University of Huddersfield Repository

University of Huddersfield Repository
University of Huddersfield Repository
Agurto Goya, Alan
New Proposal for the Detection of Concealed Weapons: Electromagnetic Weapon Detection for
Open Areas
Original Citation
Agurto Goya, Alan (2009) New Proposal for the Detection of Concealed Weapons: Electromagnetic
Weapon Detection for Open Areas. Masters thesis, University of Huddersfield.
This version is available at http://eprints.hud.ac.uk/7067/
The University Repository is a digital collection of the research output of the
University, available on Open Access. Copyright and Moral Rights for the items
on this site are retained by the individual author and/or other copyright owners.
Users may access full items free of charge; copies of full text items generally
can be reproduced, displayed or performed and given to third parties in any
format or medium for personal research or study, educational or not-for-profit
purposes without prior permission or charge, provided:
•
•
•
The authors, title and full bibliographic details is credited in any copy;
A hyperlink and/or URL is included for the original metadata page; and
The content is not changed in any way.
For more information, including our policy and submission procedure, please
contact the Repository Team at: [email protected]
http://eprints.hud.ac.uk/
New Proposal for the Detection of Concealed Weapons:
Electromagnetic Weapon Detection for Open Areas
ALAN AGURTO GOYA
A thesis submitted to the University of Huddersfield
in partial fulfillment of the requirements for
The degree of Master of Philosophy
School of Computing and Engineering
University of Huddersfield
July 2009
COPYRIGHT
i.
ii.
iii.
The author of this thesis (including any appendices and/or schedules to this thesis)
owns any copyright in it (the “Copyright”) and s/he has given The University of
Huddersfield the right to use such Copyright for any administrative, promotional,
educational and/or teaching purposes.
Copies of this thesis, either in full or in extracts, may be made only in accordance
with the regulations of the University Library. Details of these regulations may be
obtained from the Librarian. This page must form part of any such copies made.
The ownership of any patents, designs, trademarks and any and all other
intellectual property rights except for the Copyright (the “Intellectual Property
Rights”) and any reproductions of copyright works, for example graphs and tables
(“Reproductions”), which may be described in this thesis, may not be owned by
the author and may be owned by third parties. Such Intellectual Property Rights
and Reproductions cannot and must not be made available for use without the
prior written permission of the owner(s) of the relevant Intellectual Property
Rights and/or Reproductions.
2
ABSTRACT
Terrorist groups, hijackers, and people hiding guns and knifes are a constant and increasing
threat. Concealed weapon detection (CWD) has turned into one of the greatest challenges
facing the law enforcement community today. Current screening procedures for detecting
concealed weapons such as handguns and knives are common in controlled access settings
such as airports, entrances to sensitive buildings and public events. Unfortunately screening
people in this way prior to entering controlled areas is ineffective in preventing some
weapons from getting through and also produces bottle-necks in crowded environments.
Also the screening technologies employed have a high rate of false alarms due to poor
discrimination capability. A reliable CWD that is able to work in open areas and a robust
method capable to discriminate between ferromagnetic weapons are necessary.
This thesis reviews recent developments in the area of CWD using largely electromagnetic
methods. The advantages and disadvantages of these approaches are discussed and a new
research direction in CWD is presented. This thesis proposes a cost-effective weapon
detection system based on pulse induction technology which is able to work in open areas
without invading individual privacy. This approach employs a uniform magnetic field
generator to transmit pulses that cause eddy currents to flow in any metal object carried by
people. The induced eddy currents decay exponentially following sudden changes in the
exciting magnetic field with a characteristic decay time (time constant) that depends on the
size, shape, and material composition of the object. The decay currents generate a
secondary magnetic field and the rate-of-change of the field is detected by the sensors. This
thesis introduces models based on finite element analysis (FEA) to study the potential use
of the time constant as a signature for weapon discrimination. Experimental work is also
presented that confirms the theoretical predictions obtained from FEA. It is shown that
further work on signature extraction and signal processing needs to be done to build the
weapon signature database necessary for classification.
3
ACKNOWLEDGEMENTS
I would like to express my sincere gratitude to Dr Martin Sibley for his advice, guidance
and encouragement throughout this work.
The author is especially thankful to his wife and daughter for being his biggest motivation
on this work.
.
4
TABLE OF CONTENTS
COPYRIGHT ....................................................................................................................... 2
ABSTRACT .......................................................................................................................... 3
ACKNOWLEDGEMENTS................................................................................................. 4
TABLE OF CONTENTS..................................................................................................... 5
LIST OF FIGURES ............................................................................................................. 7
LIST OF TABLES ............................................................................................................... 9
LIST OF ABBREVIATIONS ........................................................................................... 10
Chapter 1: INTRODUCTION .............................................................................................. 11
Chapter 2: LITERATURE SURVEY .................................................................................. 13
2.1 Review of Current CWD Technologies ..................................................................... 13
2.1.1 Metal detection using Earth magnetic field distortion ........................................ 13
2.1.2 Inductive magnetic field methods ....................................................................... 16
2.1.3 Acoustic and ultrasonic detection ....................................................................... 21
2.1.4 Electromagnetic resonances ................................................................................ 23
2.1.5 Millimetre waves (MMW) .................................................................................. 26
2.1.6 Terahertz (THz) imaging..................................................................................... 27
2.1.7 Infrared imaging .................................................................................................. 30
2.1.8 Hybrid millimetre-wave and infrared imager ..................................................... 31
2.1.9 Electromagnetic resonances and phased array .................................................... 32
2.2 Discussion .................................................................................................................. 33
2.3 Challenges and Research Perspective ........................................................................ 37
2.4 Conclusions ................................................................................................................ 40
Chapter 3: FEA OF GUNS AND KNIVES DETECTION APPROACHES ....................... 41
3.1 Earth Magnetic Field Distortion Method ................................................................... 42
3.1.1 Sensitivity to external shapes .............................................................................. 43
3.1.2 Sensitivity to inner shapes................................................................................... 43
3.1.3 Sensitivity to orientation ..................................................................................... 45
3.2 High Frequency EM Wave Illumination Method ...................................................... 46
3.2.1 Investigation of EM wave in weapon characterization ....................................... 46
3.2.2 Electromagnetic resonances ................................................................................ 51
3.3 Conclusions ................................................................................................................ 52
Chapter 4: NEW PROPOSAL FOR THE DETECTION OF CONCEALED WEAPONS:
EM WEAPON DETECTION FOR OPEN AREAS ............................................................ 53
4.1 Theory for Electromagnetic Weapons Detection and Identification.......................... 53
4.2 Proposed Concealed Weapon Description ................................................................. 60
4.3 Proposed Design ........................................................................................................ 61
4.3.1 Metal detector subsystem .................................................................................... 62
4.3.2 Detection & classification subsystem ................................................................. 75
4.3.3 Control and operator interface subsystem ........................................................... 78
4.4 FEA Simulations ........................................................................................................ 78
4.4.1 Sensitivity to outer shape .................................................................................... 78
4.4.2 Insensitivity to object orientation ........................................................................ 81
4.4.3 Sensitivity to size ................................................................................................ 83
4.4.4 Sensitivity to material composition..................................................................... 84
5
4.4.5 Target response to stand off ................................................................................ 85
4.5 Experimental Work .................................................................................................... 88
4.5.1 Driver circuit for UMFG ..................................................................................... 88
4.5.2 Testing of the proposed weapon detection. ......................................................... 94
4.6 Conclusions .............................................................................................................. 102
Chapter 5: CONCLUSIONS .............................................................................................. 103
Chapter 6: FUTURE WORK AND RECOMMENDATIONS .......................................... 106
REFERENCES................................................................................................................... 108
APPENDICES ................................................................................................................... 113
Appendix A: Hall Sensor Technology for CWD .......................................................... 113
A.1 Offset cancellation............................................................................................... 113
A.2 Design ................................................................................................................. 114
Appendix B: Cell Phone Radiation Levels ................................................................... 115
6
LIST OF FIGURES
Figure 2.1. The CWD developed by INL [7]. ...................................................................... 15
Figure 2.2. The principle of EMI technique for CWD [8]. .................................................. 16
Figure 2.3. PI metal detection. ............................................................................................. 17
Figure 2.4. CW metal detection. .......................................................................................... 18
Figure 2.5. Cumulative signal effects in active walkthrough weapon detector. .................. 19
Figure 2.6. Block diagram of a 3D steerable magnetic field (3DSMF) sensor system [12]. 21
Figure 2.7. Non-Linear Acoustic CWD [14]. ...................................................................... 23
Figure 2.8. Enhancement of Radar Cross Section in the resonance region [17]. ................. 24
Figure 2.9. MMW images (QinetiQ imaging system) [24].................................................. 27
Figure 2.10. THz reflection image of a person carrying a gun [26]..................................... 28
Figure 2.11. Atmospheric attenuation of THz rays against frequency [22]. ....................... 30
Figure 2.12. Image fusion [31]............................................................................................. 31
Figure 2.13. CWD system for open areas using phased arrays............................................ 32
Figure 2.14. EM methods for CWD in the EM spectrum. ................................................... 33
Figure 2.15. Diagram of challenges and research perspective of CWD. ............................ 38
Figure 2.16. Diagram of a tentative weapon detection system. ........................................... 39
Figure 3.1. Cross sectional of the samples. .......................................................................... 41
Figure 3.2. FE model of CWD based on earth field distortion. ........................................... 42
Figure 3.3. Magnetic flux density patterns of three samples with different outer section. .. 43
Figure 3.4. Magnetic flux density patterns of three samples with different inner section. .. 44
Figure 3.5. Target responses showing sensitivity to internal cross sectional change. ......... 44
Figure 3.6. Signal response of a square bore barrel rotated 0 and 45 degrees. .................... 45
Figure 3.7. FE model to investigate weapon response to EMW illumination. .................... 46
Figure 3.8. RCS response change of orientation of bore gun barrel. ................................... 47
Figure 3.9. Radiation patterns of scattered wave for (a) 0-degree rotated bore; (b) 45-degree
rotated bore. ................................................................................................................. 48
Figure 3.10. RCS response to change of orientation of external shape gun barrel. ............. 49
Figure 3.11. Radiation patterns of scattered wave for (a) 0-degree rotated barrel; (b) 45degree rotated barrel..................................................................................................... 49
Figure 3.12. RCS response to barrel shapes......................................................................... 50
Figure 3.13. Cross section of gun barrels employed in the FE simulation. ......................... 51
Figure 3.14. The RCS values against frequencies for each gun barrel. ............................... 52
Figure 4.1. Sphere with radius “a” and conductivity “ σ ” irradiated with an electromagnetic
wave in the form of step pulses. ................................................................................... 54
Figure 4.2. Components of the magnetic fields caused by currents induced in the sphere. 55
Figure 4.3. Four terms of Eq. (4.2) for s=1, 5, 10 and 50 (top). The full step-response decay
computed for summations of 8100 terms of Eq. (4.2) [45] (bottom). .......................... 57
Figure 4.4. Time constant from sensor signals in logarithm scale. ...................................... 58
Figure 4.5. Functionality of the proposed CWD system...................................................... 61
Figure 4.6. Proposed CWD structure. .................................................................................. 62
Figure 4.7. Magnetic field covering area from active part of UMFG (top plane) and return
path (bottom plane). ..................................................................................................... 63
Figure 4.8. Effective magnetic field covering area. ............................................................. 64
7
Figure 4.9. Routing the return path to the sides of the box maximizes the effective magnetic
field covering area. ....................................................................................................... 65
Figure 4.10. Pulse Control Circuit. ...................................................................................... 66
Figure 4.11. Layout of a block of intended CWD system including five UMFG................ 68
Figure 4.12. Time diagram of switch trigger signal (LM555 set a 20Hz, 10% duty cycle). 69
Figure 4.13. Cross sectional view of magnetic field from a coil loop (top); Cross sectional
view of magnetic field from UMF (bottom). Current flow into the paper................... 70
Figure 4.14. Magnetic field uniformity and Target signal response [12]. ........................... 71
Figure 4.15. UMFG and Loop Coil...................................................................................... 72
Figure 4.16. Magnetic field from UMFG and a loop coil. ................................................... 73
Figure 4.17. Sensor array distributions for the proposed CWD system............................... 75
Figure 4.18. Scanning of interrogation area for detection and classification....................... 76
Figure 4.19. Proposed algorithm for extraction of the Time Constant. ............................... 77
Figure 4.20. Samples for test. .............................................................................................. 79
Figure 4.21. 3-D Model (dimensions in metres). ................................................................. 79
Figure 4.22. Scattered magnetic field on y-axis (top) and z-axis (bottom).......................... 80
Figure 4.23. Model to test sensitivity to weapon orientation (domain dimension in metres).
...................................................................................................................................... 81
Figure 4.24. Scattered magnetic field sensed from a distance 15 cm (top) and 35 cm below
the gun (bottom). .......................................................................................................... 82
Figure 4.25. Model to test size sensitivity. .......................................................................... 83
Figure 4.26. Time constant response to size of guns. .......................................................... 84
Figure 4.27. Time constant profile of a gun made of steel, copper and iron. ...................... 85
Figure 4.28. Model to measure target response at different stand off.................................. 86
Figure 4.29. Signal amplitude of the target at different stand off. ....................................... 86
Figure 4.30. Gradient of target signals response at different stand off. ............................... 87
Figure 4.31. Driver circuit for UMFG (design 1). ............................................................... 89
Figure 4.32. Current and Voltage in UMFG during one excitation pulse (design1). .......... 90
Figure 4.33. Pulse current generator (design 2). .................................................................. 91
Figure 4.34. Current and Voltage in UMFG during one excitation pulse (design2). .......... 92
Figure 4.35. Back EMF reduction comparison between design 1 and 2.............................. 93
Figure 4.36. Current fall time comparison between design 1 and 2. ................................... 93
Figure 4.37. Search coil device. ........................................................................................... 94
Figure 4.38. Solder kit responses at different orientation. ................................................... 95
Figure 4.39. Zoomed views of solder kit responses in logarithm scale. .............................. 96
Figure 4.40. Target response signals from search coil (top). Zoomed view in logarithm
scale (bottom). .............................................................................................................. 97
Figure 4.41. AMR sensor. .................................................................................................... 99
Figure 4.42. Timing diagrams. ........................................................................................... 100
Figure 4.43. Time constant decays profile of metallic objects using AMR HMC1001..... 101
Figure 4.44. Close up of Time decay profile of metallic objects using AMR HMC1001. 101
Figure A.1. Offset cancellation. ......................................................................................... 113
Figure A.2. Hall Magnetometer Design. ............................................................................ 114
8
LIST OF TABLES
Table 2.1. Summary of the Different Technologies being developed for CWD. ................ 34
Table 2.2. Summary of Main Issues of CWD. ..................................................................... 35
Table 4.1. Magnetic Sensor Technology Field Ranges [46]. ............................................... 74
Table 4.2. Performance of Sensor Technologies. ................................................................ 74
Table 4.3. Coil parameters. .................................................................................................. 95
Table B.1. SAR level of cell phones. ................................................................................. 115
9
LIST OF ABBREVIATIONS
AMMW
Active Millimetre Wave
AMR
Anisotropic Magnetoresistive
AWG
American Wire Gauge
CCTV
Closed Circuit Television
CW
Continuous Wave
CWD
Concealed Weapon Detection
DC
Direct Current
EM
Electromagnetic
EMF
Electromotive Force
EMW
Electromagnetic Wave
FEA
Finite Element Analysis
GMR
Giant Magnetoresistance
HMF
Horizontal Magnetic Field
INL
Idaho National Laboratory
IR
Infrared
MMW
Millimetre Wave
NAC
Nonlinear Acoustic
PI
Pulse Inductive
PML
Perfectly Matched Layers
PMMW
Passive Millimetre Wave
RCS
Radar Cross Section
RF
Radio Frequency
SQUID
Superconducting Quantum Interference Devices
THz
Terahertz
UMFG
Uniform Magnetic Field Generator
WAMD
Wide Area Magnetic Detection
3DSMF
3D steerable Magnetic Filed Sensor
10
Chapter 1: INTRODUCTION
From year 2000 there has been an increased security threat in public areas. Today, terrorist
groups, hijackers and people hiding weapons are a constant and increasing threat. There is
an immediate requirement for law enforcement and homeland security to identify concealed
weapons, which may present a threat to official personnel and the general public. This
involves suicide bomb vests, handguns, knife blades and other threatening weapons.
