Exeley Inc.

Xiangchu, Li

wei, Zhang

International Journal of Advanced Network, Monitoring and Controls          Volume 03, No.04, 2018 
35 
Solutions for Governance and Suppression of Power Harmonic in Cities 
Li Xiangchu 
Hunan Nonferrous Metals Vocational and  
Technical College 
Zhuzhou, Hunan, 412006, China 
e-mail: hnysjxjwk@126.com 
 
Zhang wei 
Hunan Nonferrous Metals Vocational and  
Technical College 
Zhuzhou, Hunan, 412006, China 
 
 
Abstract—Based on the current status of cities’ power sup
ply and distribution system, this paper conducts theoretica
l and experimental investigations over the generation, harz
ards as well as the restraining and suppression techniques
 of power harmonics. A solution of scheme selection for c
ontrolling and suppressing power harmonics suitable for t
he current power supply and distribution system of big a
nd medium-sized cities is proposed. This study not only p
rovides important implications for big and medium-sized c
ities, but also provides substantial reference value for cont
rolling and suppressing power harmonics in the public util
ity system in China. 
Keywords-Power Harmonic; Governance and Suppression; 
Solution 
I. INTRODUCTION  
With the transformation and upgrading of industrial cities, 
the application of nonlinear load like converter, rectifier and 
inverter that use power semiconductor-cored insulated gate 
bipolar transistor (IGBT) and intelligent power module (IPM) 
has been widely used. While plenty of high power nonlinear 
converter equipments have their advancement and reliability 
in energy saving and controlling technology, its inherent 
characteristics will generate a large amount of harmonic, 
causing voltage and current waveform distortion in power 
supply system. [12] The consequence is that the power 
quality of utility grid is not up to standard or even worse, and 
the normal running and safety of utility grid and electrical 
equipment is badly influenced. Therefore, researches on the 
restraining and suppressing of power harmonics in urban 
utility grid have sound theoretical and practical values over 
improving power quality so as to better service the 
transformation and upgrading of the city. 
II. CAUSE OF POWER HARMONIC 
Power harmonic in city utility grid is mainly produced in 
the nonlinear load power equipment such as impact load, 
asymmetry load and the nonlinear load, with nonlinear load 
as the most important factor to produce electric power 
harmonic. [9] Inverter, uninterruptible power supply (UPS) 
and power converter equipment (such as rectifier and 
inverter), industrial computer and other high power nonlinear 
load are  the harmonic source of utility grid. [7] With the 
transformation and upgrading of the inverter control in 
industrial control and other fields, the application of 
frequency converter is undergoing a wide spread. Therefore, 
inverter has become the main source of electric power 
harmonic in urban public grid. 
III. HAZARD OF POWER HARMONICS 
The harm caused by harmonic pollution of urban public 
power grid is reflected in many aspects. The main contents 
are as following. 
1) Reducing the utilization rate of electrical equipment 
makes transformer, motor, power capacitors and other 
electrical equipments, as well as cables, Low pressure 
neutral line, busbar and other conductors working under the 
state of overload operation such as vibration, fever, abnormal 
sound, and etc. It shortens the service life of the electric 
power and increase electric energy loss.[13] 
2) Interfering with relay protection, automatic devices 
and computer systems; Making the precision electronic 
equipment work abnormal, or even burn; Increasing the error 
of measuring and measuring instruments.[4] 
DOI: 10.21307/ijanmc-2019-019
International Journal of Advanced Network, Monitoring and Controls          Volume 03, No.04, 2018 
36 
3) Reducing the quality of signal transmission and 
interfering communication system. Usually harmonics of 
2000~5000Hz generates communication noise, while 
harmonics above 5000Hz lead to the disoperation of the 
signal of the telephone circuit. [8] 
IV. GOVERNANCE AND SUPPRESSION OF POWER 
HARMONIC 
In engineering, the governance and suppression of power 
harmonic in city utility grid is mainly divided into the 
following three types:  
1) Series detuning reactor harmonic resonance amplifier 
is the main purpose of the measures to prevent because of 
reactive power compensation device (such as electric power 
capacitor, etc.) access to enlarge the excessive power 
harmonic and resonance occurs, but smaller filtering effect. 
[2-3] 
2) Using passive power filter (PF) for harmonic 
governance. The passive power filter (PF) is a filter circuit 
that makes use of the combination of inductance, capacitance 
and resistance to filter out one or multiple harmonics. It Is 
currently the basic ways of city utility grid management and 
power harmonics restraining. Despite of its merits such as 
simple structure, low cost, reliable running and low-cost 
operating, still the harmonic governance effect is not ideal, 
and could lead to new problems such as oscillation of power 
supply and distribution system and harmonic amplification. 
[14] 
3) Using active power filter (APF) for harmonic 
governance. The active power filter (APF) is a new type of 
special equipment based on modern power electronic 
technology and digital signal processing technology to 
govern electric power harmonics.[11] The basic principle of 
harmonic elimination is electricity generated during runtime 
equals the current amplitude of power harmonic, reversal 
polarity of harmonic current into the power supply system 
and compensate or offset the electric power harmonic current, 
and take the initiative to eliminate electric power harmonic. 
It has the merits of high control precision, fast response, 
good harmonic elimination effect and etc. Active power filter 
(APF) is a new research that enjoys a promising prospect in 
the field of future electric power harmonic governance and 
comprehensive optimization of power quality. 
V. SOLUTION FOR GOVERNANCE AND SUPPRESSION OF 
POWER HARMONIC IN CITIES 
A. About General Solution  
With increasingly attention paid to the governance of 
electric power harmonic, related industries get fast 
development during the period of “China's 13th Five-Year 
Plan”. There are a large variety of measures and products for 
the governance of electric power harmonic. Establishing 
reasonable selection schemes can guide technical personnel 
to choose measures and products according to the actual 
circumstance of electric power harmonic in power supply 
and distribution system. [15] Based on the harmonic voltage 
limit and harmonic current allowable value regulated in the 
utility grid national standard GB/T14549-1993 Utility Grid 
Harmonic Power Quality, [1] this paper sets the nominal 
voltage of 0.38 kV for example, and establishes a model 
selection scheme for the power harmonic governance and 
suppression of urban utility grid, as is shown in figure 1. 
 
 
 
 
 
 
 
 
International Journal of Advanced Network, Monitoring and Controls          Volume 03, No.04, 2018 
37 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 1. Model Selection for Power Harmonic Governance and Suppression 
 
B. Specific Solution for Power Harmonic Governance and 
Suppression 
By using technological upgrading projects as engineering 
case, this paper conducts researches on solutions for power 
harmonic governance and suppression. 
1) About technological upgrading projects: A utility grid 
in Hunan is mainly composed of one transformer S9-1250, 
equipped with 480 kvar PGJ type of reactive power 
compensation device. Load is composed of 2 sets of 200 kva 
UPS,  switching power supply and air conditioning. When 
the system runs, output cable of the transformer reaches 45 
degrees centigrade, cable vibration and electronic equipment 
interference become bigger. [10] Customer’s technological 
upgrading requirements are as following: a) lowering the 
temperature rise of the cable and eliminating cable vibration; 
b) reducing interference of electronic devices. 
2) Intended Solutions: Based on customer’s technological 
upgrading requirements, the project group did a lot of 
resourcing work and field investigation to make sure the 
design rationality of this power supply and distribution 
system. Field test shows the distortion rate of voltage total 
harmonic was 7.8%, the distortion rate of current total 
harmonic was 68.9%. UPS and six pulse rectifier  are the 
main nonlinear load that generate harmonics. Comparative 
analysis of these data proves that this project is under serious 
harmonic pollution and hazard. According to Figure 1, the 
choice of active power filter can effectively suppress the 
electric power harmonic. Detailed solutions are as following: 
a) retaining reactive power compensation device PGJ type 
480 kvar; b) one set of LEAPF4200-0.4 active power filter 
installed into two sets of UPS respectively. 
3) Verification: After installation and debugging, voltage, 
current waveform and harmonic analysis of the power supply 
and distribution system are tested on the spot, and data 
before and after the power harmonic control and suppression 
are shown in Figure 2 and Figure 3. 
Harmonic pollution is 
slight    
International Journal of Advanced Network, Monitoring and Controls          Volume 03, No.04, 2018 
38 
       
a) Voltage and current waveform-before                      b) Voltage and current waveform-after 
       
c) Voltage harmonic analysis-before                          d) Voltage harmonic analysis-after  
Figure 2. Voltage Waveform and Harmonic Analysis Before and After Power Harmonic Governance and Suppression 
       
a) Current Rms and waveform-before                        b) Current Rms and waveform-after 
         
