2

0 0
0 10 20 30 40 50 60 09:30 10:00 12:00 14:00 16:00 18:00 19:30
Time (s) UTC Time

(a) (b)
Figure 7. Performance comparison of VMD-DFA, CEEMD-MFDFA and VMD-MFDFA methods on amplitude
scintillation: (a) synthetic scintillation data and (b) Manado station scintillation data
 Manado: PRN 23, 10 April 2013

Original
S4
0.5

0
09:30 10:00 12:00 14:00 16:00 18:00 19:30

1 VMD-DFA
CEEMD-MFDFA
Denoised 0.5 VMD-MFDFA
S4
0

-0.5
09:30 10:00 12:00 14:00 16:00 18:00 19:30

1
VMD-DFA
CEEMD-MFDFA
Component

0.5 VMD-MFDFA
Noise

-0.5
09:30 10:00 12:00 14:00 16:00 18:00 19:30
UTC Time

Figure 8. GNSS signal amplitude scintillation denoising with VMD-DFA, CEEMD-MFDFA and VMD-MFDFA.
 Fig. 7 and 8 shows the efficient comparison of the CEEMD- and VMD-
MFDFA denoising methods in eliminating the scintillation components
from synthetic and real data;

 Ionospheric scintillation index measured in the GNSS signal by GISTM

receivers observed at 13:47, 14:04, and 14:35 (UTC) were 1.3, 1.2, and
1.1;

 VMD-MFDFA denoised (reduced) the noisy component values to 0.58,

0.75, and 0.62, respectively
 CEEMD-MFDFA reduced the noise to 0.67, 0.78 and 0.71 respectively;

 Similarly, VMD-MFDFA denoised the synthetic scintillated signal from

1.64, 1.62, and 1.76 to 0.65, 0.56, and 0.74 at 6s, 11s, and 54s,
respectively;

 whereas CEEMD-MFDFA denoised them to 1.12, 1.04, and 0.95,

respectively.
 Outline
Introduction;
Methodology;
Merits of VMD method;
Experimental Results;
Discussion;
Conclusion
Table 1. Synthetic scintillation index values before and after denoising.

Before After denoising

denoising
Time (s)
S4 CEEMD- VMD-DFA VMD-MFDFA
MFDFA

6 1.64 1.12 1.01 0.65

11 1.62 1.04 0.80 0.56

54 1.76 0.95 0.82 0.74

Table 2. Ionospheric scintillation index values for PRN 23, 10 April 2013, before and after denoising .

After denoising
UTC time Before
(hh:mm) denoising
(S4) CEEMD- VMD-DFA VMD-
MFDFA MFDFA

13:47 1.3 0.67 0.62 0.58

14:04 1.2 0.80 0.78 0.75

14:35 1.1 0.71 0.66 0.62

 Table 3. Performance evaluation of VMD-MFDFA for synthetic and real data.
Before After denoising
denoising
CEEMD- VMD-DFA VMD-MFDFA
MFDFA
Data PRN STD RMSE STD RMSE STD RMSE STD RMSE
Syn. data 0.23 0.48 0.21 0.46 0.20 0.45 0.17 0.43
10 Apr 2013 23 0.23 0.13 0.21 0.11 0.19 0.10 0.17 0.09

26 Sep 2013 12 0.24 0.10 0.22 0.08 0.20 0.07 0.10 0.05

12 Oct 2013 29 0.13 0.05 0.11 0.03 0.10 0.02 0.09 0.01
Table 4. Statistical performance evaluation for VMD-MFDFA method using Monte Carlos (MC) simulation.
Before After denoising
denoising
CEEMD- VMD-DFA VMD-MFDFA
MFDFA
Data PRN STD RMSE STD RMSE STD RMSE STD RMSE
Syn.
data 0.23 0.48 0.21 0.46 0.20 0.45 0.17 0.43
10
Apr 23 0.23 0.13 0.21 0.11 0.19 0.10 0.17 0.09
2013
26
Sep 12 0.24 0.10 0.22 0.08 0.20 0.07 0.10 0.05
2013
12
Oct 29 0.13 0.05 0.11 0.03 0.10 0.02 0.09 0.01
2013
 Outline
Introduction;
Methodology;
Merits of VMD method;
Experimental Results;
Discussion;
Conclusion
 Ionospheric scintillation contributes significantly to the GNSS
positioning errors, ranging errors and degrades GNSS signals;
The denoising performance of VMD-MFDFA and other methods were
analyzed and compared;
The VMD-MFDFA-based approach showed that the RMSE and the STD
accuracy significantly improved after denoising;
Thus, VMD-MFDFA is a more efficient and suitable decomposition
technique in estimating and mitigating the effect of ionospheric
scintillation on satellite signals;
The satellite positioning accuracy also improved.
THANKS FOR LISTENING