Concealed weapons detection is one of the greatest challenges facing the law enforcement
community today [1].
Current security-detection systems, which include portals and hand-held devices for
detecting concealed weapons such as handguns, knives and explosives, are common in
controlled access settings like airports, entrances to sensitive buildings, and public events.
The presence of a portal weapon detection system and hand held devices in a security check
point warn in advance those individuals trying to hide threatening weapons. Current
security detection systems also need to be near an individual to work. They generally
provide sufficient warning when it comes to detecting a knife, but they cannot detect
weapons that can kill beyond arm’s reach. For instance by the time a handgun or a bomb
vest is detected, it generally is too close to be dealt with safely. It is clear that screening
people in this way using current security detection systems prior to entering secured areas is
ineffective in preventing some weapons from getting through. It is an almost impossible
task to achieve 100% success given the equipment screeners have available today. To
detect concealed weapons requires not only simple metal detectors in controlled
environments as used today, but detectors from a standoff distance on streets and any safety
critical environments. It is vital for law enforcement agencies to be able to detect and
respond to weapons at a sufficient distance to allow officers to make decisions and take
actions that deal safely with the situation.
There must be reliable ways to detect and identify not only metal but also non-metallic
weapons or other threatening objects that may be concealed under clothing or vehicles.
11
Many approaches have been attempted since terrorist attack to different targets in USA of
11th September 2001. The Department of Homeland Security established Homeland
Security
Advanced
Research
Projects
Agency
(HSARPA)
(http://www.dhs.gov/xres/grants/#1) to promote revolutionary technologies that would
address homeland security vulnerabilities. It is an important component of the US response
to terrorist attacks of all kinds. In the meantime, the UK in common with other European
countries has initiated several research programs on crime prevention including gun
detection. It is a technological challenge that requires innovative solutions in sensor
technologies and image processing.
This thesis has reviewed the recent progress of research and development of the CWD
technologies. As a part of the collaboration project partners (Manchester University,
Manchester Metropolitan University, Newcastle University and Queen Mary London
University) in the weapon detection research, this investigation is focused on providing a
cost effective and reliable solution employing technologies with working radiation
frequency within 0 and 1GHz. As a result of the research a new weapon detection system
based on pulse induction technology able to work in open areas without invading individual
privacy has been proposed. The eventual outcome of this work and other approaches at
several non-overlapping wavelengths could result in a reliable integrated multimodal
sensing system.
The thesis is organised as follows. Chapter 2 will review and discuss individual methods at
different frequencies within the wide electromagnetic spectrum. Chapter 3 presents Finite
Element (FE) based models for guns and knives characterisation using Passive (Earth field
distortion) and Active (EM wave illumination) weapon detector technologies. Chapter 4
presents the theory for the proposed CWD, system description, FE model simulations and
preliminary practical work.
In the last chapter the conclusions and future work are
presented.
12
Chapter 2: LITERATURE SURVEY
This chapter reviews recent developments in the area of CWD. These methods largely use
electromagnetic means including metal detection, magnetic field distortion, electromagnetic
resonance, acoustic and ultrasonic inspection, Millimeter waves (MMW), Terahertz
imaging, Infrared, X-ray etc. The advantages and disadvantages of these approaches are
discussed and research challenges and perspectives are presented. The chapter is organised
as follows: section 2.1 reviews and discuss individual methods working at different
frequencies within the electromagnetic spectrum; section 2.2 discusses the advantages and
disadvantages of these techniques; then research challenges and perspectives are presented
in section 2.3.
2.1 Review of Current CWD Technologies
Current CWD approaches range from the EM spectrum based detection to X-ray imaging.
X-ray imaging has been used for border control and luggage inspections in airports but,
because X-ray is harmful to humans, this technique will not be discussed in this review.
2.1.1 Metal detection using Earth magnetic field distortion
This technology is based on the passive sampling of the distortion of the Earth’s magnetic
field. The local aberrations in the magnetic field produced by ferromagnetic objects such as
guns and knives can be detected by extremely sensitive magnetometers. Detector systems
based on this technology exist in walk-through type devices and portable devices, currently
in development, which are called Gradiometer Metal Detectors [2].
The gradiometer metal detector system is a passive system and has been deployed as a
permanently installed part of a secure entry system for courthouses. It is a walk-through
type device that uses a series of vertically arranged gradiometers located on either side of
its portal. A gradiometer consists of two magnetometers connected electrically in what is
13
called differential mode. The gradiometer responds to changes in the local magnetic field of
the earth caused by a moving ferromagnetic object. The differential mode connection is
required to reduce the effects of normal background fluctuations that would otherwise
cause false alarms. The gradiometer portal is very similar to a zoned walk-through metal
detector. Each gradiometer pair responds to the presence of the ferromagnetic object and,
based on the magnitude of these interactions, the system displays the location of the metal
object within the portal. The system responds to moving ferromagnetic metal objects so that
placing the system near stationary steel objects will not affect its operation [3].
There are some drawbacks in the gradiometer systems:
1. The gradiometer systems become expensive if spatial resolution is an issue. The
resolution is dependent on the number of pair of gradiometers.
2. Those systems can only detect ferromagnetic weapons. Obviously some metal
weapons are not made of ferromagnetic metals. Knives made of stainless steel and
thread weapons made of aluminium, copper or brass can not be detected.
3. The system needs to be deployed in a permanent installation to avoid vibration
which could lead to false alarms. It may be possible to use extra-hardware to keep
tracking of position of the magnetometers, but it implies extra cost. Portable devices
intended to be carried by officers or mounted on vehicles are in development [4,5]
CWD and Classification based on Earth’s magnetic field distortion is a new application of
existing magnetometer sensors. Sensors in the system simultaneously collect data,
providing a top-to-bottom profile of an individual. Reasonable suspicion will be dictated by
the location and magnitude of the recorded magnetic anomalies. A database of magnetic
signatures has to be established through the collection of magnetic profiles of a variety of
weapons in differing locations and a number of non-weapon personal artifacts. These
signatures will later be used in the analysis that will determine the presence, location, and
potentially, type of weapons. Recent analysis schemes include advanced signal processing
algorithms that perform pattern recognition and calculate the probability that the collected
magnetic signature correlates to a known database of weapon versus non-weapon responses
using neural networks. In [6] a method to reduce false alarms was developed by using a
Joint Time Frequency Analysis digital signal processing technique.
14
From 2003, the Idaho National Laboratory (INL) has developed a new generation of CWD
system – a passive portal that senses disturbances in the ambient Earth’s magnetic field. It
uses gradiometers to detect and locate position of metals. The CWD system uses 16
magnetic gradiometer sensors, arranged on both sides of the portal aperture. Data are
collected from each of the gradiometers, and the change in the magnetic field over ambient
background is determined. After the individual sensor responses are collected, the data
from all of the sensors are processed to determine the location and size of a detected object.
Figure 2.1 shows the graphical interface of the INL portal provided to the operator. It uses
freeze-frame video capture technology and places filled circles — dependent upon the
number of items detected — over the video image indicating where suspected weapons
may reside on a person. The circle sizes vary in proportion to the strength of the measured
signal. This approach is normally used for a controlled environment e.g. walk through gates
at airports.
Figure 2.1. The CWD developed by INL [7].
15
2.1.2 Inductive magnetic field methods
The principle behind Electromagnetic inductive techniques for CWD is shown in Figure 2.2.
An alternating current over the frequency range of 5 kHz-5MHz generates a time-varying
magnetic field around the coil. This field induces currents in a nearby metal object which,
in turn, generate a time-varying magnetic field of their own. These fields induce a voltage
in the receiver coil which, when amplified, reveal the presence of the metal object or target.
Figure 2.2. The principle of EMI technique for CWD [8].
There are two broad categories of EM induction approaches for CWD, which are classified
by the type of magnetic field generated by its excitation coil: Pulse induction (PI) and
Continuous wave (CW).
Pulse Induction (PI) detectors typically generate a transmitter current which is turned on for
a time, and is then suddenly turned off. The collapsing field generates pulsed eddy currents
in the target, which are then detected by analysing the decay of the pulse induced in the
receiver coil. Conductive objects show a unique time-decay response, so a signature library
of these objects can be developed. When concealed metals are encountered, its time-decay
signature can be compared to those in the library and, if a match is found, the object can
potentially be classified [8].
16
Continuous Wave (CW) detectors generate a transmitter coil current which alternates at a
fixed frequency and amplitude. Small changes in the phase and amplitude of the receiver
voltage reveal the presence of concealed metal targets.
Pulse Induction approach detects metal objects by analysing the time-decay response of the
pulse induced in the receiver coil, whereas Continuous Wave approach does this based on
changes in the amplitude and phase of the receiver coil voltage.
The following plots illustrate the two concepts of inductive metal detection methods by
showing graphs of received signals with position on the horizontal axis. The scales are
notional as the intention is just illustrate concepts. Figure 2.3 shows a change in decay rate
of the signal received by PI detector (blue) with respect the reference signal (red) when
passing over a metal located about position 10. Figure 2.4 shows a change in phase and
amplitude of the signal received by CW detector (blue) relative to transmitted signal (red)
when passing over metal located about position 10.
Position
Figure 2.3. PI metal detection.
17
Position
Figure 2.4. CW metal detection.
Most of these detection systems are used in airports throughout the world for weapon
detection to protect restricted areas of the airports. Active Walkthrough Weapon Detectors
like portal shown in Figure 2.5 belong to these systems. Figure 2.5 shows how signals from
harmless objects are processed by active walkthrough metal detector. Pulsed magnetic
fields produced by a transmitter (left panel) pass through the person being screened to the
receiver panel. As a pulse passes through a metal object it causes eddy currents to be
induced in its surface. Once the pulse passes beyond the object its field strength decreases
and eddy currents in the metal starts to decay. As the currents decay they generate a low
intensity magnetic field. This field is detected by the receiver coils with ferrite core to
enhance the signal [9], and then processed into an electronic signal.
18
Figure 2.5. Cumulative signal effects in active walkthrough weapon detector.
Walkthrough weapon detectors based on inductive magnetic field differ from those based
on gradiometers as they are active methods and can not identify the position of an object,
whereas the latter ones are passive and can locate a hidden object.
CWD using this method has a problem with accuracy. A human body has a small
conductivity thus the sensitivity of the detector cannot be high enough to be able to detect
the magnetic field coming from the body. When dealing with materials that are not of very
high conductivity (in general non-metallic materials) [10] or of very small dimensions, the
human body could give a stronger signal than the material. This would cause the material to
pass undetected. Thus, such CWD devices do not have a great reliability.
19
One of the most recent applications of inductive magnetic field approach is called WideArea-Metal Detection (WAMD) [11]. This is a CWD system which enables the detection of
metal and can be used to locate people carrying metal weapons in a crowd that may require
further investigation. WAMD use pulse induction to generate time varying magnetic field
and use the time decay of the magnetic field in the target as a signature. WAMD system
sensors use a 3D steerable magnetic field sensor (3DSMF) [12] to create and measure 3D
time decay response of the magnetic field of the object. The strength of the magnetic field
vector in 3D can be controlled by varying the current in each antenna element and
employing superposition of the fields of each antenna. In this way false rate for
classification decreases because most of target identification algorithms assume that target
is excited with uniform field, which differs by far from complex magnetic field excitation
of coil loops.
The WAMD sensor system is based on Horizontal magnetic field (HMF) metal detection
antenna concept. This concept is based on sensors that attempt to generate threedimensional magnetic fields and measure the target’s 3D response by using magnetic field
antennas that have complex spatial magnetic field distributions. The 3D steerable magnetic
field sensor (3DSMF) (see Figure 2.6) orients the excitation magnetic field into the primary
axis of the target. Once the primary axis is found, the antenna’s magnetic field is rotated
into the secondary axis of symmetric objects. For no symmetric objects, the 3DSMF sensor
measures the object’s response in 4 π steradians. The classification algorithm then tries to
match the object’s 3D response to a target library.
To better understanding the above concept basic physics need to be reviewed. It is known
that, for an infinitive conducting sheet current in free space, the magnetic field is in the
direction perpendicular to the sheet current. Thus the sheet current is a horizontal magnetic
field (HMF) generator or antenna. This magnetic field is constant. To take advantage of this
feature, we create a practical approximation of an infinitive sheet current by placing closely
spaced current-carrying wires in a plane (See illustration of three 3HMF in Figure 2.6). The
result is a magnetic field with a relatively uniform horizontal shape.
20
To conceptualize the 3DSMF sensor, we need only to imagine two single-axes HMF
antenna co-located at right angles to each other. This arrangement forms a 2D horizontal
field-generating antenna. The third dimension to the magnetic field is created by adding a
horizontal loop antenna to the two HMF antennas as shown in Figure 2.6. Thus a magnetic
field in 3D is created by varying the current in each antenna and by employing
superposition of the fields of each antenna.
y
x
z
Z-axis
antenna
X-axis HMF antenna
Y-axis HMF antenna
Computer control
( current control
system)
Figure 2.6. Block diagram of a 3D steerable magnetic field (3DSMF) sensor system [12].