c) Current harmonics analysis-before                         d) Current harmonics analysis-after 
Figure 3. Current Waveform and Harmonic Analysis Before and After Power Harmonic Governance and Suppression 
International Journal of Advanced Network, Monitoring and Controls          Volume 03, No.04, 2018 
39 
It can be seen from Figure 2 Figure 3 that after harmonic 
control and suppression, the total harmonic distortion rate of 
the power supply and distribution system is reduced to 2.1%; 
the total harmonic distortion rate of current is reduced to 
8.8%; the temperature in output-end cable of the transformer 
is 37 degree centigrade when room temperature is 26 degree 
centigrade; the cable vibration and electronic equipment 
interference are eliminated. [5-6] After field tests and 
function verifications, the solution above turns out to be 
reasonable and achieves the intended goals. 
VI. CONCLUSION  
Based on the harmonic voltage limit and harmonic 
current allowable value regulated in the utility grid national 
standard GB/T14549-1993 Utility Grid Harmonic Power 
Quality, a model selection scheme for the power harmonic 
governance and suppression of urban utility grid is 
established after theoretical and experimental studies on the 
causes, hazards as well as the governance and suppression 
technology. By using technological upgrading projects as  
engineering case, this paper proposes a solution that uses 
active power filter (APF) as the core device. After field test 
and functional verification, the scheme proves to be both 
reasonable and practical. It is a reference for the power 
supply and distribution system design of urban utility grid. 
REFERENCES 
[1] GB/T14549-1993, Quality of electric energy supply Harmonics in 
public supply network [S]. 
[2] Pan Zhaodong. Design of single-phase harmonic control system in 
mine [J]. Industrial and mining automation，2016.1. 
[3] Chen Xuemei. Analysis and management of frequency conversion 
harmonics of low voltage power grid in oil field [J] .Electrical 
applications，2016.3. 
[4] Li Lanfang. Harmonic analysis and hazard management of coal mine 
variable frequency speed control system [J] .Science and technology 
of Coal，2015.12. 
[5] Li Honghui. Research on harmonic management of coal mine dc hoist 
system [J] .Coal project engineering，2015.12. 
[6] Zhang Zhicheng. Research on harmonic management method of 
isolated island micro - grid [J] .power electronics technology，
2015.12. 
[7] Jiang Youhua. Decoupling and stability optimization of 
multi-harmonic source control system with tree distribution [J] .The 
grid technology，2015.3. 
[8] Ge Shaoyun. An empirical method for harmonic loss reduction of 
distribution network [J] .Journal of electrical systems and 
automation，2015.3. 
[9] Wang Huiwu. Power harmonic detection and estimation based on 
nonlinear theory [J] .Electrical measurement and meter，2016.3. 
[10] Li Demin. Power harmonic analysis based on FFT interpolation of 
four spectral line of Nuttal window [J] .Power system protection 
pre-control，2016.3. 
[11] Chen Dongyi. New parallel active power filter with current control 
[J] .Experimental technology and management，2016.2. 
[12] Zheng Kang. Discussion and application of harmonic wave and its 
control technology in mine power supply and distribution system 
[J] .The information of Power，2014.3.  
[13] Sun Yunfei. Research and application of harmonic control technology 
in Jiao-jia gold mine [J] .Electric power energy saving，2013.12. 
[14] Zhang Jingwei. Power quality control of power supply and 
distribution system in metallurgical plant [J] .Water conservancy and 
Power Industry，2013.12. 
[15] Zhao Yuman. Research on harmonic suppression and reactive power 
compensation in power system [D] .Liaoning University of 
Technology，2014.8. 
 


English

Solutions for Governance and Suppression of Power Harmonic in Cities

Zhou, Xiaobo

International Journal of Advanced Network, Monitoring and Controls       Volume 03, No.04, 2018 
87 
Research on Noise Removal in Fiber Grating Sensing Signal 
Xiaobo Zhou * 
Xijing university,  
Xi’an, china 
e-mail: 369146317@qq.com  
 
 
 
Abstract—Signal processing is one of the key technology of 
fiber grating sensor, noise interference on signal transmission, 
influence of fiber grating sensor practical application effect, in 
order to solve the noise problem of fiber grating sensing signal, 
design a denoising method of fiber grating sensing signal based 
on improved contourlet transform. The acquisition of fiber 
grating sensing signal firstly, and remove the unwanted signal, 
select the useful signal, then using contourlet transform, by 
filtering the noise filter, finally through the concrete 
experiment to test the effect of denoising. The results show that 
the improved contourlet transform can completely remove the 
fiber Bragg grating sensing signals, and improve the quality of 
FBG sensor communication, and verify the effectiveness of the 
proposed method. 
Keywords-Fiber Grating; Sensing Signal; Noise Interference; 
Contourlet Transform; Filter 
I. INTRODUCTION 
With the continuous development of optical 
communication technology, fiber grating sensing technology 
obtained the unprecedented development of fiber Bragg 
grating sensor with strong ability of resisting electromagnetic 
interference resistance to corrosion, high sensitivity, small 
body Low cost advantages, has the incomparable advantages 
of traditional sensors, in the national defense military 
aerospace and other fields has been widely used in [1-3] in 
the practical application of fiber Bragg grating sensor and 
sensing signal is very susceptible to noise interference, the 
measurement precision of the fiber Bragg grating sensor is 
affected, so to remove the noise of the fiber Bragg grating 
sensor research has the vital significance. 
At home and abroad in recent years, some studies 
mechanism of fiber Bragg grating sensor signal denoising 
studies [7, 8], but the study also not enough in-depth the 
earliest using Fourier transform to deal with the noise of the 
fiber Bragg grating sensor signal, in order to improve the 
quality of the fiber Bragg grating sensor signals, but as a 
result of the Fourier transform of signal decomposition scale 
is coarser, cannot effectively eliminate the noise of the fiber 
Bragg grating sensor signal [9-11] In order to solve the 
limitations of Fourier transform, some scholars wavelet 
transform denoising method was adopted to realize FBG 
sensing signal quality of ascension, the decomposition scale 
is more detailed, more thoroughly remove noise, but the 
algorithm of wavelet threshold denoising process, the 
threshold selection is critical, whether hard or soft threshold 
threshold was adopted to realize, the current disagreed on 
threshold selection criteria, and the implementation process 
is very complicated, not easy to operation. 
Outline of the wavelet transform is a method of 
multi-scale decomposition can effective denoising signal in 
noise, in order to solve the problem of the current fiber 
Bragg grating sensor signal noise, design a fiber Bragg 
grating sensor based on contour wavelet transform signal 
denoising method, the experimental results, improve the 
outline of the wavelet transform can eliminate noise in the 
fiber Bragg grating sensor signals, won the high quality of 
the grating sensor signal. 
II. WORKING PRINCIPLE OF FIBER GRATING SENSING 
Fiber Bragg grating sensor is a kind of wavelength 
modulated sensor. As the wavelength is a constant, it is not 
affected by the optical fiber loss and energy of the light 
source, so its performance is better than other types of fiber 
sensor. Its working principle is shown in figure 1. 
DOI: 10.21307/ijanmc-2019-026
International Journal of Advanced Network, Monitoring and Controls       Volume 03, No.04, 2018 
88 
 
Figure 1. WorkingPrinciple of Fiber Grating Sensing 
 
Appens, reflection wavelength will have corresponding 
change, so that we can according to the detection wavelength 
to be the measured value of the fiber grating, such as 
temperature and pressure center wavelength drift (  ) and 
the relationship between the longitudinal strain (  ) can be 
described as  
(1 )B e
B
P




  

1
/eP dn d
n
  denotes the elastic modulus. 
The relationship between the center wavelength drift 
(  ) and temperature change ( T ) of the FBG can be 
described as 
( )B t
B
T

 


  

t And  respectively represent the coefficient of 
thermal expansion and the coefficient of thermal light. 
 
The relationship between the center wavelength drift 
(  ) and pressure change ( P ) of the fiber Bragg grating 
is 
( ) 1 1n n
P
n P n P

 
     
      
    
Let E be the elastic modulus of the fiber Bragg grating, 
and the formula for calculating the refractive index change of 
the fiber Bragg grating is 
 
(1 2 )L v P
L E
 


2
12 11(1 2 )(2 )
2
n n P
v
n E
 

  

In the process of fiber Bragg grating sensing, the 
refractive index is related to many factors, especially the 
interference of noise is the biggest, which brings adverse 
effects to the measurement results such as temperature and 
pressure. For this reason, contour transformation is 
introduced in this paper to carry out denoising processing on 
the fiber Bragg grating sensing signal, so as to improve the 
measurement accuracy such as temperature and pressure. 
III. THE METHOD OF RASTER SENSING SIGNAL 
DENOISING BY CONTOURLETTRANSFORM 
A. Contourlet Transform 
Aiming at the limitation of the wavelet transform, some 
scholars raised the outline of the wavelet transform, this 
method can from two aspects of measure frequency, by the 
signal processing has a certain direction, make the effective 
signal to concentrate for a signal containing noise, outline of 
the wavelet transform of the original signal by base structure 
approaching, outline of the wavelet transform at the same 
time also has a certain anisotropic characteristics through the 
Laplacian pyramid to multi-scale decomposition of noise 
signal, find some of the specific signal and noise, through the 
direction of the filter to filter noise operation, after denoising 
International Journal of Advanced Network, Monitoring and Controls       Volume 03, No.04, 2018 
89 
signal Firstly, the fiber Bragg grating sensor signal is 
sampled and the low frequency information of the fiber 
Bragg grating sensor signal is obtained. Then, the fiber 
Bragg grating sensor signal is sampled, and the high 
frequency information of the fiber Bragg grating sensor 
signal is obtained. In figure 2, X represents the fiber Bragg 
grating sensor signal, and H G is a low-pass filter. 
H ¯M M G -
x
a
b
+ H ¯M M G+-
x
a
b
+
 
(a) Decomposition of fiber Bragg grating signals           (b) Reconstruction of fiber Bragg grating signals 
Figure 2. Diagram of decomposition and reconstruction of fiber Bragg grating signals 
The process of three-level directional subband division of 
direction filter is shown in figure 3. The sub-bands of the 
upper level are sampled by resampling operator and the 
sampling results are output. 
 