2.1.3 Acoustic and ultrasonic detection
The detection of weapons using acoustic and ultrasonic detectors is dependent on the
acoustic/ultrasonic reflectivity of materials that make up an object and the shape and
orientation of the object. Basically, hard objects provide a high acoustic/ultrasonic
reflectivity and soft objects a small reflectivity. The important detection parameters for
these technologies are size of the target, diameter of the detector antenna, wavelength of the
wave emitted, and the emitted power. The antenna size and wavelength affect the size of
the smaller object that can be detected. Acoustic/ultrasonic power will not travel in a
vacuum; it is attenuated less as its travels in dense medium (solids and liquids) and is
21
attenuated more as it propagates in the air. Also humidity in the air reduces the attenuation.
The attenuation is frequency dependent; it is greater for higher frequencies. Therefore, there
is a trade-off between the required spatial resolution and attenuation. Ultrasonic detectors
operate from 40 kHz until frequencies well below 1 MHz because of increasing attenuation
at higher frequencies. Acoustic detectors operate at audio frequencies.
Ultrasonic (high frequencies) detectors [13] have problems penetrating thick clothing
whereas acoustic (low frequencies) detectors can propagate more easily through clothing
and “see” a concealed object.
Conventional acoustic and ultrasonic based detectors are sensitive to hard objects in general,
and thus, it can not differentiate between weapons and innocuous hard objects.
Consequently devices based on these technologies produce many false-positive detections.
Presently, there are hand held weapon detectors based on acoustic wave phenomena
operating at 1m -5m distance. A combination of radar and ultrasound is being explored by
JAYCOR, advanced-technology Company providing services to US Department of
Defence. The system produces an ultrasound image and can operate at 5m-8m distance.
From the combination of the ultrasonic/acoustic approach, a nonlinear acoustic method
(NAC) for CWD has been developed [14]. Figure 2.7 shows the principle. This technology
uses ultrasonic beams of frequencies f1 and f2 to project sound onto a small spot area on a
person at a distance and convert that energy probe from ultrasound to audio frequencies.
The non linear interaction in the mix zone produces the following frequencies: f1, f2, f1-f2,
f1+f2. The difference frequency (f1-f2), tuned in the audio range, is used to interrogate the
subject with a beam able to penetrate clothing. Parametric acoustic arrays [15] can be used
to produce nonlinear acoustic effects and the detection of a concealed weapon can be based
on signatures. The nonlinear acoustic method for CWD uses correlation algorithms to
perform patterns matching and classification to display the nature of a hidden weapon. In
general this technique is harmless because does not involve ionizing radiation, is sensitive
to metals and non-metals, and is able to penetrate clothing [13]. However, fast scanning is
required for ultrasonic beams to focus on and find a target.
22
Detection
f1
Ultrasonic beams
f1-f2
Acoustic beam
f2
Figure 2.7. Non-Linear Acoustic CWD [14].
2.1.4 Electromagnetic resonances
This is an active technique which uses EM resonance as “fingerprints” or signature to
distinguish weapons and nuisance objects. EM resonances in the objects are determined by
its size, shape, physical composition. These resonances occur over the range 200MHz2GHz. The detector uses radar to sweep through this range of frequencies and record the
resonant response.
The radar return or resonance based scattering exhibits some features which makes it
attractive for object identification schemes:
1. Scattering return is larger in the resonance region.
2. The natural resonances seen in a scattered return are independent of the orientation of
the object.
3. A few natural resonances characterize an object over a large frequency band.
4. An object’s resonance patterns uniquely identify it within a class of objects.
23
To induce a resonant response in an object, it is necessary to illuminate it in the frequency
band of the natural resonances. The radar cross section (RCS) scattering falls in three
regions, depending on the ratio of wavelength λ to body size L. The three regions are
Rayleigh, resonance and optics corresponding to λ much higher than L, λ proportional to
L, and λ much lower than L [16]. Figure 2.8 shows the RCS of a sphere of radius a as a
function of its circumference measured in wavelengths, 2πa / λ = ka , where k is the
wavelength number. RCS has been normalized to the projected area of the sphere, πa 2 .
When the wavelength is much greater than the sphere circumference, its RCS is
proportional to a 2 (ka) 4 , which shows that RCS increases as fourth power of frequency and
sixth power of radius. When the circumference length is between 1 and 10 wavelengths the
RCS exhibits oscillatory behaviour with several peaks which correspond to the natural
electromagnetic resonance of the sphere [17]. When the circumference is large compared to
a wavelength, the oscillatory behaviour fades out and the RCS is now independent of
frequency and equal to physically cross section of the sphere.
Figure 2.8. Enhancement of Radar Cross Section in the resonance region [17].
24
The theory of operation of this method is as follow. The target space can be illuminated by
either a pulse or swept frequency source [2]. The signal reflected by objects in the target
space provides an electromagnetic signature (EM resonances), a unique spectrum, for that
object. The object signatures are then compared to known signatures to determine whether
or not objects in the target space are threat items. Neural network processing is used to
classify the difference between weapons and nuisance objects [18]. The person carrying the
object will also exhibit a unique electromagnetic signature which must be subtracted from
the composite person-object signature.
This method has an operation range of 6.096m, allows the detection of concealed weapons
even behind a human body, operates at a safe limit power for human exposure and does not
invade privacy of individuals.
The problem found with this technique is noise corresponding to the signature of people.
Signatures of individuals vary from one to another and they also vary when a weapon is
present. Unfortunately, the signature of an individual with a weapon is very similar to one
without a weapon and so there is a problem with classification and a high rate of false
alarm.
During 5th International Conference on Antenna Theory and techniques on May 2005 an
approach based on EM resonance was presented incorporating circularly polarised, focused
antenna arrays [19] to produce a very small common antenna beam footprint on the suspect
under surveillance. This type of antenna makes the system less sensitive to the orientation
of the concealed weapon. For the weapon classification, signal processing using Fourier or
a Wavelet transform [20] is used.
25
2.1.5 Millimetre waves (MMW)
MMW based screening systems are basically of 2 types: passive and active. Passive sensors
simply observe and report whatever detects in local environment. In the RF spectral range,
natural surfaces will emit different amounts of radiation depending on parameters such as
temperature and emissivity. In addition, metals are strongly reflective to RF, which reduces
a metal surface’s emissivity and allows it to produce reflections of other sources in the
scene with the most significant being the sky. Passive sensors have the great advantage of
producing valuable information without emitting any signals from people.
Active sensors typically stimulate the environment by generating and emitting known
signals. These signals propagate out to the objects or targets of interest, interact with them,
and reflect or scatter energy back to the sensor. Because the self-generated signals have
known properties, it is often possible to use signal processing to extract very weak emitted
target signals from competing sources of noise.
According to safety views, MMW systems utilise very low radiation power to generate
detection capability. The system uses radiation power level 10,000 times less than that of a
cell phone (maximum specific absorption rate (SAR) level of cell phone at 2009 is
1.6~2W/kg depending on region (see appendix B). The use of millimetre wave technology
eliminates issues associated with use of ionizing radiation such as those seen with x-ray
systems [21-23].
Figure 2.9 show how MMW images look. Passive millimeter wave (MMW) sensors
measure the apparent temperature through the energy that is emitted or reflected by sources.
The output of the sensors is a function of the emissivity of the objects in the MMW
spectrum as measured by the receiver. Clothing penetration for CWD is made possible by
MMW sensors due to the low emissivity and high reflectivity of objects like metallic guns.
26
Figure 2.9. MMW images (QinetiQ imaging system) [24].
2.1.6 Terahertz (THz) imaging
The THz imaging technique is based on the use of THz electromagnetic waves to
spectroscopically detect and identify concealed explosives, chemical/biological agents, and
metals through characteristic transmission or reflectivity spectra in the THz range.
Generally, non-polar and non-metal solids are at least partially transparent and reflective to
waves of frequencies 0.2 to 5 THz. Different materials have different effects on the THz
wave. Typical clothing items and paper and plastic packaging should appear transparent in
the THz regime whereas metals completely block or reflect THz waves. Ceramic guns and
knives would partially reflect the THz signal. Skin, because of its high water content,
would absorb nearly all T-Rays with the energy being harmlessly dissipated as heat in the
first 100 microns of skin tissue. A THz reflection image of a person as shown in Figure
27
2.10 would show the outline of clothing and the reflection of objects beneath, such as
weapons or key chains, but the person’s skin would appear substantially dark [25].
Figure 2.10. THz reflection image of a person carrying a gun [26].
There are some advantages of this technique which are attractive for CWD:
1- The spatial resolution of THz waves is excellent for CWD. THz waves can separate
objects less than 1mm apart, which is more than enough to tell a weapon apart from its
surroundings. This resolution is roughly 10 times better then that of Millimetre Waves
(MMW), due to the smaller wavelength.
2- Many materials of interest for security applications including explosives, chemical
agents, and biological agents have characteristic THz spectra that can be used to
fingerprint and thereby identify these concealed materials. .
3- THz waves are non-dangerous. The penetrating ability of THz waves may seem
harmful to health, similar to the x-ray, but in actual fact, it is totally harmless as T-rays
are non-ionizing - they do not alter molecules in the air or in humans. On the other hand,
28
x-rays can cause diseases such as cancer in humans if over-exposed and could cause
damage to the eyes if shown directly [25].
However, THz imaging has some issues which have to be eventually solved:
1- The feasibility of the use of terahertz imaging for the detection of concealed weapons is
questionable. The main issues are cost, processing and complexity. THz systems are
expensive because it requires special power sources. The recent introduction of near
infrared ( =800nm) femtosecond laser as radiation source, has helped to bring the cost
of such systems down below 50K Euros [27]
2- Though T-ray detection is already possible, it is not, by far, perfect yet. Using close
range imaging, it is still difficult to develop video output because the scanning is still
slow, leading to a poor frame rate.
3- The most significant limiting factor of the capabilities of T-ray imaging at stand-off
range (3m to 100m) is the atmosphere which causes attenuation and turbulence to the
waves [28]. However oxygen, water and other components of the atmosphere absorb
THz power at specific frequencies only, leading to high attenuation at certain
frequencies (Figure 2.11). Therefore, all stand-off THz imaging systems are tuned to
use frequencies that are less attenuated.
4- Proper guidelines for using these imaging systems have to be finalised and put into
action, as they might be harmful at some specific conditions of exposure [29] or have
legal implications.
5- Privacy invasion issue, because T-ray can penetrate clothes.
29
Figure 2.11. Atmospheric attenuation of THz rays against frequency [22].
2.1.7 Infrared imaging
Infrared imaging is another commonly used detection system. Human bodies as well as any
other material emit radiation provided they are at a temperature above 0°K. The wavelength
of the radiation peak is dependent on the temperature of the body, and the total power
emitted from the body is dependent on the size and emissivity of the body. Most infrared
sensors are designed to have peak sensitivity near the peak emission wavelength of human
body which is 10 μm . This technology is normally used for a variety of nigh-vision
applications.
Infrared radiation emitted by people is absorbed by clothing. This absorbed radiation heats
the clothing and is then re-emitted by the clothing. Consequently, the image of a concealed
weapon will be blurred, at best, assuming the clothing is tight and stationary. For normally
loose clothing, the emitted infrared radiation will be spread over a larger clothing area
thereby significantly decreasing the ability to image a weapon. The difficulty in observing
an infrared signal of a concealed weapon becomes worse as the weapon temperature
approaches that of the body [30].
30
2.1.8 Hybrid millimetre-wave and infrared imager
Hybrid millimetre-wave and infrared imagers are systems that exploit both mm-wave and
infrared imaging techniques. The images from these two separate imaging subsystems are
brought together algorithmically in what is called a fusion process. The fusion of the mmwave and infrared images results in a fused image [31].
The fusion process requires that the images be aligned properly (Registration). Before
alignment each image is filtered to remove extraneous or noisy data and enhanced to
augment certain image features. This process is illustrated in Figure 2.12.
Figure 2.12. Image fusion [31].
The most straightforward approach to image fusion is to take the average of the source
images, but this can produce undesirable results such as a decrease in contrast. Many of the
advanced image fusion methods involve multi-resolution image decomposition based on
the wavelet transform [32]. The fused data is the result of infrared and MM-wave images. If
the infrared image cannot contain information on the concealed item, it will provide no
useful information. In most situations, clothing will prevent acquisition of any information
on a concealed weapon. Consequently, the fused image provides no further information on
the concealed weapon than that provided by the MM-wave imager. MM-wave images take
a few seconds to acquire and during this time the person must be stationary. The fused
31
image may exhibit no more than a MMW image of the concealed weapon with a highresolution image of the surface of the clothing. However, the infrared image does facilitate
location of human subjects, on which an object may be hidden, and this may accelerate the
subsequent process of detecting a weapon on the subject.
2.1.9 Electromagnetic resonances and phased array
A portable CWD systems based on finding electromagnetic resonances was developed
with relative good performance [17]. This concept can be adapted in order to screen people
in an open area. This novel idea is based around a phased array located in the floor of the
scanning place, with a sensor array in the roof and a phase shifter control system to steer
the direction of the illuminating beam. Using a phased array [33] as a transmitter offers the
advantage of illuminating the metal body in multiple directions and in any desired
direction.
This might give evidence of electromagnetic resonances as well as extra
information regarding the location of the target.
Hence, by finding electromagnetic
resonances and keeping track of the beam as it scans, it should be possible to pinpoint the
concealed weapon. The system could operate as shown in Figure 2.13. The arrays of
sensors are distributed in a grid pattern in order to show the location of gun/knife carrying
people.
Sensor array
Main direction of EM
wave
Floor
line
Phased array antenna
Figure 2.13. CWD system for open areas using phased arrays.
32
2.2 Discussion
After reviewing different detection methods for CWD, a summary is illustrated in Figure
2.14 and tables 2.1 and 2.2. Figure 2.41 shows EM spectrum used to illuminate the
detection space by different CWD methods. Table 2.1 gives a brief summary of the
technologies being developed for CWD, including issues such as type of illumination of the
target, type of energy used for interrogation of the detection space, portability and
proximity of use. Table 2.2 gives a summary of the main issues of CWD methods,
including the parameter to be assessed, distance achievable, attenuation factor, depth of
penetrability, and detectable weapons.
Figure 2.14. EM methods for CWD in the EM spectrum.
33
Table 2.1. Summary of the Different Technologies being developed for CWD.
DESCRIPTION
ILLUMINATION PROXIMITY PORTABILITY
ACOUSTIC OBJECT DETECTOR
active
far 1
portable
active
near 2
portable
active
near
portable
IMAGING PORTAL
active
near
portable
METAL OBJECT LOCATOR
passive
far
portable
active
far
portable
active
far
portable
MM-WAVE RADAR DETECTOR
active
far
portable
EM PULSE DETECTOR
active
far
portable
MM-WAVE IMAGER
passive
far
portable
IR IMAGER
passive
far
portable
WALK-THROUGH METAL OBJECT
DETECTOR
HAND HELD METAL OBJECT
DETECTOR
PULSE RADAR/ SWEEP FREQUENCY
DETECTOR
THZ-WAVE
IMAGER
1
More than three-meter-range
2
Less that one-foot-range
34
Table 2.2. Summary of Main Issues of CWD.
DETECTED
CWD
ASSESSED
OPERATING
ATTENUATION
TECHNIQUE
PARAMETERS
RANGE (M)
FACTORS
PENETRABILITY
WEAPON
CONTENTS
EM
EM natural
RESONANCES
frequencies
Atmospheric
wavelength
metal and
conditions
dependent
non-metal
~8
air
medium
3~50
water
medium/low
4.5~7.62
water
high
<3
-
high
> 10
Impedance,
NON-LINEAR
ACOUSTIC
resonance,
non- linear
acoustic
metal and
non-metal
interactions
THz
TERAHERTZ
radiation
IMAGING
absorbed or
metal and
non-metal
reflected
Emissivity
MMW
and
brightness
metal and
non-metal
temperature
INDUCTIVE
Conductivity,
METHODS,
permeability
MAGNETIC
(magnetic
SIGNATURE
signatures)
Metal
objects
There is a large gap in the current detector market. Most pre-used gun detection is through
intelligence and surveillance, along with targeted stop and search. CCTV does not often
assist and is only really effective post event in providing identification and evidence, – i.e.
it is not pro-active.
The police do not have a deployable portable detector that can
unambiguously determine whether someone is carrying a gun. Such a device would assist
them in producing more effective stop and search strategies providing the number of false
35
positive events were kept within acceptable limits. Currently, research effort into
developing a sensor is concentrated in large parts of the EM spectrum, plus acoustic and
magnetic detectors. Focus now is on stand-off detection rather than proximity detection
with portals or hand held metal detectors [34].
Wild [35] has developed a portable ultrasound concealed weapons detector that works at
distances of up to 6m. There are reports of its partial effectiveness, and it is easily
deployable. However, this makes no use of the resonant cavities present in a gun to
generate a unique gun signature and it is falsely triggered by non-gun items such as leather
wallets. Millimetre waves are known to penetrate clothing well, and both passive and active
MMW imagers are being developed. These devices are generally large and more suited to
portals at airports; although a passive MMW video surveillance camera has been developed
[36]. Active through wall imaging systems, operating at 94 GHz, have been developed that
are capable of identifying a gun in the hand of a suspect inside a room. At 94 GHz
microwave radiation is attenuated but not scattered by building material. It is difficult to
identify a weapon on MMW image, but some authors concentrate on fusing images of
MMW and Infrared images to better reveal hidden guns [37] with some success.
THz sensing and imaging systems are currently being developed that are capable of
forming 3D images of concealed weapons [38]. However, the penetration of THz through
the atmosphere for stand-off detection and through some types of clothing leads to poor
results (e.g. a wool sweater produces diffuse scattering of THz, leading to the blurring out
of any concealed gun). Infrared imagers can potentially be used for detecting concealed
guns at night. A challenge identified for infrared imager is clothing radiation absorption.
The absorbed radiation heats the clothing and is then re-emitted by it, spreading over a
large area and decreasing the ability to image the weapon.
Research is also ongoing into Wide-Area metal detectors designed to locate people in a
crowd potentially carrying guns [39]. This technique utilises a pulsed magnetic field
radiating across a large area, coupled with a receiver and CCTV to identify suspects. It
does rely on CCTV operators to observe potential suspects once an alarm is triggered. At
36
the moment there is a terahertz imaging system commercially available which is sold by
TeraView Ltd, in Cambridge, UK. The imager uses photomixing technology [40, 41] with a
femtosecond pulsed laser illuminating low temperature GaAs as Terahertz source. Finally,
work is ongoing on portable and handheld MMW radar detectors that utilise frequency
modulated CW radar and complex signal processing to form an image of a concealed
weapon [42]. A small body of work does exist that has systematically examined the
opacity or otherwise of fabrics and other materials to microwaves and THz but these are
not in a form that would easily assist the police in the deployment of suitable sensors for
given conditions.
2.3 Challenges and Research Perspective
For development of deployable concealed weapon detectors, three issues should be solved:
sensing distance, high penetration and weapon recognition. Enough sensing distance is
required for open, outdoor environments e.g. high streets. High penetrations are required
for detection of metals and explosive under clothes or garments. Alternatively, penetration
is also required to detect targets beyond walls. Robust weapon recognition is required for
various weather conditions and multiple objects or crowd environment. The above review
has identified current detection methods are complementary approaches. No single
approach can meet all the requirements for a comprehensive CWD.
To solve these issues, taking advantage of different CWD technologies, a combination of
these is proposed.
37
Perspective of Research
Comprehensive CWD system
Challenges of CWD
system
Open and Outdoor
Environments
Large
Sensing
Distance
High
Penetrability
Complementary
CWD
approaches
(Sensor Fusion)
High penetrability:
clothes, garment, walls
Robust
Weapon
Recognition
Robust under weather conditions,
crowd environments, low false
alarms
Figure 2.15. Diagram of challenges and research perspective of CWD.
Figure 2.15 shows a diagram of the CWD research perspective arising from challenges of a
deployable CWD. Large sensing distance, high penetrability and a robust weapon
recognition system can be achieved by using complementary approaches. A comprehensive
CWD system is expected working in this direction.
To broaden and improve weapon recognition to explosives, chemical agents, and biological
agents, the characteristic THz spectra of these threats can be exploited. The spectra can be
used to fingerprint and thereby identify these concealed materials.
The literature survey has identified that current detection methods are complementary
approaches and so the best approach is a combination of some of them. A tentative weapon
detection system is showed in the next diagram (Figure 2.16). Complementary CWD
approaches in low and high frequency of EM spectrum (EM induction methods, Earth’s
magnetic field distortion, and EM resonance) could be used to meet all requirements for
comprehensive concealed gun/knife detection system
38
Figure 2.16. Diagram of a tentative weapon detection system.
39
2.4 Conclusions
A general concealed weapon detector collects data about the detection space and transmits
this information to the operator. Different sensor technologies are employed to capture
different sort of radiating energy (magnetic field, electromagnetic wave, acoustic wave, and
ultrasonic wave) that contains information about the detection space or about objects in the
detection space. Some technologies provide an image and other an indication regarding the
objects inspected.
All technologies show advantages and disadvantages in issues such as operating range,
material composition of the weapon, penetrability and attenuation factors. Each method
has some advantages over another.
It is clear that no single method can meet all
requirements for a comprehensive CWD system. Most of the methods described here are
complementary.
There are several challenges to the development of a comprehensive concealed weapon
detector. Three issues are the most critical and must be solved: Detection at a distance for
open and outdoor environments; High penetration for detection of weapons, explosive
under clothes, back or even walls, and a robust weapon recognition for various weather
conditions, crowd environment and able to discriminate accurately threat/non-threat items.
New CWD systems need to be an amalgamation of the techniques mentioned above
permitting a reduction in the number of false-positive trigger rates.
40
Chapter 3: FEA OF GUNS AND KNIVES DETECTION
APPROACHES
Following the review of recent technologies related to CWD, FEA is applied to study the
detection capability of CWD technologies working with radiation of 0-1Ghz frequency,
which is the part of EM wave spectrum that this research is focus on. In this chapter for the
given EM spectrum, the study of two CWD technologies has been chosen. The first study is
done on technology based on sensing disturbances in the earth magnetic field and the
second one on technology based on Radar Cross Section (RCS) resonance response under
EM wave illumination. In each case the sensitivity of the method to external and inner
characteristics, and orientation of the weapon is analysed in order to study detection
capability. To simplify the analysis and without loss of generality, gun barrels and knives
have been modelled as uniform bars with different external and inner shape cross sections
as shown in Figure 3.1.
Figure 3.1. Cross sectional of the samples.
Figure 3.1 shows the samples used in this test. Top row show different cross sectional of
gun barrels. Bottom row show external shapes: tubular barrels (diameter 17mm), knife
(0.8mm width) and square barrel (17mm side).
41
3.1 Earth Magnetic Field Distortion Method
In this section the local disturbances in the Earth’s magnetic field (≈50 μT ) arising from
the presence of different sort of gun barrels and knives is analysed.
In order to emulate the Earth’s magnetic field magnetic potential condition is applied to a
finite volume (box: 1m x 0.5m x 0.5m). The top boundary is set to a magnetic potential of
1989A whereas the bottom boundary is set to a magnetic potential of zero (see Figure 3.2).
This setting corresponds to a vertical magnetic flux density of about 50 μT .
Domain dimensions in metres
20 cm
Magnetic flux density
measured along this axis
Figure 3.2. FE model of CWD based on earth field distortion.
Magnetic flux density norm is measured along a line in the axial direction of the
interrogating weapon located 20cm below this (see Figure 3.2). Potential use of the
resulting magnetic flux density profile as distinguishable magnetic signatures is analyzed in
the next sub-sections.
42
3.1.1 Sensitivity to external shapes
Perturbations in the earth’s magnetic field caused by weapons (bars for this model) with
different external shapes show different amplitude pattern when magnetic flux density
norm is measured. It could be used as a magnetic signature for weapons with different
external shapes. Figure 3.3 shows Magnetic flux density pattern for three samples.
-5
5.045
x 10
5.04
Magnetic flux density norm (T)
5.035
tubular barrel
knife
square barrel
5.03
5.025
5.02
5.015
5.01
5.005
5
0
0.02
0.04
0.06
0.08
0.1
0.12
position (m)
0.14
0.16
0.18
0.2
Figure 3.3. Magnetic flux density patterns of three samples with different outer section.
3.1.2 Sensitivity to inner shapes
Perturbations in the earth’s magnetic field caused by gun barrels with the same external
shape but different bore show a slight difference in amplitude pattern. This means the
method is sensitive to inner shapes and could be distinguish different classes of guns.
Figure 3.4 and 3.5 shows Magnetic flux density norm distribution for gun barrels with
different bore shape.
43
-5
5.044
x 10
Magnetic flux density norm (T)
5.042
5.04
5.038
5.036
5.034
4-groove bore
8-groove bore
octagon bore
5.032
5.03
5.028
0
0.02
0.04
0.06
0.08
0.1
0.12
position (m)
0.14
0.16
0.18
0.2
Figure 3.4. Magnetic flux density patterns of three samples with different inner section.
-5
5.05
x 10
smoothbore gun barrel
square bore gun barrel
Magnetic flux density norm (T)
5.045
5.04
5.035
5.03
5.025
5.02
5.015
0
0.02
0.04
0.06
0.08
0.1
0.12
position (m)
0.14
0.16
0.18
0.2
Figure 3.5. Target responses showing sensitivity to internal cross sectional change.
44
3.1.3 Sensitivity to orientation
The earth magnetic field distortion caused by a square gun barrel at 0 degree and 45
degrees rotation around its axial axis is measured as shown in Figure 3.6. The signals
measured for these two scenarios show slight different pattern, especially in the middle of
axial length. Thus this technology shows sensitivity to orientation of the weapon which
makes more difficult weapon classification process.
O deg
45 deg
-5
5.03
x 10
Magnetic flux density norm (T)
5.028
5.026
5.024
5.022
square bore 0 degree rotation
square bore 45 degree rotation
5.02
5.018
0
0.02
0.04
0.06
0.08
0.1
0.12
position (m)
0.14
0.16
0.18
0.2
Figure 3.6. Signal response of a square bore barrel rotated 0 and 45 degrees.
45
3.2 High Frequency EM Wave Illumination Method
3.2.1 Investigation of EM wave in weapon characterization
The radar cross section (RCS) is a measure of sensitivity of an object to reflect illuminated
signal in the direction of the receiver. In the next model RCS response is used to
characterize guns and knives under EM wave illumination. The sensitivity of RCS respond
at shapes (external and internal) and orientation is analyzed. Since the scattered wave from
objects radiates in every direction, the ‘perfectly matched layers’ i.e. PMLs [43] are set up
in cylinder coordinates. The modeling domain is 200mm by 200mm (see Figure 3.7). The
frequency varies from 3GHz to 30GHz with a wavelength step of 10mm. The standoff
distance is 90mm from the symmetric axis of gun barrel.
Surface: Scattered electric field, z component [V/m]
Weap
Measuring
Figure 3.7. FE model to investigate weapon response to EMW illumination.
46
•
Orientation of inner shape vs. RCS
The orientation of non-circular bore and material of gun against RCS pattern is analyzed.
To simplify the problem, the bore shape is defined as a square. Two orientations i.e. 0
degree and 45 degree with respect to the radiation direction of incident wave are compared.
The results are shown in Figure 3.8.
2.5
square bore; 0 degree; Brass
square bore; 45 degree; Brass
square bore; 45 degree; Steel
2
RCS
1.5
1
0.5
0
0.5
1
1.5
Frequency (Hz)
2
2.5
Figure 3.8. RCS response change of orientation of bore gun barrel.
47
3
10
x 10
(a)
(b)
Figure 3.9. Radiation patterns of scattered wave for (a) 0-degree rotated bore; (b) 45-degree rotated
bore.
As can be seen in Figure 3.8, the RCS pattern is insensitive to material and the orientation
changes of gun and barrel bore, respectively. It is because the conductive objects within
EMW is quite reflective and can hardly penetrate into the barrel to interrogate the
orientation or shape changes of bore.
•
Orientation of outer shape geometry vs. RCS
The orientation of non-circular outer geometry of barrels against RCS pattern is
investigated. The square barrel is modeled. Two orientations i.e. 0 degree and 45 degree
with respect to the radiation direction of incident wave are compared. The results are shown
in Figure 3.10.
48
3.5
square; 0 degree
square; 45 degree
3
2.5
RCS
2
1.5
1
0.5
0
0.5
1
1.5
Frequency (Hz)
2
2.5
3
10
x 10
Figure 3.10. RCS response to change of orientation of external shape gun barrel.
(a)
(b)
Figure 3.11. Radiation patterns of scattered wave for (a) 0-degree rotated barrel; (b) 45-degree rotated
barrel.