 
Figure 3. Subband division of direction filter 
B. The De-noising Principle of Fiber Bragg grating (FBG) 
Sensor in Contour Wave Transform 
As the useful FBG sensor signal and noise show different 
characteristics in contour wave transformation, the FBG 
sensor signal is transformed in different levels, and the noise 
in the FBG sensor signal is removed according to the 
appropriate threshold value. 
Set the original fiber Bragg grating sensor signal
 ( ); 1,2, ,g x x N 
,Signal with noise fiber grating 
sensor
 ( ) ( ) ( ); 1,2, ,f x g x N x x N   
,Where, 
the mean value of noise is 0, the variance is
2 , and the 
formula for calculating the standard deviation of each layer 
is 
1,2, ,
( ( ) )
ˆ ( )
0.5
f
j J
Median w j
j




( )fw j means the decomposition coefficient of the
( )f x
JTH layer. 
C. Selection of Threshold Value of Signal Denoising by 
Fiber Bragg Grating Sensor 
There are many threshold selection methods for contour 
wave transformation. Due to the universal and easy 
implementation of the Universe threshold method, this paper 
selects the Universe threshold method to determine the 
threshold of the optical fiber grating sensor signal denoising, 
and the high-frequency sub-band threshold of the Universe 
threshold method is determined as follows 
( ) ( ) 2lg( ( )* ( ))TH j j M j N j

Where, 
( )* ( )M j N j
represents the number of contour 
wave decomposition coefficients of j layer. 
( )* ( )M j N j
,the larger the size, the denoising 
threshold will set the high frequency coefficient of the 
decomposition result of the fiber Bragg grating sensor to 0. 
In this way, a large number of useful fiber Bragg grating 
sensor signals will be removed. Therefore, equation (7) will 
be corrected as follows 
( )
( )
( )* ( )
TH j
TH j
M j N j


D. The Process of Signal Denoising by Fiber Bragg 
Grating Sensor 
Step1: contour wave transform is performed on signal f(x) 
of fiber Bragg grating sensor with noise. The coefficients of 
contour transform domain are as follows: 
( ( )) _ _w CT f x CT f CT n  

Step2: according to the denoising rules, the contour 
transformation coefficient ( wˆ ) of the original image g(x) is 
International Journal of Advanced Network, Monitoring and Controls       Volume 03, No.04, 2018 
90 
obtained by applying the denoising rules to the original 
optical fiber grating sensor signal w. The denoising rules are 
as follows: 
, ( )
ˆ
0, ( )
w if w TH j
w
if w TH j
 
 
 
Step3: perform contour wave reconstruction and 
transformation on the estimated value wˆ of contour wave 
transform coefficient, and obtain the signal of fiber Bragg 
grating sensing
'( )g x
after de-noising. 
1 ˆ( )g x CT w 

IV. TEST AND ANALYSIS OF THEDENOISING EFFECT OF 
FIBER BRAGG GRATING SENSOR 
A. The Signal of Fiber Grating Sensor WithNoise 
In order to analyze the denoising effect of the fiber Bragg 
grating sensing signal of contour wave transform, Matlab 
2014 was used as the test platform to conduct the 
denoisingexperiment. The signal of fiber grating sensor with 
noise is shown in figure 4. 
 
 
Figure 4. The signal of fiber grating sensor with noise 
B. Results and Analysis 
1) The denoising effect of this paper. Contour wave 
transform is used to de-noising the signal-sensing signal of 
the fiber Bragg grating containing noise in figure 4. The 
signal-sensing signal of the fiber Bragg grating after 
de-noising is shown in figure 5. As can be seen from figure 5, 
contour wave transform can remove the noise in the fiber 
Bragg grating sensing signal very well and improve the 
quality of the fiber Bragg grating sensing signal. It is an 
effective method to remove noise from the fiber Bragg 
grating sensing signal. 
 
 
Figure 5. The denoising results of fiber Bragg grating sensor in contour 
wave transform 
2) Performance Comparison with Current Classical 
Methods. In order to test the advantage of contour wave 
transform in denoising, wavelet transform [16] and literature 
[17] were used for comparison experiments, and peak signal 
to noise ratio (PSNR) and root mean square error (MSE)[18] 
were used to evaluate the result of signal de-noising of fiber 
Bragg grating sensor. PSNR and MSE of fiber Bragg grating 
sensor after de-noising were shown in table 1. Can be seen 
from table 1, the outline of the fiber Bragg grating sensor 
signals after wavelet transform denoising PSNR, said fiber 
Bragg grating sensor signal quality is better, noise removal 
more thoroughly, and contrast methods cannot completely 
eliminate the noise in the fiber Bragg grating sensor signals 
effectively, the interferences with fiber Bragg grating sensor 
measurement accuracy, MSE is far less than contrast method 
at the same time, said contour wavelet transform denoising 
results more stable, and the method of denoising time is 
shorter, can satisfy the mass of fiber Bragg grating sensor 
signal denoising practical application requirements. 
TABLE.I. COMPARES THE PERFORMANCE OF THE CLASSICAL RASTER 
SENSING SIGNAL DE-NOISING METHOD 
Denoising method PSNR MSE 
The wavelet 
transform 
20.33 8.00 
The literature[17] 22.19 7.72 
Contour wave 
transform 
24.73 7.24 
V. SUMMARY 
Serious impact on the quality of the fiber Bragg grating 
signal noise, its application scope and bring certain actual 
application value of interference, in order to eliminate the 
adverse effect, was proposed based on wavelet transform 
outline of fiber grating sensing signal denoising method, 
according to outline the multi-scale wavelet transform and 
directional advantages, the fiber Bragg grating sensor signal 
decomposition, found in the decomposition of the signal 
noise data, and through the corresponding filter to remove 
noise, and through the inverse operation outline of the 
wavelet transform denoising signal after refactoring, 
improves the signal-to-noise ratio of the fiber Bragg grating 
sensor signals, and compared with other denoising method 
has carried on the experiment. The contrast method can 
International Journal of Advanced Network, Monitoring and Controls       Volume 03, No.04, 2018 
91 
remove some useful signals while removing noise, while the 
method in this paper may keep the useful signal as much as 
possible, and at the same time improve the real-time 
performance of the fiber Bragg grating sensing signal. 
ACKNOWLEDGEMENT 
Xijing University Scientific Research Foundation 
Project(XJ140236). 
 
REFERENCES 
[1] L Li ,A J Xia : The advantages and applications of fiber Bragg 
grating sensing technology [J]. Optical communication technology, 
2007(7): 62-64. 
[2] K.W. Zhang, S.W. Zhang and S. X.Zhao: Application of fiber Bragg 
grating strain sensor in bridge structure monitoring [J]. Optical 
instrument, 2014, 36(1):15-19. 
[3] A.Q. LI, G.D.ZHOU: Research progress and prospect of optical fiber 
Bragg sensor testing technology [J]. Journal of southeast university 
(natural science edition),2009,39(6):1298-1306. 
[4] Jeanno F, Joel C, John B: Low energy impact damage monitoring of 
composites using dynamic strain signals from FBG sensors-part II: 
damage identification [J]. Composite Structures, 2012,94:593-600. 
[5] Z. Q.LIN, H. YANG and H.S. CHEN: Temperature compensation of 
fiber grating strain sensor [J]. Journal of southeast university (natural 
science edition), 2007, 37(2): 310-314. 
[6] C.XNING, S.S. ZHANG :Experimental study on temperature 
compensation of fiber grating [J]. Journal of hebei university of 
science and technology, 2006, 27(2): 155-157. 
[7] P.X.ZHENG, Y.L. SONG and D.S. ZHANG:  Experimental study 
on temperature and strain sensing characteristics of fiber Bragg 
grating [J]. Instrument technology and sensor,2008,22(11):12-15. 
[8] X.F. ZHOU,L.LIANG: Experimental study on stability of fiber Bragg 
grating sensor [J]. Sensor and microsystem, 2007, 26(11): 25-27. 
[9] Loutas T H, Kostopoulos V, Ramirez-Jimenez C, et al: Damage 
evolution in center-holed glass/polyester composites under 
quasi-static loading using time/frequency analysis of acoustic 
emission monitored waveforms [J]. Composite Science and 
Technology, 2006, 66:1366-1375. 
[10] L.Q.HOU,X.F.ZHAO;Z.P.LENG andT.SUN :Improved calculation 
value of temperature compensation of fiber Bragg grating strain 
sensor [J]. Journal of sensor technology, 2017, 27(1): 70-74. 
[11] J.P.LI, X.L.ZENG and J.W. GUAN: Study on refractive index 
sensing of fiber Bragg grating [J]. Laser and infrared, 2016, 46(6): 
737-741. 
[12] Hampshire T A, AdeliH.:Monitoring the behavior of steel structures 
using distributed optical fiber sensors. Journal of Constructional Steel 
Research, 2000, 53(3): 267-281. 
[13] Hill K O, Fujii Y, Johnson D C, et al. Photosensitivity in optical fiber 
waveguides: Application to reflection filter fabrication[J]. Applied 
Physics Letters, 2008, 32(10): 647-649. 
[14] J. ZHOU, T.G.NING: Study on demodulation of fiber Bragg grating 
sensing signals [J]. Optical communication technology, 2010,8:8-11. 
[15] X.LYU, Z.B.ZHANG and Y. QU:  Study on curved long-period 
photonic crystal fiber grating sensing [J]. Laser technology, 2015, 
39(4):571-575. 
[16] J. J.CAO, L.L.HU and R. ZHAO: An optical fiber grating sensing 
signal de-noising method that improves the wavelet threshold 
function [J]. Journal of sensing technology, 2015, 28(4): 521-525. 
[17] Y. P.WANG, L.L.HU and B.WANG: Optical fiber grating sensing 
signal de-noising based on digital filtering and its FPGA 
implementation [J]. Journal of xi 'an university of technology, 2015, 
31(1): 95-100. 
[18] L. LIU,M. YU and R.J. YANG:Wavelet de-noising for optical fiber 
Raman temperature sensing system [J]. China laser, 2013, 40(6): 1-5. 
 