49
Unlike comparison illustrated in Figure 3.8, in Figure 3.10, significant variations in RCS
pattern can be found with the orientation of non-circular barrel changed. The reasoning is
readily obtained by looking into the radiation patterns of scattered waves from the object.
The wave pattern is entirely changed when the reflective surface under wave illumination
varies. Interestingly, the sharp section of reflective surface makes the resonance frequency
of the entire object decrease when it is directly illuminated.
•
Shape changes in inner and outer geometry vs. RCS
The RCS response to the shape of bore and outer geometry of barrel is compared. Three
types of barrels are investigated: round barrel with round bore, round barrel with square
bore, and square barrel with round bore. The comparison results are presented in Figure
3.12. From this Figure, it is noticeable that RCS pattern is insensitive to the internal shape
variation of barrel. The RCS patterns for round bore and square bore nearly overlapped
each other, though a little deviation can be found from 25GHz to 30GHz. In contrast, the
pattern for square barrel is prominently distinctive from those for the other two scenarios.
3.5
round; round bore
square; round bore
round; square bore
3
2.5
RCS
2
1.5
1
0.5
0
0.5
1
1.5
Frequency (Hz)
2
2.5
Figure 3.12. RCS response to barrel shapes.
50
3
10
x 10
3.2.2 Electromagnetic resonances
Section 3.2.1 has shown that RCS profile can characterize objects. In this section gun
barrels with different cross sections will be illuminated with high frequency EMW
(200MHz to 20GHz), corresponding to resonance region of most common weapons
according to the literature review (section 2.1.4). The same 2- dimension model of 3.2.1 is
set to sweep through 200MHz-20GHz and record resonance respond. Figure 3.13 shows the
geometry the gun barrel (actually the cross section is employed for simulation) and cross
section profile for the simulations.
Figure 3.13. Cross section of gun barrels employed in the FE simulation.
51
In Figure 3.14 three smoothbore barrels with different diameter and a square barrel has
been illuminated with EMW swept from 200MHz to 20GHz. The results imply the samples
can be differentiated in terms of the fundamental resonance (around 500 MHz) and the
magnitude of RCS
Figure 3.14. The RCS values against frequencies for each gun barrel.
3.3 Conclusions
First study using FE simulations have shown that the presence of metallic objects in an
interrogation space causes local distortion of amplitude of the earth magnetic field. This
distortion field response is sensitive to geometry (external and internal shape) of the
interrogated object and also to the orientation.
Measuring earth magnetic field distortion requires sensor technology with high
resolution (order of microtesla) which implies high cost. Another drawback is the
sensitivity to the object orientation which makes the detection more difficult.
Second study with FE simulations have shown the RCS of weapons under EMW wave
illumination can characterize weapons. The sensitivity of RCS for weapons is larger in
the resonance region (200MHz-2GHz) [17]. Thus RCS exhibits resonance when objects
are illuminated by EM wave with frequencies in resonance region. Study has shown the
sensitivity of RCS to outer geometry but insensitive to inner geometry, orientation and
material composition.
52
Chapter 4: NEW PROPOSAL FOR THE DETECTION
OF CONCEALED WEAPONS: EM WEAPON
DETECTION FOR OPEN AREAS
4.1 Theory for Electromagnetic Weapons Detection and
Identification
The detection and identification of the proposed method is based on the scattered magnetic
field from a weapon when it is illuminated with low frequency electromagnetic fields. The
computation of the scattered magnetic field is based on suitable solutions of Maxwell’s
equations. The subject has been analysed in detail by Kaufman and Keller in 1985 [44].
Rather than repeat details of the developments of the authors in [44], selected results
pertinent with identification of CWD are described here.
To simplify the computation of the scattered magnetic field, a weapon is approximated by a
sphere with radius “a” and conductivity “ σ ” and is irradiated with an electromagnetic
wave in the form of step pulses. When the pulse occurs, the weapon is illuminated by a step
change in the amplitude of the magnetic field vector which is considered to be planar in the
vicinity of the weapon. The magnetic field in the vicinity of the weapon is defined as “H 0 ”.
Figure 4.1 shows a sphere of radius “a” irradiated with an electromagnetic wave in the form
of step pulses from group of energized wires.
53
Dimensions in metres
Figure 4.1. Sphere with radius “a” and conductivity “ σ ” irradiated with an electromagnetic wave in
the form of step pulses.
54
Using a quasi-static solution of Maxwell’s equations, it is found that the primary magnetic
field around the sphere is increased by an amount due to currents induced flowing within
the sphere and given by the following three equations:
Eφa =
BRa =
Bθa =
3KB0 a 3 sin θ
R2
6 KB0 a 3 cos θ
R3
3KB0 a 3 sin θ
R3
qs e − qs t
∑
2 2
s =1 [ k s a + ( K − 1)( K − 2)]
∞
e − qs t
∑
2 2
s =1 [ k s a + ( K − 1)( K − 2)]
(4.1)
∞
e − qs t
∑
2 2
s =1 [ k s a + ( K − 1)( K − 2)]
(4.2)
∞
(4.3)
where Eφa , BRa , Bθa are the only components of the anomalous fields caused by currents
induced in the sphere. These components are expressed in a spherical coordinate system
centred on the sphere, with “R” being in the direction of the incident field as shown in
Figure 4.2.
Figure 4.2. Components of the magnetic fields caused by currents induced in the sphere.
55
The other quantities in the equations are defined as:
K = μ / μ 0 (The relative magnetic permeability of the weapon)
q s = k s2 / σμ
(4.4)
(4.5)
k s = πs / α (s=1, 2, 3, ∞ )
(4.6)
and “t” is time, following the initiation of the current step.
The form of each electromagnetic field component is that of a sum of exponentially
decaying transients. The asymptotic behaviour of these transients is now examined. During
the early part of the transient response of the electromagnetic field (tÆ0), the expressions
for the anomalous magnetic field components are as follow:
BRa = B0 (a / R )3 cos θ [1 − 6 / π (α t )1/ 2 ]
(4.7)
Bθa = B01/ 2(a / R)3 sin θ [1 − 6 / π (α t )1/ 2 ]
(4.8)
During the late part of the transient decay, the field is almost entirely determined by the
first exponential terms:
−
6B ⎛ a ⎞
B = 20 ⎜ ⎟ cos θe τ 0
π ⎝R⎠
t
3
a
R
(4.9)
−
3B ⎛ a ⎞
Bθ = 20 ⎜ ⎟ sin θe τ 0
π ⎝R⎠
t
3
a
(4.10)
−
3B a ⎛ a ⎞
Eφ = 0 2 ⎜ ⎟ sin θe τ 0 , where τ 0 = σμα 2π 2 a is a time constant.
τ 0π ⎝ R ⎠
3
t
a
(4.11)
The computation of step-response decay curves of a sphere with parameters: radius a=0.2m,
K=150, σ =3.54e6 S/m is shown in Figure 4.3. Four terms of equation (4.1) and the
summation of first 8100 terms have been computed. The full step response decay curve
shows three zones: early-time behaviour (flat), intermediate-time behaviour (power-law),
and the late-time behaviour (exponential decay).
56
Early- t im e
I nt erm ediat e- t im e
Lat e- t im e
Figure 4.3. Four terms of Eq. (4.2) for s=1, 5, 10 and 50 (top). The full step-response decay computed
for summations of 8100 terms of Eq. (4.2) [45] (bottom).
These results are the basis for the design of a highly effective weapons detection system.
Each of the three expressions for field components is of the form of a product of two terms,
57
the first term of which involves only the geometry and magnitude of the primary field in the
vicinity of the weapon, and a second term involving a time constant, but independent of the
geometry and strength of the field incident on the weapon. Measurement of the time
constant provides a means for weapon detection which is free of false alarms due to
variations in primary field strength.
A time constant can be determined from a plot of the transient magnetic field with
amplitude in logarithm scale (the slope of the curve is the time constant as shown in Figure
4.4.
Sensor
to logarithmic scale
τ1
τ3 τ 2
Figure 4.4. Time constant from sensor signals in logarithm scale.
58
The time constant can also be determined from the expression:
τ0 = −
Bθa
BRa
=
−
∂BRa / ∂t
∂Bθa / ∂t
(4.12)
Weapons have a complex geometry, so a question arises as to whether or not conclusions
for simple geometries extend to more complicated shapes. Kaufman and Keller (1985)
extended the analysis to axial symmetry and arrive at an expression for the scattering time
constant as follows:
τ 0 = σμa 2 / q ,
where “q” depends on the shape of the conductive body.
(4.13)
The expression shows the potential of using the time constant in identifying a specific
weapon of an even more complicated shape. All weapons of the same size, shape and
metallic composition will be characterised by a scattered electromagnetic field. Time
constant will be the same, no matter what the strength of the incident field or the distance of
the weapon from the transmitter or the receiver array. Thus weapons can be classified at
least within the precision with which the time constant of the decaying magnetic field can
be determined.
The appropriate way to determine the time constant for a real weapon is by illuminating
that weapon with an electromagnetic field using step pulses and measuring the time
constant from the decaying field. Usually the weapon will be carried on the body of a
person. However the time constant associated with the human body might obscure that of
the weapon. An estimate as to whether this will happen can be made by substituting
numbers in the expression (4.13). For example, assume that the conductivity-permeability
of the steel in a weapon is of the order of 1 SHm 2 , the radius of the sphere enclosing a
weapon is 0.1m, the form factor “q” is 10 (the more convoluted the shape of the metallic
object the greater will be the form factor). With these numbers, the order of magnitude
estimate for the time constant of the hypothetical weapon is 1 millisecond. In the human
body, flesh and bone are conductive and the conductivity-permeability product for a human
59
body is of the order of 10SHm 2 . Thus the corresponding order of magnitude time constant
for a human body is 0.1 μs . Although these are only estimates, the two time constants are
so different that there is no likelihood that one will obscure the other.
4.2 Proposed Concealed Weapon Description
The system comprises a Uniform Magnetic Field Generator (UMFG), consisting of a group
of parallel current carrying cables located below a walking surface to produce a horizontal
magnetic field and arrays of magnetic field sensors located within the sensing area as can
be seen in Figure 4.5. The system works based on pulsed-EMI technique. Currents in the
UMFG flow for a sufficiently long time to allow turn-on transients in the sensing object to
dissipate. The current in the cables are then turned off. According to Faraday’s law, the
collapsing magnetic field induces an electromotive force in the metal target. This force
causes eddy currents to flow in the metal. Because there is no energy to sustain the eddy
currents, they begin to decrease with a characteristic decay time that depends on the size,
shape, and electrical and magnetic properties of the metal object. The decay currents
generate a secondary magnetic field, and the time rate-of-change of the field is detected by
magnetic field sensors. The time decay response is used for classification. The arrays of
sensors are distributed within sensing area in a grid pattern in order to show the location of
gun/knife carrying people.
60
Figure 4.5. Functionality of the proposed CWD system.
4.3 Proposed Design
The new proposition of CWD system comprises:
1. Metal detector subsystem.
•
•
UMFG, Switches and Pulse control
Sensor arrays
2. Control subsystem.
3. Detection & classification subsystem.
4. Operator interface.
61
Figure 4.6. Proposed CWD structure.
4.3.1 Metal detector subsystem
•
Uniform Magnetic Field Generator (UMFG)
It is known from basic physics textbooks that for an infinite conducting sheet current, the
field in the direction parallel to the sheet is given by:
B = μ0v / 2
(4.14)
Where v is the current density in the sheet and μ 0 is the permeability of free space (
0
=
4π×10−7 N·A−2).
The magnetic field perpendicular to the sheet is zero. It implies that the sheet current is a
horizontal magnetic field generator. Sheet current can be approximated by closely spaced
parallel current carrying wires. The important feature of (4.14) is the fact that the magnetic
field is constant, a feature that improve weapon detection as it will be shown in the next
section.
An approximation to a sheet current is created by a series of closely spaced parallels wires
(see 4.3.1). These wires form the active surface of UMFG. The UMFG consists of a single
cable AWG10 (American Wire Gauge, 2.58 mm diameter) wiring around a box forming
rectangular 20 loops (3 m x 0.60 m), 2cm between loops in the way showed in Figure 4.7.
62
Electronics
Btop
Bbottom
I
I
Dimensions in
Figure 4.7. Magnetic field covering area from active part of UMFG (top plane) and return path
(bottom plane).
In Figure 4.7 the cable is connected to high-speed electronic (turn-off delay time 180ns)
switches. A single power supply provides current to the UMFG. The excitation field is
generated by the plane of wires closest to the ground (top of the box). However the primary
field is reduced by the superposition of the opposite magnetic field generated by the return
path (bottom of the box). The effective magnetic field covering area is shown in Figure 4.8.
63
Figure 4.8. Effective magnetic field covering area.
Magnetic field covering area will not be reduced if magnetic shielding is added to the
return path. An alternative solution without using magnetic shielding is to route return path
to the sides of the box instead of the bottom as shown in Figure 4.9.
64
Figure 4.9. Routing the return path to the sides of the box maximizes the effective magnetic field
covering area.
The current in the UMFG is controlled by the electronic switches. The switches are power
MOSFETs (IRF150). A pulse generator controls the opening/closing of the switches. The
circuit is shown in Figure 4.10. The circuit consists of low power pulse generator (LM555
set a 20Hz, 10% duty cycle in order to not overheat the wires.) switching a power MOSFET
(IRF150) in order to drive 6.7A into the UMFG. To prevent electrical failures the low
power circuit (pulse generator) and MOSFET are decoupled by an opto-isolator.
To protect the MOSFET from back electromotive force (EMF) a flywheel diode series with
three Zener diodes are connected across the ends of UMFG wires. Besides protecting the
MOSFET the set of diodes reduce considerably the switch off time whose effect on target is
to enhance the signal back to the sensors.
65
66
Figure 4.10. Pulse Control Circuit.
The proposed weapon detection system needs to cover a larger area. To accomplish that an
option could be to use a single UMFG with a longer wire; however doing this maximizes
the magnetic field collapse time, inducing weaker eddy currents in the metal and making
detection difficult. In the time domain, the strength of the excitation magnetic field is
dependent of rate of change of its collapse time. The decay of the excitation magnetic field
is governed by the decay of the current in the wires after the switch is opened. Thus a way
to use a longer wire in which currents decays in a short time needs to be found.
Ignoring capacitive effects, the decay time on the UMFG is given by L/R, where L is the
inductance of the loop and R is the resistance of the loop. For this application, L can be
minimized using several shorter loops. L decreases more than R so the reduction of R does
not prevent the minimization of the time decay.
For the proposed weapon detection it is planned to use five of UMFG (144mx5) connected
in parallel as shown in Figure 4.11 with sensor arrays in between to cover an area of 3x3 m.
This block area could be used as a module for a larger system. In the whole system each
UMFG has its own switch because each UMFG will be excited sequentially as shown in
Figure 4.12 by a pulse controller. It improves the detection by giving location of the target
and avoiding exciting other metal in the vicinity.
67
Power Source
PG1 (Switch 1)
Sensor Array
PG2 (Switch 2)
Sensor Array
PG3 (Switch 3)
Sensor Array
PG4 (Switch 4)
Sensor Array
PG5 (Switch 5)
Pulse Control
Figure 4.11. Layout of a block of intended CWD system including five UMFG.
68
Figure 4.12. Time diagram of switch trigger signal (LM555 set a 20Hz, 10% duty cycle).
•
Advantages of uniform excitation magnetic field
There are some advantages in terms of detection and classification according to the way
that targets are excited. Targets could be illuminated by magnetic field generated by a large
coil (3m diameter) or group of parallel wires. Figure 4.13 shows streamlines of magnetic
field density from a loop (a) and those from closely spaced current carrying wires (b)
69
Figure 4.13. Cross sectional view of magnetic field from a coil loop (top); Cross sectional view of
magnetic field from UMF (bottom). Current flow into the paper.
70
Illumination of objects with uniform magnetic field makes easier the classification process
and provides a strong source of illumination.
[1] Improvement in target classification
The advantages of exciting an unknown metal with uniform magnetic field can be better
explained by a description of the basic physics of the problem. Consider a simplified
extended target model made of two separated dipoles as shown in Figure 4.14. Each dipole
decay time constant is modelled as a single exponential decay and the sum of the two
dipole time decays is the target’s decay signature. Dipole 1 has a decay time constant τ a
and amplitude A a and dipole 2 has a decay time constant τ b and amplitude A b . This
signature is an inherent property of the target and is used for classification. Now let’s
consider the case as showed in Figure 4.14a, where exciting the target with a spatially
varying (non-uniform) magnetic field that excites the two dipoles with different magnetic
fields, B 1 and B 2 , the target signature is dependent on the excitation field and is not simply
the sum of two exponential decays. This potentially makes target classification more
difficult. In Figure 4.14b, the excitation magnetic field is uniform (B u ) over the target. In
this case the theoretical response of the target is preserved because the magnetic field can
be factored out of the target response equation. So target signal expression still contains
decay properties, independent of the strength of the uniform magnetic field.
Target signal ≈
Target signal ≈
B1 Aa exp(−t / τ a ) + B2 Ab exp(−t / τ b ) Bu [ Aa exp(−t / τ a ) + Ab exp(−t / τ b )]
B
B
Target m odel
Dip 1
Target m odel
Dip 1
Dip 2
Dip 2
Bu
B2
B1
X1
X2
Posit ion
Posit ion
B: Uniform Magnet ic Field
A: Non- Uniform Magnet ic Field
Figure 4.14. Magnetic field uniformity and Target signal response [12].
71
[2] Slow reduction of field strength with distance
The uniform magnetic field generator has a stronger magnetic field intensity compared to
that of conventional loop coils as a function of distance from the plane of the magnetic field
generators [11]. Figure 4.16 shows a plot that compares Bx from the Uniform magnetic
field generator UMFG (1m x 3m) to Bz of a loop coil (1m of diameter) versus distances
from the plane of UMFG. The magnetic field for each case has been normalized to one at
the distance of 10cm to show the relative field intensity fall-off with distance.
Bx
Bz
Bx
Figure 4.15. UMFG and Loop Coil.
72
1.4
Bx of UMFG
Bz of Coil
1.2
normalized B @ 10cm
1
0.8
0.6
0.4
0.2
0
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
vertical height from centre of plane (m)
0.9
1
Figure 4.16. Magnetic field from UMFG and a loop coil.
•
Sensor array
The magnetic field strength for the proposed weapon detection system is in the order of
microtesla (10 −2 gauss). Looking on table 4.1 there are plenty of magnetic field sensor
technologies which can cope with this requirement [46]. Some sensor technologies have
more advantages than others in term of minimum detectable field, cost and circuit
complexity. For this research three magnetic sensor technologies could be suitable: search
coil, hall-effect devices and magnetoresistive devices.
73
Table 4.1. Magnetic Sensor Technology Field Ranges [46].
Magnetic Sensor Technology
Detectable Field (Tesla)
10 −12
10 −8
10 −4
10 0
10 4
1. SQUID 1
2. Fiber-Optic
3. Optically pumped
4. Nuclear-procession
5. Search coil
6. Anisotropic Magnetoresistive
7. Flux-Gate
8. Magnetotransistor
9. Magnetodiode
10. Magneto-Optical sensor
11. Gian magneto-impedance
12. Hall-effect Sensor
1
Superconductive Quantum Interference Device
Table 4.2 shows an analysis of capabilities of those sensor technologies for the present
work.
Table 4.2. Performance of Sensor Technologies.
Characteristics
Resolution (T)
Search coil
10 −12
Hall device Magnetoresistive (AMR)
10 −10
10 −4
Cost1
Size
Temperature Stability
Input Power
Low
~cm
Good
0
Moderate2
~mm
Poor
~mW
High
~mm
Acceptable
~mW
1
Cost depends on the scale of manufacture; the relative cost here is estimated for large-scale manufacturing.
2
Relative cost include bias magnet.
74
In the proposed weapon detection the array of sensors are distributed within sensing area in
a grid pattern in order to show gun/knife carrying people location as shown in Figure 4.17.
3m
Sensor array (1,1)
Sensor array (1,2)
Sensor array (1,3)
Sensor array (2,1)
Sensor array (2,2)
Sensor array (2,3)
Sensor array (3,1)
Sensor array (3,2)
Sensor array (3,3)
Sensor array (4,1)
Sensor array (4,2)
Sensor array (4,3)
Sensor array (5,1)
Sensor array (5,2)
Sensor array (5,3)
60 cm
3m
Figure 4.17. Sensor array distributions for the proposed CWD system.
4.3.2 Detection & classification subsystem
The eddy currents excited in the metal body take the form of an exponentially decaying
transient immediately following sudden changes in the exciting magnetic field. This decay
curve can be used to obtain a time constant of the current decay. It has been found (G.V.
Keller, 1985) that time constant decay is independent of the strength of the field incident
magnetic field and highly dependent on the size, shape and material composition. For
gun/knife classification a library of potential threat metal objects can be developed. When a
metal body is encountered in the field, its time decay response can be compared to those in
the library, and if a match is found, the metal body can be classified. If an unknown target’s
75
response signature is significantly different from any weapon signature in the library, the
target can be considered a clutter.
In the proposed system each UMFG and sensor array needs to be excited and read
independently following a given order. This facilitates finding the location of the suspect
and saving memory and computation time for classification. This process is shown in
Figure 4.18.
Weapon Library
Signature
Sensor array (1,1)
Sensor array (1,2)
Sensor array (1,3)
Sensor array (2,1)
Sensor array (2,2)
Sensor array (2,3)
Sensor array (3,1)
Sensor array (3,2)
Sensor array (3,3)
Sensor array (4,1)
Sensor array (4,2)
Sensor array (4,3)
Sensor array (5,1)
Sensor array (5,2)
Sensor array (5,3)
Weapon/Clutter
Pulse Controller
Pulse generator
1 (switch 1)
Pulse generator
21 (switch 2)
Pulse generator
3 (switch 3)
Pulse generator
4 (switch 4)
Pulse generator
5 (switch 5)
Figure 4.18. Scanning of interrogation area for detection and classification.
•
Extraction of time constant decay
One of the proposed algorithms for the determination of a time constant consists of
transformation of the signal strength by using the logarithm of the signal. With this
transformation, an exponential decay curve appears as straight line, with the slope being the
76
time constant. The time constant (slope of the line) can be extracted by curve fitting (bestfit one –parameter linear function) of the transformed data. This process is illustrated in
Figure 4.19.
Figure 4.19. Proposed algorithm for extraction of the Time Constant.
77
4.3.3 Control and operator interface subsystem
The control subsystem gives the sequence order for the UMFG excitation, reads the sensors,
and performs the signal processing for the detection and classification.
Set of alarms, location indicator, video monitors and displays interfaces with the operator
giving the result of the detection and classification.
4.4 FEA Simulations
To test the proposed method, a simplified 3-D model in COMSOL 3.3 (finite element
analysis and solver software package) as shown in Figure 4.21, has been used. The model
consists of a UMFG (20 parallel cables laid on the floor, 3cm in between) for transmitting
step pulse of current of 10A, 1.5ms width pulse, falling time 170us, causing eddy currents
to flow in a nearby metal object. The eddy currents scatter a signal that will be detected by
a three-axis magnetic sensor array (hall or magneto-resistive sensors) placed between the
cables. Parameter values used in the model are based on previous studies and simulations
performed to achieve a strong enough magnetic field. Simulations have been run to analyse
the sensitivity/insensitivity of the target response to changes in shapes, orientation, sizes
and material composition of interrogated objects. A study of target response at different
stand off is also included.
4.4.1 Sensitivity to outer shape
Four different samples: Gun1 with smooth barrel and smooth bore, Gun2 with square barrel
and smooth bore, and two knives with different cross section as showed in Figure 4.20 are
tested respectively. In Figure 4.22, the magnetic flux density (B) along z-axis and y-axis is
sensed from a distance of 35cm from the object along perpendicular line between object
and floor. Following the intended algorithm for time constant extraction, the logarithm of
magnetic flux density is plotted against time in order to show more clearly the time
constant for each interrogated object.
78
Figure 4.20. Samples for test.
target
Pick up signal point
Figure 4.21. 3-D Model (dimensions in metres).
79
Figure 4.22. Scattered magnetic field on y-axis (top) and z-axis (bottom).
80
4.4.2 Insensitivity to object orientation
In the following test a gun (steel) is oriented in three positions as seen in Figure 4.23.
Results showed that even if the amplitude of the signal changes, the time of decay stays
invariant. Also measurements have been taken at different distance from the gun (15cm and
35cm as shown in Figure 4.24) to show how the time constant is independent of the
strength of the field incident.
Figure 4.23. Model to test sensitivity to weapon orientation (domain dimension in metres).
81
Figure 4.24. Scattered magnetic field sensed from a distance 15 cm (top) and 35 cm below the gun
(bottom).
82
4.4.3 Sensitivity to size
Three guns of the same type but different sizes are tested. In this test a coil is added to the
model to sense the time decay of induced current in the target caused by the abrupt change
of the magnetic field. Results in Figure 4.25 show that time constant is a function size of
the sample. Because large metallic objects store more energy than small ones, the time
decay of collapsing magnetic field (or induced current in the coil) takes longer. This
characteristic could be potentially used to discriminate object size.
Figure 4.25. Model to test size sensitivity.
83
100%
200%
400%
-1
Induced current [A]
10
-2
10
-3
10
1
1.05
1.1
1.15
time (s)
1.2
-3
x 10
Figure 4.26. Time constant response to size of guns.
4.4.4 Sensitivity to material composition
The same sample as Figure 4.25 with different material composition (copper, iron, and
steel) was tested. The time constant profile is clearly different for each case confirming the
potential of the new weapon detection system to identify objects with a different material
composition.
84
Figure 4.27. Time constant profile of a gun made of steel, copper and iron.
4.4.5 Target response to stand off
The measurements are taken in a point located over the cables. The target (aluminium) is
placed at different distances (15mm, 25mm, 35mm, 45mm and 55mm) from the sensor as
shown in Figure 4.28. The signal measured by the sensor without the metal object has been
subtracted from all measurements.
85
Figure 4.28. Model to measure target response at different stand off.
Figure 4.29. Signal amplitude of the target at different stand off.
86
Figure 4.29 shows how peak amplitude of induced current in a search coil is proportional to
target location.
Figure 4.30. Gradient of target signals response at different stand off.
From Figure 4.29 and 4.30 it can be seen that even though amplitude of the signal response
changes at different stand off, the time constant remains invariant. This simulation proves
that time constant is independent of the strength of illuminating and scattering magnetic
field.
Tests have shown that time decay measurements obtained from simulations of the proposed
weapon detection system are sensitive to the shape, size, and material composition of the
target. Furthermore, the time decay profile is independent of the strength of the illuminating
magnetic field. Measurements taken in any direction also show an agreement in time decay
profile. All the mentioned are requirements of a comprehensive CWD system, so the results
look promising in terms of effectiveness and reduction of false alarm level.
87
4.5 Experimental Work
4.5.1 Driver circuit for UMFG
To improve the detection the UMFG needs to induce eddy currents in the target with
sufficient strength to generate magnetic fields strong enough to be measured by sensors.
UMFG has to excite the target with a magnetic field with strength and rate of change high
enough to sense distance targets and induce strong eddy currents. Thus UMFG needs to
carry enough current and switch faster to meet those requirements.
Based on FE simulations the UMFG needs to carry a current 6~10A and switch in around
250us to induce strong enough eddy currents in the target such that the magnetic field back
to the sensors is in the order of μT . A driver circuit for the UMFG consists of a MOSFET
IRF150 that can carry the required current and a step pulse generator to switch the
MOSFET at a frequency adjusted to the acquisition and signal processing time of available
hardware (in this research the frequency used is 20Hz).
Figure 4.31 shows the first design for the driver of the UMFG. The pulse generator (low
power circuit) and the MOSFET (high power circuit) connected to the UMFG are
electrically decoupled by an opto-isolator to prevent electrical failures. The pulse generator
consists of timer LM555 set at 20Hz, 10% duty cycle in order to not overheat the wires.
88
89
Figure 4.31. Driver circuit for UMFG (design 1).
To prevent back electromotive force (EMF) in the MOSFET a flywheel diode is connected
across the wire. The pulse signal obtained from the 10 turn cable has a fall time of 1.2ms
and amplitude 6.7A. As explained before the MOSFET needs to be switched off in less
than 250ms to induce high eddy current density in the target. The clamped voltage and the
current are shown in Figure 4.32.
16
14
Voltage (V)
Current (A)
12
Back EMF
10
8
6
4
2
0
0.075
0.076
0.077
0.078
Time (s)
0.079
0.08
Figure 4.32. Current and Voltage in UMFG during one excitation pulse (design1).
To speed up the fall time a modification in the circuit was done (Figure 4.33). Three Zener
diodes were added to the fly-back diode in order to reduce energy stored in the inductance
L and allows it discharge faster. The fall time is now 200us with amplitude 6.7A. The
reverse transient voltage is raised but not enough to damage the MOSFET (IRF150 has a
VDss=100V). Voltage and Current in the new design are shown in Figure 4.34.