Research on Noise Removal in Fiber Grating Sensing Signal

Geesa, Rachel Louise

Elam, Nicholas P.

Mayes, Renae D.

McConnell, Kat R.

McDonald, Kaylee M.

School leaders’ perceptions on comprehensive school counseling (CSC) evaluation processes: Adherence and implementation of the American School Counselor Association (ASCA) National Model

Matzen, Julia

Schmitt, Friedhelm C.

Kreissl, Michael C.

Voges, Jürgen

Heinze, Hans-Jochen

Galazky, Imke

43
C
a
s
e
 r
e
p
o
r
t
Ictal asystole with reduced cardiac sympathetic function in new-onset 
symptomatic epilepsy
Julia Matzen 1*, Friedhelm C. Schmitt 1*, Michael C. Kreissl 2,  
Jürgen Voges 3,4, Hans-Jochen Heinze 1,4, Imke Galazky 1
1 Department of Neurology, University of Magdeburg, Magdeburg,  
 Germany
2 Department of Radiology and Nuclear Medicine,  
 University Hospital Magdeburg, Magdeburg, Germany
3 Department of Stereotactic Neurosurgery,  
 University of Magdeburg, Magdeburg, Germany
4 Leibniz Institute for Neurobiology (LIN), Magdeburg, Germany
Correspondence
Dr. med. Julia Matzen, MD
Otto-von-Guericke University, Department of Neurology 
Leipziger Str. 44, D-39120 Magdeburg, Germany
e-mail: julia@matzen-net.de 
Phone +49 391 67 15 001
Fax +49 391 67 15 032
SUMMARY
Introduction. So far, cardiac sympathetic dysfunction, demonstrated in pharmacoresistant epilepsy pa-
tients with ictal bradycardia or asystole by I-123 metaiodobenzylguanidine (I-123 MIBG) imaging has been 
attributed to repeated occurrence of seizure activity.
Aim. Discussion of the mechanisms of cardiac sympathetic dysfunction associated with ictal asystole un-
der consideration of a case with new onset epilepsy.
Materials and methods. We describe the occurrence of a cardiac asystole during a complex-partial sei-
zure in an antiepileptic-drug-naïve patient with new-onset symptomatic epilepsy.
Results. MIBG imaging showed reduced tracer accumulation in cardiac sympathetic nerve endings in this 
patient with right parietotemporal glioblastoma.
Discussion and Conclusion. To our knowledge, this is the first report of impaired cardiac sympathetic 
function in new-onset symptomatic epilepsy without antiepileptic drug treatment. MIBG imaging should 
be considered in patients with ictal bradycardia or asystole.
Keywords: ictal asystole • MIBG imaging • symptomatic epilepsy • glioblastoma
Received April 7, 2018 
Accepted for publication on-line February 8, 2019 
Published on-line February 16, 2019
INTRODUCTION
Involvement of the autonomic nervous system often 
occurs during epileptic seizures. Frequent autonom-
ic symptoms include for example sweating, pupil dil-
atation or tachycardia. In contrast to the latter, ictal 
bradycardias or even asystoles are rare. In retrospec-
tive analyses of data from presurgical video-electro-
encephalography-monitoring, ictal asystoles of 5 to 
60s duration were found in 0.27 to 0.59% of patients 
(Kerling et al., 2009; Rocamora et al., 2003; Schuele et 
al., 2007). Nevertheless, ictal bradycardias and asysto-
les are of relevance for the affected patients, since they 
usually go along with syncopes, often accompanied by 
falls and risk of injury. A higher risk of sudden unex-
pected death in epilepsy (SUDEP) in these patients has 
been widely discussed but could not be proven yet. The 
pathophysiology underlying ictal bradyarrhythmias is 
not entirely clear, however, a key role is attributed to an 
imbalance in the autonomic nerve system (Sforza et al., * Both authors contributed equally.
Copyright and photocopying by Foundation of Epileptology, 2019
Journal of Epileptology • 2019 • 27 • 43–47 • 10.21307/jepil-2019-001
44 Julia Matzen et al.
2014). In patients with long-lasting temporal lobe ep-
ilepsy (TLE), however, cardiac sympathetic denerva-
tion has been shown by means of I-123 metaiodoben-
zylguanidine (MIBG) imaging, a scintigraphic method 
using an analogue of norepinephrine, that is retained 
in sympathetic nerve endings. Among the investigated 
patients with TLE, the finding was significantly more 
pronounced in those experiencing asystoles with their 
seizures (Druschky et al., 2001; Kerling et al., 2009).
AIM
Case presentation and discussion of the mechanisms 
of cardiac sympathetic dysfunction associated with ic-
tal asystole under consideration of a case with new on-
set epilepsy.
MATERIAL AND METHOD
Here, we report on an antiepileptic-drug-naïve patient 
suffering ictal asystoles as the sole seizure manifestation 
of new-onset symptomatic epilepsy due to glioblasto-
ma of the right parietotemporal lobe, in whom MIBG 
scanning demonstrated marked sympathetic denerva-
tion of the heart.
RESULTS
Case
A 71-year old male patient was admitted to our de-
partment with recurrent glioblastoma WHO °IV of 
the right parietotemporal lobe for brachytherapy. The 
tumour was diagnosed 11 months before by MRI and 
histopathological analysis after a hemianopsia to the 
left side had been found during a routine ophthalmo-
logical examination. First-line therapy had consisted 
of incomplete surgical resection with implantation of 
carmustine wafers and subsequent radiotherapy with 
concomitant and adjuvant temozolomide chemother-
apy. The decision for implantation of iodine-125-seeds 
was made because of further progression of the tumour. 
Clinical neurological and neuropsychological exami-
nation revealed homonymous hemianopsia to the left, 
reduced alertness and mild to moderate executive dys-
function. There was no history of epileptic seizures in 
terms of overt clinical seizure manifestation. Howev-
er, the patient reported on two events with loss of con-
sciousness preceded by nausea and dizziness shortly be-
fore the diagnosis of tumour. Otherwise, medical his-
tory revealed no relevant diseases, especially no his-
tory of cardiovascular or Parkinson’s disease; the pa-
tient was not taking any regular medication besides the 
above mentioned chemotherapeutics.
Figure 1. (A) EEG showing right temporal (T4) delta slowing (B) a seizure pattern evolves consisting of rhythmic 3/s activity, 
no clinical abnormalities (C) 50 s later: ongoing seizure pattern, patient complains an olfactory sensation, bradycardia (D) 
asystolia, generalised EEG flattening, patient unconscious (E) asystolia ends spontaneously after 23 s, clinically hypoxic 
myoclonus (F) normalisation of EEG pattern, patient awake.
45Sympathetic cardiac function in ictal asystole
During preoperative routine-EEG, the patient sud-
denly complained about a funny smell, then became un-
responsive with staring followed by loss of conscious-
ness with loose muscle tone and single bilateral asym-
metric jerks before awakening. The patient was quick-
ly reoriented and was amnestic for the episode. The 
EEG (fig. 1) initially showed continuous right fronto-
temporal delta-slowing maximal at F4 without epilep-
tiform discharges, ECG showed sinus rhythm with 72 
bpm. Subsequently, rhythmic right temporal theta-ac-
tivity developed 10 s before clinical seizure manifesta-
tion which included bradycardia. The patient became 
asystolic 66 s after clinical seizure onset, cardiac ac-
tivity reoccurred after 23 s. Flattening of the EEG last-
ed 26 s (fig. 1).
Subsequent cardiological examination including 
chest X-ray, 24 h Holter-ECG and transthoracal echo-
cardiography was unremarkable. MIBG imaging re-
vealed a low myocardial MIBG uptake (heart/mediasti-
num ratio 1.6 after 15 min and 1.8 after 4 hours post 
injection (fig. 2) reflecting a decreased post-ganglion-
ic cardiac catecholamine reuptake
After the patient exhibited another asystole requir-
ing cardiac resuscitation during his hospital stay, a car-
diac pacemaker was implanted and antiepileptic drug 
therapy was initiated (levetiracetam 1500 mg daily). 
The patient did not experience any asystoles or other 
seizure manifestations until his death 8 months later.
DISCUSSION
Ictal asystole is a rare manifestation of seizures most-
ly of temporal origin (Rocamora et al., 2003; Schuele 
et al., 2007; Tenyi et al., 2017). The exact pathomecha-
nisms underlying this seizure mediated cardiac brad-
yarrhythmia are unclear. However, in patients with 
temporal lobe epilepsy (TLE), cardiac sympathetic in-
nervation has been shown to be compromised using 
MIBG imaging (Druschky et al., 2001; Kerling et al., 
2009) a finding that is significantly more pronounced 
in those TLE patients exhibiting ictal asystoles (Kerling 
et al., 2009). According to the literature there are two 
hypothesis concerning the pathophysiological mean-
ing of this finding: It has been proposed that enhanced 
sympathetic outflow from brain regions activated dur-
ing seizure propagation leads to a compensatory down-
regulation of sympathetic nerve endings (Druschky et 
al., 2001, Kerling et al., 2009). Also, an inherent distur-
bance of cardiac sympathetic innervation accompany-
ing neuronal migration disorders in TLE patients has 
been discussed (v. Manitius-Robeck et al., 1998). How-
soever caused, deficiency of cardiac sympathetic inner-
vation probably limits counterregulation of vagal acti-
vation thus facilitating cardiac bradyarrythmias.
In its clinical appearance, the case presented here is 
typical for ictal asystoles as depicted in the literature: 
the patient experiences an aura, in this case in the form 
of a smelling sensation, followed by a complex-par-
Figure 2. Planar (123)I metaiodobenzylguanidine scans with low myocardial tracer uptake (heart/ mediastinum ratio 1.6 
after 15 min and 1.8 after 4 hours post injection). This is often found in deficient post-ganglionic cardiac catecholamine 
uptake in idiopathic Parkinson´s disease (left: anterior view; right: posterior view).
46 Julia Matzen et al.
tial seizure with staring and unresponsiveness. Both 
the semiology and the ictal EEG (fig. 1) are suggestive 
for a temporal lobe seizure. He then loses conscious-
ness due to cerebral hypoperfusion and subsequently 
myoclonic jerks before awakening as an expression of 
cerebral reperfusion (Duplyakov et al., 2014). MIBG 
scanning in our patient showed a low post-ganglion-
ic sympathetic innervation in a comparable degree as 
described in the literature (Druschky et al., 2001; Ker-
ling et al., 2009). In the aforementioned studies, how-
ever, patients with ictal asystole and pathological find-
ings on MIBG imaging had been suffering from chron-
ic pharmacoresistant epilepsy either of unknown etiol-
ogy or due to long-lasting epileptogenic lesions such as 
hippocampal sclerosis or low grade tumours. In con-
trast, the patient reported here suffered from new-on-
set symptomatic epilepsy, which had not been treated 
with antiepileptic drugs yet. Contrary to our case re-
port, Tenyi et al. (2017) found in their metaanalysis of 
156 patients that the group of new onset ictal asystole 
patients (defined as occurrence of ictal asystole with-
in a year) had a significant higher preponderance of fe-
male patients and patients with heart disease.
For temozolomide, which the patient had been tak-
ing in the course of radiochemotherapy, cardiac side 
effects have not been described. There is one patient 
reported by Druschky et al. (2001) who had been suf-
fering from TLE due to hippocampal gliosis for only 
one year, but in contrast to our patient she had a high 
seizure frequency of 15/month and was taking carba-
mazepine 1200 mg daily, a drug that is known to po-
tentially cause bradycardia and other cardiac arrhyth-
mias. In the other 21 patients-described by Druschky 
et al. (2001) – mean duration of epilepsy was 21.8 years, 
and 10 patients were taking carbamazepine. The find-
ing of a compromised sympathetic innervation of the 
heart in a patient with a newly acquired epilepsy with 
ictal asystoles challenges both of the above mentioned 
hypotheses: neither compensatory downregulation of 
sympathetic nerve endings due to repeated, seizure-
driven, enhanced cerebral sympathetic outflow nor 
combined malformations of the peripheral and central-
nervous system can explain the occurrence of ictal asys-
toles on the basis of post-ganglionic sympathetic den-
ervation of the heart in this case. Furthermore, a medi-
cation side effect is excluded in this drug-naive patient. 
Interestingly, only 18% of the patients with new-onset 
ictal asystole in the cohort of Tenyi et al. (2017), were 
pretreated with AEDs. Also, this group had a signifi-
cant higher proportion of epileptogenic lesions in MRI, 
so that it could be speculated that in these acute symp-
tomatic subgroups of patients ictal asystole is a marker 
of different pathological mechanism compared to the 
“typical”, long-lasting epilepsy patient, who are more 
prone to SUDEP.
CONCLUSION
Dysfunctional cardiac sympathetic innervation should 
also be considered in patients with sporadic seizure oc-
currence, with a recent seizure onset and without intake 
of drugs with potential cardiac side effects. For further 
discrimination of the pathomechanism of cardial au-
tonomic seizures, MIBG imaging can be helpful in pa-
tients with this seizure type.
CONFLICT OF INTERESTS
All authors have no conflict of interest: neither the au-
thors nor the author’s institution have a financial or 
other relationship with other people or organizations 
that may have influenced the presented work.
ACKNOWLEDGEMENTS
We thank Dr. med. J. Ruf for providing the results and 
images of figure 2.
REFERENCES
Duplyakov D., Golovina G., Lyukshina N., Surkova E., El-
ger C.E., Surges R.: Syncope, seizure-induced bradycardia and 
asystole: Two cases and review of clinical and pathophysiologi-
cal features. Seizure, 2014, 23: 506–511. doi: 10.1016/j.seizu-
re.2014.03.004
Druschky A., Hilz M.J., Hopp P., Platsch G., Radespiel-Tröger 
M., Druschky K. et al.: Interictal cardiac autonomic dysfunc-
tion in temporal lobe epilepsy demonstrated by (123)I metaio-
dobenzylguanidine-SPECT. Brain, 2001, 124: 2372–2382. doi: 
10.1093/brain/124.12.2372
Kerling F., Dütsch M., Linke R., Kuwert T., Stefan H., Hilz 
M.J.: Relation between ictal asystole and cardiac sympathe-
tic dysfunction shown by MIBG-SPECT. Acta Neurol. Scand., 
2009, 120: 123–129. doi: 10.1111/j.1600-0404.2008.01135.x
Manitius-Robeck S.v., Schüler P., Feistel H., Platsch G., 
Stefan H.: Iktuale Synkopen – Kardiale sympathische Innerva-
tionsstörung als Ursache? [Ictal syncopes. Cardiac sympathe-
tic innervation disorder as the etiology?]. Nervenarzt, 1998, 
69: 712–716.
Rocamora R., Kurthen M., Lickfett L., von Oertzen J., Elger 
C.E.: Cardiac asystole in epilepsy: Clinical and neurophysiolo-
gical features. Epilepsia, 2003, 44: 179–185.
47Sympathetic cardiac function in ictal asystole
Schuele S.U., Bermeo A.C., Alexopoulos A.V., Locatelli 
E.R., Burgess R.C., Dinner D.S., Foldvary-Schaefer N.: Vid-
eo-electrographic and clinical features in patients with ic-
tal asystole. Neurology, 2007, 69: 434–441. doi: 10.1212/01.
wnl.0000266595.77885.7f
Sforza E., Pichot V., Gschwind M., Roche F., Maeder-In-
gvar M.: Bradycardia and asystole during generalised interic-
tal EEG discharges. Epileptic Disord., 2014, 16: 506–509. doi: 
10.1684/epd.2014.0710
Tenyi D., Gyimesi C., Kupo P., Horvath R., Bone B., Barsi P. 
et al.: Ictal asystole: A systematic review. Epilepsia, 2017, 58: 
356–362. doi: 10.1111/epi.13644
•


Ictal asystole with reduced cardiac sympathetic function in new-onset symptomatic epilepsy

Pimentel, José

Lopes Lima, José Manuel

Rational polytherapy: Myth or reality?

Bravo-Valenzuela, Nathalie Jeanne

Peixoto, Alberto Borges

Carrilho, Milene Carvalho

Siqueira Pontes, Ana Letícia

Chagas, Caroline Cevante

Simioni, Christiane

Júnior, Edward Araujo

Fetal cardiac function by three-dimensional ultrasound using 4D-STIC and VOCAL – an update

Peide, Lou

Zhang, Liu

Review of Development and Application of Future Network (IPV9)

El-Sadi, Haifa

Cole, Matthew R.

Denis, Aaron M.

Fernandes, Derek

Benhamu-Chocron, Alberto

International Journal of Advanced Network, Monitoring and Controls                      Volume 03, No.03, 2018 
DOI: 10.21307/ijanmc-2019-005                                              47 
 
Design and Analysis of Thermoplastic Metal Detector  
RC Car with Wireless Charging 
 
Haifa El-Sadi 
Wentworth institute of technology 
Mechanical engineering Department 
Boston, MA 
e-mail: elsadih@wit.edu 
Matthew R. Cole, Aaron M. Denis 
Wentworth institute of technology 
Mechanical engineering Department 
Boston, MA 
Derek Fernandes  
Wentworth institute of technology 
Mechanical engineering Department 
Boston, MA 
 
Alberto Benhamu-Chocron 
Wentworth institute of technology 
Mechanical engineering Department 
Boston, MA
 
 
Abstract—The RC Car industry is a growing industry that will 
always be a past time for the older generation. This research is 
focused on a specific type of RC Car which is a Metal Detecting 
one. Metal Detectors have been developed for many years. 
Through research there aren’t many metal detecting RC Cars 
on the market. Currently it’s extremely limited in range and 
depth of detection. 
The goal of this research is to improve, build and test a new 
RC-metal detector technology. A thermoplastic-ABS used to 
build the chassis of the RC- metal detector. ABS is easily 
machined, sanded, glued and painted. Finite element analysis is 
a powerful tool that allows us to quickly analyze and refine a 
design. When the Chassis fixed at front Differential 
(Displacement), the value of the maximum stress and the 
deflection are higher than when the Chassis Fixed at Rear 
Differential. The maximum deflection is shown as about 1.339 
inches. This value is the resultant of the deflection in all three 
directions. This research includes, a cost analysis for each piece 
of the car, group goals for the RC Car and Metal Detector. A 
working prototype is designed before moving into the final 
design phase in order to assure that the best possible product 
can be produced. There is a wireless charging station that ease 
the process of recharging the battery. 
Keywords-Metal Detector; Finite Element; Thermoplastic; 
RC 
I. INTRODUCTION 
The challenge for this project is to design and build a 
remote-controlled car that incorporates two features that no 
other car possesses. RC Car owners have tedious work to do 
every time the car needs to be charged. Some components 
must be taken apart to retrieve the battery to charge it. The 
wireless charging feature being implemented in the design 
will eliminate this extra work, while providing a smooth 
charging operation where no disassembling will take place. 
The design and development of remotely operated solar-
powered mobile metal detector robot is a rescue robot to 
autonomously operate in detecting the threat of land mines. 
During the First and Second World War, military forces 
deployed many bombs on land filed to fight between soldiers 
on the battlefield. There were many countries like Libya, 
Cambodia and Laos had explosive weapons that did not 
explode when fired or dropped on the ground. In fact, more 
than twenty thousand people have been killed or injured by 
unexploded bombs [1]. A remotely solar-powered mobile 
metal detector robot has been designed and implemented. 
The system is using RF communication with Atmega32 
MCU in embedded system domain. The robot moves in 
particular direction using the handheld remote. The 
experimental work has been carried out carefully. The metal 
detector sensor worked as the required specification for the 
metal detection sensor. The testing demonstrated that the 
robot would not pose any performance problem for the 
installation of the metal detection robot such as the merits 
and drawbacks of mounting the sensor, cost, support vehicle, 
handling the cable between the robot and also easiness of the 
adjustment [2]. Nation et al. [3] demonstrated the accuracy 
of HHMD in the identification and localization of metallic 
foreign bodies. They proposed an emergency room foreign 
body protocol that uses HHMD as an early screening tool in 
triage in order to expedite the process of obtaining 
Otolaryngology consultation and potentially shorten the wait 
time to the operating room or discharge. In instances were 
outside films are previously performed, HHMD use may be 
able to minimize the overall radiation exposure to children 
by obviating the need for repeat radiographs. As the 
sensitivity is not 100%, a negative HHMD screening does 
not negate the need for a standard radiograph in order to 
avoid missed MFBs. HHMD is best suited for detection of 
coins, which accounts for the majority of the MFB ingestions, 
and may not be suitable for all metallic objects since the 
amount of metal may decrease its sensitivity. 
Holm, Katja F et. al. [4] evaluated a commercially 
available metal detector for detecting CIEDs. Design. 
Observational study including pacemaker patients (n = 70) 
International Journal of Advanced Network, Monitoring and Controls                      Volume 03, No.03, 2018 
48 
and a control group without pacemaker ( n = 95). The 
investigational device was a hand-held metal detector for 
detecting metal or electricity wiring. Results. The metal 
detector detected the pacemaker in all pacemaker patients 
and thus exhibited a sensitivity of 100%. The specificity of 
the metal detector was 86%, and the negative predictive 
value was 100%. Thirteen individuals without pacemakers 
were falsely identified as having an implanted device due to 
implanted prosthetic material or elements of clothing. 
The objective of this research is to design a RC car that 
will provide a smooth wireless charging process and 
incorporate a metal detector. RC Cars require tedious work 
in order to charge the battery. It will offer a wireless 
charging station where no disassembling of the car will be 
required. Apart from technical innovations, an improved 
metal detector will be incorporated that will be useful on 
beaches and in parks to search for lost jewelry. A long-term 
goal with an unlimited budget is to use this car to detect 
mines for the army. This car could drive over the field before 
infantry and passenger vehicles drive over it to protect them 
from mines or IEDs. 
II. MATERIAL AND COST 
 
Acrylonitrile Butadiene Styrene (ABS) is an opaque 
thermoplastic and amorphous polymer. ABS becomes liquid 
at a certain temperature, 221 degrees Fahrenheit. They can 
be heated to their melting point, cooled, and re-heated again 
without significant degradation. Instead of burning, 
thermoplastics like ABS liquefy which allows them to be 
easily injection molded and then subsequently recycled. ABS 
is easily machined, sanded, glued and painted. This makes it 
a great material for prototyping. Table 1 shows the required 
material to build the RC-car metal detector. 
TABLE I. THE MATERIAL AND THE COMPONENTS OF RC-CAR METAL DETECTOR 
 
 
A. Finite Element Analysis of Material 
SOLIDWORKS Simulation uses the displacement 
formulation of the finite element method to calculate 
component displacements, strains, and stresses under internal 
and external loads. The geometry under analysis is 
discretized using tetrahedral (3D), and solved by iterative 
solver. SOLIDWORKS Simulation using p adaptive element 
type, the solution has converged. The material parameters 
were obtained and the results were simulated. One of the 
most important inputs to the model is the elastic modulus E 
of the material. The elastic modulus defines the stiffness 
(resistance to deflection) of the material. Its value is 
determined from material tests. A material with a high value 
of E will deflect less than one with a lower value of E. By 
applying finite element analysis, we can accurately observe 
the stress distributions in the various layers of the material as 
shown in Figures 1, 2, 3 and 4. Figure 1 shows Chassis Fixed 
at Rear Differential, the highest stress is 4039 psi with mesh 
size was 0.2. However, the maximum deflection was 0.7 
inches as shown in Figure 2. Simulates impact on underside 
of chassis. 
B. ABS Plastic Material Data 
Elastic Modulus: 290075.4753 psi, Shear Modulus: 
46252.53454 psi, and Tensile Strength: 4351.13213 psi 
 
International Journal of Advanced Network, Monitoring and Controls                      Volume 03, No.03, 2018 
49 
 
Figure 1. Chassis Fixed at Rear Differential (Von Mises) 
 
Figure 2. Chassis Fixed at Rear Differential (Displacement) 
 
Figure 3 shows Chassis Fixed at front Differential 
(Displacement), the highest stress is 8375 psi with mesh size 
was 0.2. However, the maximum deflection was 1.339 
inches as shown in Figure 4, it Simulates impact on 
underside of chassis 
International Journal of Advanced Network, Monitoring and Controls                      Volume 03, No.03, 2018 
50 
 
Figure 3. Chassis Fixed at Front Differential (Von Mises) 
 
Figure 4. Chassis Fixed at front Differential (Displacement) 
III. DESIGN PROTOTYPE AND TESTING 
One of the biggest challenges was designing a chassis 
from scratch that would house all the components necessary 
to operate a RC Car. The chassis also needed to be strong 
enough to not only hold the weight of the car and all its 
components, as shown in Figure 5, but also handle the torque 
and power that would be outputted from the car. The 
designed chassis was more than suitable to hold all the 
components and handle the force of the motor. There was 
more room than anticipated, which was useful for placing the 
components and keeping things away from all the moving 
parts. In the end, the RC Car is programmed and moves as 
planned. The metal detector detects deeper than expected. 
Finally, the wireless charging station works better than 
anticipated. Figure 7 shows the Wireless Charging Ramp 
Setup. 
International Journal of Advanced Network, Monitoring and Controls                      Volume 03, No.03, 2018 
51 
 
Figure 5. Detailed Picture of all Components 
 
 
Figure 6. Layout of the Electrical Circuit 
 
Figure 6 shows the layout of the electrical circuit of the 
RC-car metal detector. The main circuit is powered by a 
battery through a mechanical switch. If the switch is turned 
ON, power is supplied to a motor and a servo. The speed of 
the motor is regulated by an Electronic Speed Control Circuit 
(ESC). The servo on the hand is controlled by a Raspberry Pi 
computer board. The circuit also has four terminals that are 
used to connect to an external charger once the battery has 
depleted. 
 
Figure 7. Wireless Charging Ramp Setup 
 
Figure 8. Final prototype 
IV. CONCLUSION 
This project aimed at creating a thermoplastic RC Car 
designed with a purpose, more than just an average 
recreational toy. This was bringing up the reality of a much 
greater idea. The wireless charging aspect of the RC Car will 
mostly be aimed towards small scale vehicles. However, the 
metal detection can be a very realistic and useful application 
for much bigger vehicles. The idea that an unmanned vehicle 
can be used for detecting not only metals, but harmful 
objects could potentially be a breakthrough for something 
like the military. IED’s are a very serious issue for the 
military, being able to sniff these out without having to risk 
the lives of anyone would be huge. Completing this project 
with a limited budget would show a lot, especially when 
someone like the military could spend endless amounts of 
money in order to perfect the system for more dangerous and 
life like scenarios. FEA analysis was used to test the chassis, 
When the Chassis Fixed at front Differential (Displacement), 
International Journal of Advanced Network, Monitoring and Controls                      Volume 03, No.03, 2018 
52 
the value of the maximum stress and the deflection are 
higher than when the Chassis Fixed at Rear Differential. 
V. ACKNOWLEDGEMENTS 
The authors sincerely appreciate the immense and highly 
valued support from the Dean of Engineering, Dean Fred 
Driscoll 
 
REFERENCES 
[1] Handicap International, Fatal Footprint: The Global Human Impact of 
Cluster Munitions. November 2006. Available from: URL: 
http://www.mineaction.org 
[2] F.Y.C. Albert, C.H.S. Mason, C.K.J. Kiing, K.S. Ee, K.W. Chan, 
“Remotely Operated Solar-powered Mobile Metal Detector Robot”, 
Procedia Computer Science, Volume 42, 2014, Pages 232-239 
[3] Nation, Javan and Jiang, Wen, “The utility of a handheld metal 
detector in detection and localization of pediatric metallic foreign 
body ingestion”, International Journal of Pediatric 
Otorhinolaryngology January 2017 92:1-6 
[4] Holm, Katja F, Hjortshøj, Søren, Pehrson, Steen, Svendsen, Jesper 
Hastrup, Riahi, Sam, “Implanted cardiac devices are reliably detected 
by commercially available metal detectors”, Scandinavian 
Cardiovascular Journal. Oct2013, Vol. 47 Issue 5, p271-274. 4p. 
 
 


Design and Analysis of Thermoplastic Metal Detector RC Car with Wireless Charging

Lou, Peide

Xu, Fei

IPV9 is the Foundation of the Future Digital World

Yu, Jun

Xian, Tong

Hu, Zhiyi

Liu, Yutong

International Journal of Advanced Network, Monitoring and Controls          Volume 04, No.02, 2019 
81 
Hazard Grading Model of Terrorist Attack Based on 
Machine Learning 
 
Yu Jun 
School of Computer Science and Technology 
Xi'an Technological University 
Xi'an 710021, Shaanxi, China 
e-mail: yujun@xatu.edu.cn 
Xian Tong 
School of Computer Science and Technology 
Xi'an Technological University 
Xi'an, 710021, Shaanxi, China 
Hu Zhiyi  
Institute of Engineering Design  
Army Academy of PLA 
Beijing, 100042, China 
 
Liu Yutong 
Engineering Design Institute  
Army Academy of PLA 
Beijing, 100042, China 
 
 
 
Abstract—In this paper, there is no unified grading standard 
for the harm of terrorist attacks. A classification model of 
terrorist incidents based on machine learning is proposed. 
First, the data related to the hazard in the Global Terrorism 
Database (GTD) is extracted and preprocessed. Secondly, the 
data is extracted by principal component analysis, and all 
events are aggregated into 5 by K-means clustering. Again, the 
entropy method is used to calculate the weighting coefficient of 
each indicator, and the comprehensive score of the hazard of 
each type of terrorist attack is calculated. Finally, the scores 
are divided into 1-5 levels of hazard grading models in order of 
high to low. The results show that the hazard grading model 
can scientifically and objectively quantify terrorist attacks. 
Keywords-Terrorist Attacks; Hazard; Hierarchical Model; 
Principal Component Analysis; K-Means Clustering; Entropy 
Method 
I. INTRODUCTION  
A terrorist attack is an aggression committed by an 
extremist or organization that is not in conformity with 
international morality and is directed against, but not 
limited to, civilians and civilian installations. It not 
only has great destructiveness and destructive power, 
but also directly causes huge casualties and property 
losses. It also brings tremendous psychological 
pressure to people, causing a certain degree of turmoil 
in society and greatly hindering economic development. 
Global terrorism is a phenomenon of public interest, 
and everyone is directly affected by it. Therefore, anti-
terrorism work is imminent. Big data is now the main 
source of counter-terrorism intelligence. The Global 
Terrorism Database (GTD) is the world's most 
comprehensive database of non-confidential terrorist 
attacks, containing more than 180,000 terrorist attacks, 
each containing at least 45 variables. An in-depth 
analysis of data related to terrorist attacks will help 
deepen people's understanding of terrorism and provide 
valuable information support for opposing terrorism 
and preventing terrorism. Data collection and 
preprocessing intelligence are the lifeblood of counter-
terrorism work. Keeping reliable information in a 
timely manner can play an active role in combating 
terrorism and effectively curb the spread of 
terrorism[2]. 
Grading catastrophic events (such as earthquakes, 
traffic accidents, meteorological disasters, etc.) is an 
important task of social management. The usual 
grading generally adopts a subjective method, and the 
authority stipulates the grading standard. The 
harmfulness of terrorist attacks depends not only on the 
two aspects of casualties and economic losses, but also 
on the timing, geography, and targeted objects. 
Therefore, it is difficult to fully reflect these factors. 
The  
hazard grading of terrorist incidents can clearly 
define the future attacks, and different levels of events 
correspond to different treatments. This will not only 
help the management of social security, but also avoid 
unnecessary waste of manpower and property. 
Combined with big data processing technology, this 
paper establishes a hierarchical model based on PCA 
algorithm, K-meas clustering algorithm and entropy 
method. First, 14 evaluation indicators related to the 
hazard of the event were selected to preprocess the 
DOI: 10.21307/ijanmc-2019-051
International Journal of Advanced Network, Monitoring and Controls          Volume 04, No.02, 2019 
82 
existing data. Secondly, the PCA method was used to 
reduce the index from 14 dimensions to 4 dimensions, 
and the reduced dimension vector was obtained by the 
clustering algorithm. Gather into 5 categories, you can 
get the category corresponding to each event. Finally, 
using the entropy method to score the hazard of each 
event and according to the average hazard score of 
each class. According to the degree of harm from high 
to low levels 1 to 5. A hazard grading model of 
terrorism events is obtained with a hazard rating of 5. 
II. DATA PREPROCESSING 
In this paper, the hazard grading model of terrorism 
events data is established from some important fields of 
the GTD original database. The selected data handling 
requires missing value processing, conversion of 
characters to numeric values and numerical processing. 
A. Important field selection 
The Important field of hierarchical is pointed out by 
the World Anti-Terrorism Incident Research. The 
Terrorism Hazard Classification Model Data Table has 
selected the following 14 fields from GTD, as shown in 
Table 1. 
TABLE I.  THE SELECTED FIELD TABLE 
Field Description 
extended Whether it is a continuous event 
latitude latitude 
longitude longitude 
success Successful attack 
suicide Suicide attack 
nkill Total number of deaths 
propextent Degree of property damage 
nwound Total number of injuries 
country country 
region area 
city city 
attacktype Attack type 
targtype Target/victim type 
weapontype Weapon type 
 
B. Missing value processing 
In the selected field, Python's function DataFrame. 
dropna  can delete rows or columns with null values, 
and retain all data that is not empty.  Then the character 
field needs to be converted to a numeric field. 
C. Converting character fields to numeric fields 
The character field that need to be converted is as 
follows: 
1) Eventid: Events in the GTD are numbered with 
12 digits. The first 8 digits are recorded in the format 
"yyyymmdd". The last 4 digits calibrate the serial 
number of the day, e.g. 0001, etc.  
2) Country: According to the developed economies 
assessment standards recognized by the United Nations, 
168 countries are divided into developed and 
underdeveloped countries. Since terrorist attacks are 
more harmful to developed countries, the relevant 
assignments are shown in Table 2.1. 
3) Region: Count the frequency of terrorist 
incidents in each region and assign the frequency to 
regional indicator values. 
4) City: The world city is divided into three levels: 
the capital, the provincial capital, and other cities. 
Since the terrorist attacks are more harmful to the 
political and economic centers, the relevant 
assignments are shown in Table 2.1. 
5) Attack type: Counting the frequency of 
occurrence of 9 types of attacks, and assigning the 
frequency to the attack type indicator value. 
6) Weapon type: Counting the frequency of 
occurrence of 13 weapon types, and assigning this 
frequency to the weapon type indicator value. 
7) Targtype: Counting the frequency of occurrence 
of 22 target types, and assigning this frequency to the 
target type indicator value. 
TABLE II.  THE STATE AND CITY ASSIGNMENT 
Index assignment 
developed countries 2 
underdeveloped countries 1 
the capital 3 
the provincial capital 2 
other cities 1 
International Journal of Advanced Network, Monitoring and Controls          Volume 04, No.02, 2019 
83 
D. Numerical processing 
In the original GTD database, the nkill field 
includes the number of all victims and terrorists who 
directly caused death from terrorist incidents. We use 
only requires the number of victims and does not 
require the death toll of terrorists. Therefore, the 
number of victims is obtained by subtracting the 
number of terrorist deaths (nkiller) from the total 
number of deaths. 
III. TERRORIST ATTACK HAZARD CLASSIFICATION 
MODEL 
In this paper, the PCA algorithm, K-means 
clustering algorithm and entropy method are used to 
classify the terrorist attacks. The process of building a 
hierarchical model is divided into four steps: 
1) The 14 indicators with greater influence is 
standardized by PCA algorithm. We construct a 14-
dimensional matrix, and then reduce the matrix from 
14 dimensions to 4 dimensions. 
2) The K-means algorithm is used to cluster all the 
terrorist events in the matrix into five major categories, 
i.e. five hazard levels. 
3) Using the entropy weight method finds the 
weights of each of the 14 indicators, and then 
weighting and summing the 14 indicators of each event 
to obtain the score of the event. For each hazard level, 
finding the average score for all events is at that level. 
4) Sorting by the average scores of the five hazard 
levels, We divide them into one to five grades from 
high to low. The higher score means the greater 
damage.  
A. Using the PCA algorithm for dimensional 
reduction 
Principal Component Analysis (PCA) extracts M-
dimensional feature matrices from N-dimensional 
matrices. First, we calculates eigenvalues and 
eigenvectors of N-dimensional matrices. According to 
the order of PCA eigenvalues from large to small, we 
select the corresponding first M eigenvectors., and then 
obtain an N*M feature transformation matrix T. In this 
paper, N=14, M=4. The dimensionality reduction is 
completed.[6] 
The order of  PCA eigenvalues generated by 14 
indicators from large to small is shown in Table 3. 
 
TABLE III.  CHARACTERISTIC VALUES CORRESPONDING TO THE 
INDICATORS 
Indicators Characteristic values 
nkill 9.82022087e-01 
nwound 8.06184462e-02 
targtype 7.91122120e-03 
country 5.20872985e-02 
attacktype 4.84991077e-03 
region 4.01240379e-02 
suicide 2.66626688e+00 
city 2.60031933e-02 
longitude 1.84972981e+02 
extended 1.63936354e+03 
latitude 1.36725606e+03 
propextent 1.06560032e-01 
success 1.04574700e+02 
weapontype 0.00000001e+00 
 
In this paper, 98686 data is reduced by the PCA 
algorithm, i.e. the original 14-dimensional matrix 
  1 2 3 4 5 6 7 8 9 10 11 12 13 14x = x ,x ,x ,x ,x ,x ,x ,x ,x ,x ,x ,x ,x ,x is reduced to a 4-
dimensional matrix 1 2 3 4Y = y ,y ,y ,y   . The corresponding 
contribution degrees of the 4-dimensional feature vectors 
are: 0.49, 0.42, 0.06, 0.03, and the sum is greater than 
0.99. Therefore, the dimension-reduced matrix preserves 
most of the original data and can be directly used for 
clustering. 
B. Using K-Means algorithm for Hazard classification 
The main idea of the K-means clustering algorithm is to 
cluster a number of discrete data points with k centroids and 
divide them into k clusters to distinguish data points with less 
similarity. Sum of the squared error (SSE) is the objective 
function of clustering, and classify data points with similar 
similarity into one class. The method finally converges to the 
optimal solution by continuously updating the centroid 
attribution and centroid position of the data points[1]. The 
algorithm process is as follows: 
1) We select 5 event objects as the initial cluster center. 
2) We calculate the Euclidean distance from each event 
to each cluster center and assign this event to the nearest 
cluster. 
3) After all the event assignments are completed, the five 
cluster centers are recalculated, and compared with the 
cluster center obtained in the previous calculation. If the 
cluster center changes, the Euclidean distance and the 
assigned category are recalculated. 
International Journal of Advanced Network, Monitoring and Controls          Volume 04, No.02, 2019 
84 
4) When the cluster center does not change, the 
clustering result is directly output. 
Calculate the cluster center to which each type of event 
belongs, as shown in Table 4. 
TABLE IV.  TABLE 4. CLUSTERING CENTER FOR EVENT 
CLASSIFICATION 
type X1 X2 X3 X4 numbers 
0 2.4843 -16.3826 -1.3464 0.3081 63122 
1 -3.3968 22.8297 -3.8782 0.0615 37848 
2 825.778 873.697 28.9316 -104.59 2 
3 13.8411 -127.794 19.7789 -2.7281 3500 
4 -9.5985 63.3898 16.7324 -1.2382 9711 
The formula for calculating each event category is as 
shown in Equations (1) to (6). 
       2 2 2 2
1 1 2 3 1
( 8256.783) ( 873.658) ( 28.915) ( 104.608)D y y y y  
        
2 2 2 2
2 1 2 3 4
( 13.840) ( 127.794) ( 19.779) ( 2.728)D y y y y  
       2 2 2 2
3 1 2 3 4
( 2.484) ( 16.382) ( 1.346) ( 0.308)D y y y y  
2 2 2 2
4 1 2 3 4
( 3.396) ( 22.829) ( 3.878) ( 0.061)D y y y y         
        
2 2 2 2
5 1 2 3 4
( 9.598) ( 63.289) ( 16.731) ( 1.238)D y y y y  
 i 1 2 3 4 5min min{D ,D ,D ,D ,D }  
Among them is 1 2 3 4Y = y ,y ,y ,y    the feature 
component vector after dimension reduction by 
PCA algorithm. Di is the Euclidean distance 
between the dimension vector and the five cluster 
centers. mini is the minimum Euclidean distance, 
and i is the final event category. 
C. Using entropy method for calculating weight 
coefficient 
The entropy method is a mathematical method used 
to determine the degree of dispersion of an indicator. 
With the great degree of dispersion comes great impact 
of the comprehensive evaluation of the indicator. The 
entropy value can be used to determine the degree of 
dispersion of an indicator. The steps of calculating the 
weight coefficient by the entropy method are as follows: 
1) We select 14 indicators of 98686 events, and use 
xij to indicate the index value of the i-th indicator in the 
j-th terrorist attack. ）；；（ 1498686n14,,1;98686,,1i  mj   
2) Normalization of 14 indicators is Normalized 
processing. The absolute values of the 14 indicators are 
conversed into relative values. It  has different 
representative meanings that the positive indicator and 
the negative indicator value (the higher the positive 
indicator value is the better), the lower the negative 
indicator value is the better), as shown in Equation (7) 
and Equation (8). 




'
1 1
min{ ,... }
max{ ,..., } min{ ,..., }
ij ij nj
ij
j nj j nj
x x x
x
x x x x  




' 1
1 1
max{ ,... }
max{ ,..., } min{ ,..., }
j nj ij
ij
j nj j nj
x x x
x
x x x x  
3) Calculating the proportion of the i-th event in the 
j-th index are shown in Equation 9. 




n
i
ij
ij
ij
x
x
1
p
 
4) Calculating the entropy value of the j-th indicator, 
are shown in Equation 10. 
 )nln(/1k            
0),ln(e
1

 

j
n
i
ijijj eppk
 
5) Calculating the information entropy redundancy 
are shown in Equation 11. 
 jj e1d  
6) Calculating the weights of each indicator are 
shown in Equation 12. 




m
j
j
j
j
d
1
d
w
 
7) Calculating the hazard weighting value of each 
event are shown in Equation 13. 
International Journal of Advanced Network, Monitoring and Controls          Volume 04, No.02, 2019 
85 
 



m
j
iji w
1
j xs
 
The weighting factors for each indicator are shown 
in Table 5. 
TABLE V.  TABLE 5. WEIGHT COEFFICIENTS OF EACH INDICATOR 
indicator x1 x2 x3 x4 x5 x6 x7 
Weight 0.25 0.01 0.26 0.15 0.17 0.08 0.01 
indicator x8 x9 x10 x11 x12 x13 x14 
Weight 0.01 0.01 0.01 0.01 0.01 0.01 0.01 
D. Hazard grading result 
All events can be divided into five hazard levels by 
PCA and K-Means clustering. The hazard score of 
each event is obtained by entropy method, and the 
average value of the hazard score of each type of event 
is obtained. After sorting the average, the five hazard 
levels are shown in Table 6. 
TABLE VI.  TABLE 6. HAZARD GRADING RESULT 
Hazard level Cluster category Hazard level 
1 2 1766.7104 
2 3 3.2596 
3 0 0.6239 
4 4 -2.6904 
5 1 -0.8788 
IV. CONCLUSION 
In this paper, 14 categories related to hazard are 
selected from the Global Terrorism Database (GTD) 
for the hazard grading of terrorist attacks; after pre-
processing the data used, through principal component 
analysis (PCA) The related data is used for feature 
extraction. The K-means clustering method aggregates 
all events into five categories. The entropy method 
calculates the weight coefficient of each indicator, and 
finally obtains the comprehensive score of the harm of 
each type of attack. According to the comprehensive 
scores of the five types of attacks, a graded to five-
level classification model was obtained. This model 
quantifies the relevant data of past terrorist attacks, 
and the obtained model has objectivity. It is necessary 
to establish more detailed grading standards. 
 
REFERENCE 
[1] Sanjun Nie. Research on Counter-terrorism based on Big Data[A]. 
IEEE Beijing Section. Proceedings of 2016 IEEE International 
Conference on Big Data Analysis (ICBDA) [C]. IEEE Beijing 
Section: IEEE BEIJING SECTION Institute of Electrical Engineers 
Beijing Branch), 2016: 5. 
[2] Strang, Kenneth David & Sun, Zhaohao. (2017). Analyzing 
Relationships in Terrorism Big Data Using Hadoop and Statistics. 
Journal of Computer Information Systems. 57. 67-75. 
10.1080/08874417.2016.1181497.K. Elissa, “Title of paper if 
known,” unpublished. 
[3]  Li Wei. Characteristics and Trends of Current International Terror 
and Anti-Terrorism Struggle [J]. Modern International Relations, 
2007 (02): 22-27. 
[4] Yu Yihan, Fu Wei, Wu Xiaoping. Privacy data metric and 
hierarchical model based on Shannon information entropy and BP 
neural network[J].Journal of Communications,2018,39(12) 
[5] He Jing. Research and Analysis of Future Anti-terrorism Situation 
Based on Big Data[J]. Economic Research Guide, 2019(05): 186-187. 
[6] Wang Qi, Li Xiaopei, Dong Xinyan. Classification model of wine 
grape based on principal component analysis[J]. China High-tech 
Zone, 2018(05): 218. 
[7] Wang Chao, Yao Min, Fu Zhanzhan. Research on Emergency 
Classification Based on Fuzzy Comprehensive 
Evaluation[J].Software Guide, 2019(04): 149-15 
[8] Lu Ronghui. Terrorism and Counter-Terrorism in the Context of 
Globalization [D]. Suzhou University, 2005. 
[9] Wang Chao, Yao Min, Fu Zhanzhan. Research on Emergency 
Classification Based on Fuzzy Comprehensive 
Evaluation[J].Software Guide, 2019,18(04):149-152. 
[10] Hou Wenjing, Jiang Xinxin, Wen Hong, Lei Wenxin, Xu Aidong. 
Terminal Security Level Grading Model of BP Neural Network 
Based on Edge Side[J].Communication Technology, 2018, 51(10): 
2455-2458. 
[11] Shi Ya, Wang Xiuhua, Yang Wei, Liu Li, Tan Zhezhen, Ouyang Wei. 
Study on the grading strategy of comprehensive evaluation model for 
long-term care of the elderly[J]. Chinese Journal of Nursing, 2018, 
53(10): 1237-1243 
 
   


Hazard Grading Model of Terrorist Attack Based on Machine Learning

Not a member yet