90
91
Figure 4.33. Pulse current generator (design 2).
35
Voltage (V)
Current (A)
30
25
Back EMF
20
15
10
5
0
0.0755
0.076
0.0765
0.077
Time (s)
0.0775
0.078
0.0785
Figure 4.34. Current and Voltage in UMFG during one excitation pulse (design2).
Changes in transient voltage and current of UMFG as a consequence of adding three Zener
diodes series with original fly-back diode are shown in Figure 4.35 and 4.36. Figure 4.35
shows the increasing of the reverse transient voltage as price to reduce switch off time.
Figure 4.36 shows the substantial reduction of MOSFET turn off time after design changes.
92
35
30
Voltage (V)
25
20
15
10
5
Flyback diode
Flyback+Zener diode
0
0.076 0.0765 0.077 0.0775 0.078 0.0785 0.079 0.0795 0.08
Time (s)
Figure 4.35. Back EMF reduction comparison between design 1 and 2.
6
Flyback diode
Flyback+Zener diode
Current (A)
5
4
3
2
1
0
0.076 0.0765 0.077 0.0775 0.078 0.0785 0.079 0.0795 0.08
Time (s)
Figure 4.36. Current fall time comparison between design 1 and 2.
93
4.5.2 Testing of the proposed weapon detection.
Experimental tests were developed on an UMFG (3 m x 0.60 m) placing metallic objects at
30 cm over the centre of the UMFG plane and reading the measurement of sensors placed
over the plane and below the targets. The tests were performed in an environment under
EM interference (electronic equipments) and noise (nearby metallic objects in the
illumination area) to resemble the worst working scenario of the proposed CWD system.
•
Metallic target response in search coils.
The following tests were performed using a search coil device as showed in Figure 4.37.
This search coil device allows choosing three loops arrangements. Previews test with each
individual loop showed different sensitivity. The best response was obtained with 11.79 Ω ,
2.525mH loop.
Figure 4.37. Search coil device.
94
Table 4.3. Coil parameters.
Diameter (mm)
Large coil
150
Medium coil
120
Small coil
70
Inductance (mH)
4.414
2.525
1.571
Resistance (Ohms)
17.833
11.796
48.158
First test: sensitivity of time constant to target orientation
In the first test the sensitivity of time constant at target orientation is analyzed. The metallic
base of a solder kit is placed 30cm over the UMFG and rotated in three positions. Search
coil responses are plotted in Figures 4.38 and 4.39. The signal measured from the search
coil in absence of target (red) is also included for reference.
2.5
No sample
Solder kit pos 1
Solder kit pos 2
Solder kit pos 3
Voltage (V)
2
1.5
1
0.5
3.5
4
4.5
5
5.5
6
6.5
Time (s)
7
7.5
8
Figure 4.38. Solder kit responses at different orientation.
95
8.5
-4
x 10
No sample
Solder kit pos 1
Solder kit pos 2
Solder kit pos 3
0
Voltage (V)
10
-1
10
5.2
5.4
5.6
5.8
6
Time (s)
6.2
6.4
6.6
6.8
-4
x 10
Figure 4.39. Zoomed views of solder kit responses in logarithm scale.
.
From Figure 4.38 it can be seen that target response for each position (blue curves) has
similar transient profile. A zoomed view of curves (Figure 4.39), plotted in logarithmic
scale, shows a first part of the transient (500-560 μs ) dominated by target signals with
linear profile. The second part of the transient (after 560 μs ) shows coupled EM
interference which is dominant because the target has already released most of the stored
energy. Gradient of the curves in first part of the transient correspond to time constant
which is similar for the threes positions. It probes the insensitivity of time constant to target
orientation.
Second test: sensitivity of time constant to target shape
In this test, the response of samples with differentiated geometry characteristics (a base of
solder kit, a mobile, and three pliers) is analysed. Measurements were taken at 30cm over
the UMFG. Search coil responses are plotted in Figure 4.40(top). The signal measured from
the search coil in absence of target (red) is also included for reference.
96
No sample
Solder kit
Plier 1
Mobile
Plier 2
Plier 3
2.5
Voltage (V)
2
1.5
1
0.5
0
3.5
4
4.5
5
5.5
Time (s)
6
6.5
7
-4
x 10
No sample
Solder kit
Plier 1
Mobile
Plier 2
Plier 3
0
10
Volt age (V)
7.5
-1
10
5
5.2
5.4
5.6
Time (s)
5.8
6
6.2
-4
x 10
Figure 4.40. Target response signals from search coil (top). Zoomed view in logarithm scale (bottom).
97
Figure 4.40 (top) shows that each sample response has different transient profile. Like first
test the zoomed view of curves (Figure 4.40(bottom)), plotted in logarithmic scale, shows a
first part of the transient (500-550 μs ) dominated by target signals with linear profile and a
dominant EM interference in the rest of the transient. Now gradients of the curves in first
part of the transient are remarkably different, which allow classifying samples. Thus this
test probes the sensitivity of time constant to target shapes.
•
Metallic target response in magnetoresistive sensors
Anisotropic Magnetoresistive (AMR) sensors convert magnetic fields to a differential
output voltage, capable of sensing magnetic fields as low as 30 μ gauss. Most low field
magnetic sensors are affected by large magnetic disturbing fields (>4~20 gauss) that may
lead to output signals degradation. In order to reduce this effect, and maximize the signal
output, a magnetic switching technique can be applied to the AMR bridge that eliminates
the effect of past magnetic history [47].
The circuit in Figure 4.41 is the implementation of the switching technique to maximize
sensitivity of AMR before UMFG illuminates the target. The circuit operates sending
master pulsed signal of 14.5Hz, 9.88% duty cycle (first LM555 working as astable) to drive
a pulsed current of 4A through P and N Darlington into the AMR coils. This pulsed current
realigns magnetic domains of the permalloy film on each set and reset increasing the
sensitivity. The master pulsed signal also pass through diodes, resistor, inverters and trigger
the second LM555 (working as mono stable, 10ms output time) to produce delay before
switch off the UMFG and excite the target. This circuit eliminate all past magnetic history
of the sensor just before start the measurements. Figure 4.42 shows the timing diagram of
signals in different part of the circuit.
98
Master pulse
signal
1
99
Switching (S/R)
circuit
2
HMC1001
Figure 4.41. AMR sensor.
3
1. Master pulse driving set/reset AMR
Voltage [V]
10
5
0
0.05
0.1
0.15
0.2
0.25
2. Set/Reset AMR
Voltage [V]
10
0
-10
0.05
0.1
0.15
Time [s]
3. Delayed pulse driving IRF150
0.2
0.25
0.1
0.15
0.2
0.25
Voltage [V]
10
5
0
-5
0.05
Figure 4.42. Timing diagrams.
Test: sensitivity of time constant to target shape
The sensitivity of AMR sensor to samples with differentiated shapes (a base of solder kit, a
mobile, three pliers, and keys) is analysed. Measurements were taken at 30cm over the
UMFG. Target responses are plotted in Figures 4.43 and 4.44. The signal measured from
the AMR in absence of target (blue) is also included for reference.
The different transient profile of each sample can be observed more clearly from zoomed
plot (Figure 4.44). Time constants can be extracted from first part of transient (80-100 μs )
where target signals are dominant.
100
No sample
Plier 1
Plier 2
Solder kit
Mobile
Keys
Small plier
3.5
Voltage (V)
3
2.5
2
1.5
1
6
7
8
9
10
Time (s)
11
12
13
14
-5
x 10
Figure 4.43. Time constant decays profile of metallic objects using AMR HMC1001.
1.4
No sample
Plier 1
Plier 2
Solder kit
Mobile
Keys
Small plier
1.3
Voltage (V)
1.2
1.1
1
0.9
0.8
0.7
1
1.05
1.1
1.15
1.2
1.25
Time (s)
1.3
1.35
1.4
-4
x 10
Figure 4.44. Close up of Time decay profile of metallic objects using AMR HMC1001.
101
4.6 Conclusions
A new proposal CWD has been proposed based on information contained in the scattered
magnetic field from objects illuminated by a low frequency magnetic field. Information
regarding physical characteristics, size and material composition of metallic object can be
extracted from one of the parameters of the scatter signal: the time constant of the decaying
magnetic field induced in the object after step excitation.
The system design discussed and proposed in the second and third part of this chapter can
serve as foundation for further development of the proposed CWD.
FEA simulations have shown that time decay are sensitive to the shape, size, and material
composition of the target. Furthermore, the time decay profile is independent of the
strength of the illuminating magnetic field. Measurements taken in any direction also show
an agreement in time decay profile. All the mentioned are requirements of a comprehensive
CWD system, so the results look promising in terms of effectiveness and reduction of false
alarm level.
Two sensor technologies were chosen to test the feasibility of the proposed CWD system.
Signals picked up from search coil and magnetoresistive sensor (AMR) technologies
showed sensitivity to shapes and insensitivity to orientation of metallic objects.
Measurements were taken in an environment full of electronic equipments where noise and
EM interference were present.
Search coils showed less immunity to those effects
comparable to those in AMR. AMR sensors could provide more space resolution detection
than search coils because of the small size however it can be achieved at cost of complexity
in the circuitry (i.e. switching circuit, signal conditioning, etc). More investigation is need
with sensor technologies to provide high resolution and low cost solution for the proposed
CWD.
102
Chapter 5: CONCLUSIONS
•
The literature survey has identified some advantages and disadvantages of CWD
systems in issues such as operating range, material composition of the weapon,
penetrability and attenuation factors. Electromagnetic (EM) resonance system,
which uses EM resonance as signatures to distinguish weapons and nuisance
objects, shows a large sensing distance. It could detect concealed weapons at
distances up to 10m. However the system shows problems with classification due to
signature of an individual with a weapon is very similar to one without a weapon.
CWD systems based on MMW have high penetrability (clothing mainly) but high
attenuation under weather condition (water). Terahertz sensing and imaging system
are currently being developed that are capable of forming images of concealed
weapons suitable for weapon classification. However, the penetration of THz
through the atmosphere for stand-off detection and through some types of clothing
leads to poor results. Infrared imagers can be used for detecting concealed guns at
nights. However infrared radiation emitted by people is absorbed by clothing. The
best results for gun detection are when the clothing is tight, which is not normally
the case. Ultrasound concealed weapons detector works at distances of up to 8m.
However, this makes no use of the resonant cavities present in a gun to generate a
unique gun signature and it is falsely triggered by non-gun items such as leather
wallets. Inductive magnetic field detectors are the cheapest and common in the
market but have a problem of sensitivity. When dealing with materials that are not
of very high conductivity or of very small dimensions, the material is hardly to be
detected. Passive CWD systems such as those based on sampling of the earth’s
magnetic field are very attractive because of lacking of radiation to interrogate the
inspection area. However those systems need very sensitive magnetometers of high
cost.
•
Literature survey has identified that a comprehensive CWD needs to meet three
main requirements: high sensing distance for open and outdoor environments; high
penetration for detection of weapons, explosive under clothes, or even walls, and
103
robust weapon recognition to discriminate accurately threat/non-threat items. No a
single method in the survey meets all these requirements. Most of the CWD
methods described here are complementary. New CWD systems need to be an
amalgamation of the techniques mentioned above permitting a reduction in the
number of false alarms.
•
Study of low cost complementary technologies has shown the feasibility of
characterisation of guns and knives. Finite element models of Passive (Earth field
distortion) and Active (EM wave illumination) weapon detection systems showed
that earth magnetic field changes are sensitive to outer and inner shapes of metallic
objects, however they are also sensitive to object orientation. RCS profile of
metallic object illuminated by high frequency radiation (300MHz to 1GHz) showed
sensitivity to outer shapes and a convenient lack of sensitivity to object orientations.
This study showed that it could be possible to obtain signatures for each threatening
objects to be used on identification.
•
A new CWD system has been proposed to discriminate threatening metallic objects
based on time constant of scattered field from pulsed magnetic field used to
interrogate an area with metallic objects.
•
The feasibility of the new CWD systems has been analyzed by experimental tests
and FE simulation studies. Experimental tests have confirmed that the proposed
CWD is sensitive to the target’s shape and insensitive to target’s orientation. FEA
has shown that the time decay profile is independent of the strength of the magnetic
flux density, so independent of the distance away from the sample; however there is
a constraint on the distance because of the sensitivity limitation of the magnetic
field sensors. Measurements at 45 cm from the target are in range of 10 −5 T. There
are magnetic sensors on the market able to detect these levels (search-coil
magnetometer; flux-gate magnetometer; SQUID magnetometer; GMR) supporting
the feasibility of the proposed CWD system.
104
•
This thesis has provided the foundation for development of new proposals on CWD
and given an input for a comprehensive multimodal weapon detector system free of
false alarms.
105
Chapter
6:
FUTURE
RECOMMENDATIONS
WORK
AND
The present stage of this work has shown the potential of the time constant as a parameter
for weapon detection and classification. The following works could be included in a
further stage.
•
Design specification for the hardware of the CWD. It should include sampling time and
bit resolution for the acquisition card, and switching frequency for the pulse controller
such that online detection and classification can be achieved.
•
Even though sensing systems using search coils and AMR have shown the relatively
good sensitivity and response, further studies with new low cost sensing technologies
need to be included to get optimum signals and improve the detection. One of the
emergent and promising technologies is the two dimensional electron gas,
AlGaAs/InGaAs/GaAs (Indium Gallium Arsenide/Indium Gallium Arsenide/Gallium
Arsenide) [48] 2DEG Hall device which could be included in a future work.
•
Multi-axis sensor array can be included in the CWD system to have more views of the
targets and improved detection.
•
Implement 3D reconstruction of concealed weapons by means of UMFG modules such
that resembles a walk through portal. The portal could be comprised of one UMFG
lying on the floor and two UMFG attached to walls. An algorithm to reconstruct a 3D
image from sensor array signals, associated with the UMFG modules, needs to be
developed.
•
More investigation needs to be done for multitarget detection and classification.
106
•
The collection of signal response of different classes of guns, knives and other
threatening metallic objects to develop signature database.
•
Test the sensitivity of the CWD to target material composition.
•
The development of robust pattern recognition system with the time constant as a main
feature. Efforts need to put in Signal Processing to achieve time constant extraction and
correlation algorithms with low computation time to ensure online detection and
classification.
•
Finally the present work needs to be complemented with other approaches at several
non-overlapping wavelengths in order to produce an integrated multimodal sensing
system, reducing the level of false alarms.
107
REFERENCES
[1]
SUTTON, V., and BROMLEY, D.A., Understanding technologies of terror,
Technology in Society, vol. 27, Issue 3, pp. 263-285, Aug 2005.
[2]
PAULTER, N.G., Guide to the Technologies of Concealed Weapons and Contraband
Imaging and Detection, National Institute of Justice Guide 602-00, pp. 33-50, Feb
2001
[3]
ROYBAL, L.G., RICE, P.M., and MANHARDT, J.M., A new approach for detecting
and classifying concealed weapons, Proceedings of the SPIE, The International
Society for Optical 54 Engineering, Conference on Surveillance and Assessment
Technologies for Law Enforcement, vol. 2935, pp. 95-107, Boston, MA, , November
1996
[4]
ALLEN, G.I., CZIPOTT, P., MATTHEWS, R., and KOCH, R.H., Initial evaluation
and follow on investigation of the Quantum Magnetics laboratory prototype, room
temperature gradiometer for ordnance location, Proceedings of the SPIE, The
International Society for Optical Engineering, Conference on Information Systems for
Navy Divers and Autonomous Vehicles Operating in Very Shallow Water and Surf
Zone Regions, vol. 3711, pp. 103-112, Orlando, FL, April 1999
[5]
CZIPOTT, P.V., and IWANOWSKI, M.D., Magnetic sensor technology for detecting
mines, UXO, and other concealed security threats, Proceedings of the SPIE, The
International Society for Optical Engineering, Conference on Terrorism and Counterterrorism Methods and Technologies, pp. 67-76, Boston, MA, November 1996
[6]
KOTTER, D.K., ROYBAL, L.G., and POLK, R.E., Detection and classification of
concealed weapons using a magnetometer-based portal, Proceedings of SPIE – Vol.
4708 Sensors, and Command, Control, Communications, and Intelligence (C3I)
Technologies for Homeland Defense and Law Enforcement, pp. 145-155, August
2002
[7]
IDAHO NATIONAL LABORATORY TECHNOLOGIES, Concealed Weapon
Detection using a magnetometer-based portal to detect potential threats [online].
Available at:
https://inlportal.inl.gov/portal/server.pt/gateway/PTARGS_0_200_816_259_0_43/htt
p%3B/inlpublisher%3B7087/publishedcontent/publish/communities/inl_gov/about_in
l/home_page_fact_sheets/sheets/concealed_weapons_detection_4.pdf [Accessed 20
November 2006]
[8]
NELSON, C.V., Metal Detection and Classification Technologies, Johns Hopkins
APL technical Digest, Vol. 24, Number 1, pp. 62-66, 2004
[9]
TIAN, G.Y., ZHAO, Z.X., and BAINES, R.W., The research of inhomogeniety in
eddy current sensors, Sensors and Actuators A, 58, pp. 153-156, 1998.
108
[10] PAULTER, N.G., Users’ Guide for Hand-Held and Walk-Through Metal Detectors,
NIJ Guide 600–00, NCJ 184433, Office of Science and Technology, U.S. Department
of Justice, Washington, DC 20531, January 2001.
[11] NELSON, C.V., Wide-Area Metal Detection System for Crowd Screening, in Proc.
SPIE AeroSense 2003 Conf., Sensors and Command, Control,Communication, and
Intelligence (C3T) Technologies for Homeland Defence and Law Enforcement II,
Orlando, FL (22–25 Apr 2003).
[12] NELSON, C.V, MENDAT, D., and HUYNH, T.B., Three-dimensional Steerable
Magnetic Field Antenna for Metal Target Classification, in Proc. SPIE AeroSense
2003 Conf., Detection and Remediation Technologies for Mines and Minelike Targets,
pp. 707–717, Orlando, FL, 22–25 Apr 2003
[13] DUCHATEAU, J.E., HINDERS, M., Using Ultrasound in Concealed Weapons
Detection, NDE Lab, Department of Applied Science, College of William and Mary,
April 2005
[14] ACHANTA, A., et al, Non-Linear Acoustic Concealed Weapons Detection,
Proceedings of the 34th Applied Imagery and Pattern Recognition Workshop, pp. 2127, 2005
[15] HAMILTON, M.F., and BLACKSTOCK, D.T., Non linear acoustics, Boston,
Academic Press, 1988
[16] KNOTT, E.F., SHAEFFER, J.F., and TULEY, M.T., Radar cross section, 2nd edition,
SciTech Publishing, pp. 82-83, 2003
[17] U.S. Department of Justice, Final Report – Demonstration of a Concealed Weapons
Detection System Using Electromagnetic Resonances, January 2001 [Online].
Available at: http://www.ncjrs.org/pdffiles1/nij/grants/190134.pdf [Accessed 10
Mach 2007]
[18] HUNT, A.R., HOGG, R.D., and FOREMAN, W., Concealed weapons detection using
electromagnetic resonances, Proc. of the SPIE, The International Society for Optical
Engineering, Conference on Enforcement and Security Technologies, Vol. 3575, pp.
62-67, Boston, MA, November 1998
[19] KOZAKOFF, D.J., and TRIPP, V., Antennas for Concealed Weapon Detection,
Proceedings of the 5th International Conference o Antenna Theory and Techniques,
pp. 65-69, Kyiv, Ukraine, May, 2005
[20] CHUI, C.K., An introduction to Wavelets, Academic Press, San Diego, CA, 1992
109
[21] APPLEBY, R., Passive millimetre-wave imaging and how it differs from terahertz
imaging, The Royal Society, Vol. 362, pp. 379- 393, Number 1815, February 15 2004
[22] YUJIRI, L., SHOUCRI, M., and MOFFA, P., Passive Millimeter-Wave Imaging,
IEEE microwave magazine, pp. 39-50, September 2003
[23] McMILLAN, R.W., CURRIE, N.C., FERRIS, D.D., WICKS, M.C., Concealed
Weapon Detection using Microwave and Millimeter Wave sensors,
Microwave and Millimeter Wave Technology Proceedings, 1998. ICMMT '98. 1998
International Conference on, Beijing, China
[24] LUUKANEN, A., Bolometer and Thz imaging, Millimetre-wave Laboratory of
Finland –MilliLab-Microsensing seminar 2-4-2008
[25] FEDERICI, J.F., GARYA, D., BARATB, R., ZIMDARSC, D., THz Standoff
Detection and Imaging of Explosives and Weapons, Proc. SPIE, Vol. 5781, 75, 2005.
[26] DICKINSON, J.C., et al., Terahertz for Military and Security Applications IV. Edited
by Woolard, D.L., Hwu, R.J., Rosker, M.J., Jensen, J.O., Proceedings of the SPIE,
Volume 6212, pp. 62120Q (2006).
[27] Optics & Laser Europe, October 2008, p15, Terahertz success relies on research
investment [online]. Available at:
http://images.iop.org/objects/optics/analysis/13/9/2/pdf.pdf [Accessed 15 March
2008]
[28] McMILLAN, R.W, Terahertz imaging, milimeter-wave radar, Conference on
Surveillance and. Assessment. Technologies for Law. Alabama, USA, 2004.
[29] THz-BRIDGE, Final report May 2004, Tera-Hertz radiation in Biological Research,
Investigation on Diagnostics and study of potential Genotoxic [online]. Available
at: http://www.frascati.enea.it/THz-BRIDGE/reports/THz-BRIDGE Final Report.pdf
[Accessed 10 March 2008]
[30] KRUSE, P.W., and SKATRUD, D.D., Editors, Uncooled Infrared Imaging Arrays
and Systems in Semiconductors and Semimetals, R.K Willardson, Eds., Academic
Press, New York, 1997
[31] SLAMANI, M.A., VARSHNEY, P.K., RAO, R.M., and ALFORD, M.G., An
integrated platform for the enhancement of concealed weapon detection sensors,
Proceedings of the SPIE, The 56 International Society for Optical Engineering,
Conference on Enforcement and Security Technologies, Vol. 3575, pp. 68-78, Boston,
MA, November, 1998
[32] CHEN, H.M., et al., Imaging for Concealed Weapon Detection, IEEE Signal
Processing Magazine, Vol. 22, No. 2, pp. 52-61, 2005.
110
[33] MAILLOUX, R.J., Phased Array Antenna Handbook, Second Edition, Artech House
Publishers; 2nd edition , March 31, 2005
[34] PAULTER, N.G., Hand-Held Metal Detectors for Use in Weapons Detection and
Contraband Detection, National Institute of Justice Guide 602-02, February 2003
[35] FELBER, F., et al., Handheld ultrasound concealed weapons detector, Proceedings of
the SPIE, The International Society for Optical Engineering, Conference on
Enforcement and Security Technologies, Vol. 3575, pp. 89-98, Boston, MA,
November, 1998.
[36] YUJIRI, L., et al., Passive Millimetre-wave camera, in Proc. SPIE Conf. on Passive
Millimeter-Wave Imaging Technology, pp. 15-22, Orlando, FL, 1997, 3064
[37] KLEIN, L.A., Millimetre-Wave and Infrared Multisensor Design and Signal
Processing Boston, MA: Artech, 1997
[38] SHEEN, D.M., McMAKIN, D.L., HALL, T.E., Three-Dimensional Millimeter-Wave
Imaging for Concealed Weapon Detection, IEEE Transactions on Microwave Theory
and Techniques, Volume 49, Issue 9, pp. 1581 - 1592, Sep 2001
[39] NELSON, C.V., Wide-Area Metal Detection System for Crowd Screening, in Proc.
SPIE AeroSense 2003 Conf., Sensors and Command, Control, Communication, and
Intelligence (C3T) Technologies for Homeland Defence and Law Enforcement II,
Orlando, FL (22–25 Apr 2003).
[40] McINTOSH, K.A., et al., Terahertz photomixing with diode lasers in lowtemperature-grown GaAs, Appl. Phys. Lett. 67, 3844 (1995).
[41] DUFFY, S.M., VERGHESE, S. and McINTOSH, K.A., Photomixers for ContinuousWave Terahertz Radiation, in Sensing with Terahertz Radiation, D. Mittleman ED,
Springer (2002).
[42] CHAN, Y.W., JOHNSON, M., Portable Concealed Weapon Detection Using
Millimeter Wave FMCW Radar Imaging, National Institute of Justice Office Science
and Technology, Final Report, , pp. 1-27, August 2001
[43] BERENGER, J., A perfectly matched layer for the absorption of electromagnetic
waves, Journal of Computational Physics, Vol. 114, No. 2, pp. 185-200, 1994
[44] KAUFFMAN, A.A., and KELLER, G.V., Inductive Mining Prospecting, Elsevier,
Amsterdam, 1985
111
[45] ASTEN, M.W., On the time-domain electromagnetic response of a conductive
permeable sphere, Journal of Applied Geophysics Volume 67, Issue 1, pp. 63-65, 1
January 2009
[46] CARUSO, M.J., and BRATLAND, T., A New Perspective on Magnetic Field Sensing,
Honeywell, SSEC Carl H. Smith and Robert Schneider, Nonvolatile Electronics, Inc.
[47] HONEYWELL, Sensor products, application note AN213, Set/reset function for
magnetic sensors [online]. Available
at: http://www.magneticsensors.com/datasheets/an213.pdf [Accessed 10 October
2007]
[48] HANED, N., and MISSOUS, M., Nano-tesla magnetic field magnetometry using an
InGaAs–AlGaAs–GaAs 2DEG Hall sensor, Sensors and Actuators A: Physical, Vol.
102, Number 3, pp. 216-222, 1 January 2003
112
APPENDICES
Appendix A: Hall Sensor Technology for CWD
Gun/Knife detection from software simulations and practical tests included in this thesis
show that magnetic field detected at 50cm from the target is very low (in the order of
microtesla). So it is necessary a sensor device able to measure low magnetic fields. Hall
sensor technology is very cheap and spread in the market. Looking in the market of hall
sensor device, the P15A from AHS technologies [http://www.ahsltd.com] is the most
sensitive. The next is the design of sensor device using DC technique able to measure
magnetic fields as low as 1uT with resolution in the low 100nT.
A.1 Offset cancellation
Hall elements exhibit piezoelectric effect, a change in electrical resistance proportional to
strain. It is desirable to minimize this effect. This is accomplished by oriented the Hall
elements on the IC to minimize the effect of the stress and by using multiple Hall elements.
Figure 1 show two Hall elements located in close proximity on an IC. They are positioned
in this manner so that both experience the same packaging stress. The first element has its
excitation applied along the vertical axis and the second along the horizontal axis.
Summing the two outputs eliminates the signal offset.
VH + Voffset (0o )
VH + Voffset (90o )
Figure A.1. Offset cancellation.
113
A.2 Design
The sensor system (see Figure 2) consists in a current source (LM334), which supply the
P15A hall sensor via a switching circuit for automatic offset cancellation (piezoelectric
effect). The commutated output from the switching circuit is connected to a low noise
instrumentation amplifier (INA126) with a gain of 1000 set by one external resistor. A low
pass active filter (MAX280) with a cut frequency of 0.1Hz averages the commutated output
of the instrumentation amplifier. Finally the filter output voltage is fed into an amplifier
with a gain of 20, which the overall gain of the circuit equal to 20000 (86dB).
Current source
LM334
Square wave
oscillator 1 kHz
Power supply
+/-5V
Switching Circuit
Instrumentation
amplifier,
G=1000
INA126
Hall sensor
P15A
Output signal
Low pass filter
Fc=0.1 Hz
MAX280
Amplifier G=20
OPA 177
Figure A.2. Hall Magnetometer Design.
114
Appendix B: Cell Phone Radiation Levels
Article published by CNET staff (updated April 21, 2009)
According to the Cellular Telecommunications Industry Association (CTIA), specific
absorption rate, or SAR, is "a way of measuring the quantity of radio frequency (RF)
energy that is absorbed by the body." For a phone to pass FCC certification, that phone's
maximum SAR level must be less than 1.6 watts per kilogram. In Europe, the level is
capped at 2W/kg while Canada allows a maximum of 1.6W/kg. The SAR level listed in our
charts represents the highest SAR level with the phone next to the ear as tested by the FCC.
Keep in mind that it is possible for the SAR level to vary between different transmission
bands and that different testing bodies can obtain different results. Also, it is possible for
results to vary between different editions of the same phone (such as a handset that is
offered by multiple carriers).
Table B.1. SAR level of cell phones.
Manufacturer and model
Motorola V195s
Motorola ZN5
Motorola VU204
Motorola W385
RIM BlackBerry Curve 8330 (Sprint)
RIM BlackBerry Curve 8330 (U.S. Cellular)
RIM BlackBerry Curve 8330 (Verizon Wireless)
Motorola Deluxe ic902
T-Mobile Shadow (HTC)
Motorola i335
115
SAR level(digital)
1.6
1.59
1.55
1.54
1.54
1.54
1.54
1.54
1.53
1.53
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement