Acknowledgment [2] J. Vey, M. M. Driscoll, E. P. EerNisse, R. T. Filler, R. M. Garvey, W. J.

The author thanks Dr. Kate Remley, Editor of IEEE Riley, R. C. Smythe, and B. D. Weglein, “The effects of acceleration
on precision frequency sources,” U.S. Army Lab. Command Elec-
Microwave Magazine, and the reviewers for their insight- tron. Technol. Dev. Lab., Fort Monmouth, NJ, Res. Dev. Tech. Rep.
ful suggestions that improved this article. SLCET-TR-91-3 (Rev. 1).
[3] G. R. Kurzenknabe, “Phase noise under vibration in crystal
References oscillators,” in Measurement Product Dig., May/June 1989, pp.
[1] S. Goldman, “Phase noise leakage through a mixer,” MSN CT, pp. 33–35.
80–96, Nov. 1987.

Correction
Due to a production error, four references were left [13] D. C. Schleher, Electronic Warfare in the Information Age. Boston:
Artech House, 1999, pp. 380–381, (6.10) and (6.6.12).
out of the Application Note by Samuel J. Caprio [1]
[14] R. L. Goodwin, “Ambiguity-resistant three- and four- channel
in the last issue. The following are the correct refer- interferometers,” Tactical Electronic Warfare Division, Naval
ences that should have appeared on the list, with Res. Lab., Washington, DC, p. 23, Table 3-1.
added information that should have appeared in the [15] S. J. Caprio, “Effects of single-tone spurious signals in ESM re-
ceivers,” Microwave J., vol. 53, no. 10, pp. 64–70, Oct. 2010, (8)
citation of the reference in text (e.g., equation num-
and Figure 1.
ber, etc.). We apologize for this error and any confu- [16] R. L. Goodwin, “Ambiguity-resistant three- and four- channel
sion it may have caused. interferometers,” Tactical Electronic Warfare Division, Naval
Res. Lab., Washington, DC, pp. 4–5, Figs. 1 and 2.
[12] W. Bishop, “Interferometer accuracy and DOA cell size trade-
offs,” private communication, Apr. 19, 1978.
Reference
[1] S. J. Caprio, “Specifying upper bounds on phase noise in phase-
locked oscillators in electronic warfare systems—Part I [Appli-
Digital Object Identifier 10.1109/MMM.2011.2178290 cation Notes],” IEEE Microwave Mag., vol. 11, no. 7, pp. 96–112,
Date of publication: 13 January 2012 2011.

A Review of Nonimaging Stand-Off Concealed

Threat Detection with Millimeter-Wave Radar
■ Stuart William Harmer, Nicholas Bowring, David Andrews,
Nacer Ddine Rezgui, Matthew Southgate, and Sarah Smith

T
here is now, more than ever before, a need for to conceal on the body, and successful detection is re-
technologies that enable the screening of people quired at considerable distance or stand-off range be-
from a distance. A wide variety of weapons can fore the bomber reaches the target area.
be easily concealed under clothing and carried into Current electromagnetic screening technologies,
crowded public sites to target national infrastructure, such as metal detectors, explosive trace detection, and
spread fear, and inflict mass murder and casualties. x-ray backscatter are very restricted in the coverage,
The most feared and devastating terrorist weapon is stand-off capability, flexibility, and efficacy they afford.
the suicide bomb or person borne improvised explo- Perhaps the most widely known of these technologies
sive device (PBIED). Such devices are relatively simple is millimeter-wave or submillimeter-wave imaging [1]–
[7], where millimeter-wave images are interpreted by a
trained human operator or by computer algorithm to
ascertain the presence or absence of concealed threats.
Stuart William Harmer ([email protected]),
Nicholas Bowring, David Andrews, Nacer Ddine Rezgui, Millimeter-wave imaging is certainly the most com-
Matthew Southgate, and Sarah Smith are with monly employed technique for stand-off detection of
The School of Electrical and Electronic Engineering at concealed threats.
Manchester Metropolitan University, Manchester, U.K. A prerequisite of a screening system based on imag-
ing is the ability to form an image in which there exists
Digital Object Identifier 10.1109/MMM.2011.2174125 a contrast between concealed items and the human
Date of publication: 13 January 2012
body. This condition requires that the portion of the

160 1527-3342/12/$31.00©2012 IEEE January/February 2012

electromagnetic spectrum utilized for image forma- Current electromagnetic screening
tion is not significantly attenuated by standard cloth-
ing and textiles and also that the objects to be resolved technologies, such as metal detectors,
have physical properties, temperature, emissivity, explosive trace detection, and x-ray
reflectance, etc., which are different from that of the backscatter are very restricted in
human body to provide contrast in the formed image.
Also important is that the image formed has sufficient the coverage, stand-off capability,
spatial detail so that the shape of the object can be rec- flexibility, and efficacy they afford.
ognized sufficiently well for the identification of con-
cealed objects to be made, for example, to enable the
user to discriminate between a benign object, such as a
mobile phone handset, and a handgun. The only por- 3 m with a good, passive millimeter imaging system.
tion of the electromagnetic spectrum where these con- Spatial resolution is limited by the frequency band that
ditions are readily met is in the millimeter-wave and the millimeter-wave camera operates over, where higher
submillimeter-wave region. In this region, imaging frequency bands give finer resolution and the aperture
systems with sufficiently high spatial resolution can size of the optical components of the imaging system are
be realized using the comparably small antenna sizes reduced. The aperture size of the receiver elements that
needed for a practical and deployable system. comprise the sensitive area for signal detection is also
The science of millimeter-wave imaging is based important for determining the spatial resolution of the
on spatially mapping the distribution of radiometric camera, however, a reduction in size of these indefinitely
contrast in a scene. For passive imaging, this contrast does not improve resolution beyond the diffraction limit
is a function of the temperature and emissivity dis- set by the system’s aperture. Sensitivity is determined by
tribution of the scene and the reflectance distribution receiver noise, the receiver collection area, and the dwell
within the scene [8]. In an outdoor environment, the time of the camera. Optical aperture and the receiver
apparent temperature of the sky is very different from aperture are typically matched to give the largest receiver
the human body temperature and, therefore, climactic aperture possible while keeping the system diffraction
conditions contribute significantly to the quality of the limited by the optical aperture of the system. This is done
image formed. Active millimeter-wave imaging uses to maximize sensitivity of the receiver while not degrad-
artificially generated radiation to illuminate the scene. ing spatial resolution. Integration or dwell time is limited
Here the formed image is dominated by the scene’s by the need for real-time video frame rate capture, how-
reflectance distribution, which is sensitive to the ori- ever, the intensity of natural emission from the human
entation of reflecting surfaces and their diffusivity and body in the millimeter region is very low compared
specular reflections. This gives rise to areas of high with that in the infrared spectrum (about two orders
brightness, known as “glinting,” which can plague the of magnitude fewer photons emitted per unit area per
active methods [9]. Recently, there has been interest in unit time). Accordingly, dwell times need to be longer
combining passive and active millimeter-wave imag- for low-noise millimeter-wave imagery than is required
ing techniques to improve imaging capability [10] and for thermal imagery, although this disadvantage is off-
in making use of the different polarizing effect of con- set by the fact that millimeter-wave receivers have better
cealed items from that of the human body to enhance thermal noise performance than their infrared counter-
discrimination [11]–[13]. parts. Typical noise levels in a state-of-the-art passive
Millimeter-wave imaging has limitations, some of millimeter-wave imaging system are approximately
which are best described by quasi-optical analysis, in equivalent to a 0.4 K temperature differential, limiting
that the imagery and the figures of merit of that imagery contrast to this level. Existing millimeter-wave imaging
can be described in a similar way to those used to char- systems periodically sample different solid angles of the
acterize the performance of an optical or infrared image total field of view, building up the entire scene by scan-
capture system. There are two main limitations to the ning and overlapping these smaller angled samples. This
image quality obtainable from millimeter-wave imaging can be done with a single receiving channel by scanning
systems: spatial resolution and radiometric sensitivity. the directional pattern of receiving antenna, modified
Spatial resolution characterizes the smallest features of by the addition of refractive or reflective quasi-optical
the scene that can be resolved in the image. For a given elements across the two dimensions of the scene or by
frequency band of operation (millimeter-wave radia- using multiple receiver channels and scanning spatially,
tion is the band occupied from 30 GHz to 300 GHz) and thus building up multiple image pixels at every scan-
camera aperture size, the smallest feature in the object ning position. Scanning millimeter-wave cameras have
resolvable in the image will scale linearly with range. At low frame rates and strict requirements are imposed by
long ranges, the image resolution will be much poorer the mechanical scanning of the beam steering optics. The
than for the same target closer in. Spatial resolution of operation of scanning can be made redundant if a suffi-
several millimeters is typically attainable at a range of cient number of receiver channels are used such that the

January/February 2012 161

 • Interpretation of millimeter-wave imagery is
The operation of scanning can be
difficult and does not lend itself well to autono-
made redundant if a sufficient mous decision making, whereas interpretation
number of receiver channels are used of scattered radar returns can, for certain types
of objects, be far quicker, more reliable, and dis-
such that the required field of view
criminatory.
can be met solely by the total viewing • Spatial resolution requirements for a radar sys-
angle obtained. tem are far more relaxed, since imagery is not be-
ing sought. Consequently, stand-off ranges can be
extended considerably beyond those achievable
required field of view can be met solely by the total view- with an imaging technique.
ing angle obtained. This approach is identical in concept • Speed of operation of a radar system is higher
to the focal plane arrays that are used to capture images than for a millimeter-wave imaging system, with
in the optical and infrared bands of the electromagnetic capture rates well in excess of 50 frames per sec-
spectrum. The disadvantages of such systems include ond possible.
the difficulty of positioning a large number of receiver • Radar systems are mechanically simpler than
channels in the focal plane of the antenna, problems aris- scanned millimeter-wave imaging cameras since
ing in constructing arrays with spatially overlapping the scene does not need to be scanned using rap-
beams, and the very high total cost of the systems, which idly moving mirrors, as the radar is a point and
is dominated by the increased number of receiver chan- detect system.
nels required to achieve the required field of view. Mil- • The radar system can be integrated with conven-
limeter-wave imaging is a very well understood, highly tional video or infrared imagery to show the po-
developed and mature technology and improvements in sition of the millimeter-wave beam in real time.
performance will now come at a high cost, both finan- • Size and weight saving of a millimeter-wave ra-
cially and in development time. dar system over an imaging system are consid-
erable as a result of the reduced number of com-
Millimeter-Wave Radar ponents, optics, mechanical drives, and power
A promising alternative method of concealed threat requirements.
detection is to avoid image formation altogether and • The cost of a radar system is significantly lower
the restrictions that an imaging approach imposes, the since fewer expensive millimeter-wave compo-
greatest of which are those of limited spatial resolu- nents are required and expensive optical compo-
tion and scene contrast, making interpretation and nents are not required.
identification of concealed objects by millimeter-wave Highly directional antennas may be made at milli-
imaging difficult and often unreliable. But the essen- meter-wave frequencies so that quasi-optical beams of
tial ability of millimeter-wave radiation to penetrate millimeter radiation may be used to probe the target at
clothing and gather information on concealed objects large stand-off ranges, where the beam width or spot
must be retained. A millimeter-wave radar system has size may be sufficiently localized to cover an area com-
the ability to probe objects concealed under clothing mensurate with, or smaller than, a typical human torso
with identification of those objects being carried out by (see Figures 1 and 2). Smaller beam sizes are required
analysis of the scattered radar return [14]–[22]. This has to detect smaller objects reliably, and therefore there
distinct advantages over imagery. is always a trade-off between aperture size, stand-off
range, frequency, and minimum resolvable target size.
The relationship between these competing param-
eters can be effectively summarized by the expression:
L # Rc/Dv, where L is the largest linear dimension of
L the concealed object to be detected, R is the range to
the target, c is the speed of light in vacuum, D is the
D
diameter of the antenna aperture, and v the frequency
of the millimeter-wave radiation. A beam that is con-
siderably smaller than the largest linear dimension of
R a concealed object will fail to characterize the whole
object fully since the radar return will not originate
Figure 1. An illustration of a millimeter-wave radar
from the entire object. Conversely a beam that is con-
system in use to individually screen people in a crowd
scenario. The beam can be aimed at an individual at siderably wider than the largest linear dimension of
standoff ranges; guidance of the beam is achieved by a concealed object will not localize the object suffi-
colocation of video imagery provided by a camera mounted ciently, and the radar return will contain the signatures
on the system. of other nearby items (see Figure 3).

162 January/February 2012

 Over the last few years, there have been several A promising alternative method
nonimaging millimeter and microwave radar-based
products released and patents filed that describe these of concealed threat detection is to
systems, which claim to be able to detect various threat avoid image formation altogether
items concealed on the human body. Currently, there and the restrictions that an imaging
are a handful of companies that are offering radar-
based concealed threat detection devices. Ariel Uni- approach imposes.
versity Research and Development Company Ltd have
developed a small, hand-held device [23]–[24]. This sys-
tem sends out a train of millimeter-wave pulses using radar information of the illuminated target. The system
a modulated Gunn oscillator and measures characteris- automatically tracks the radar beam onto the person to
tics of the scattered radar return. An autonomous threat be screened by using video imagery from a colocated
decision is made based on comparing these measure- video camera, and delivers an autonomous threat deci-
ments to a library of previously measured threat situa- sion to the user.
tions. Safe Zone Systems [25]–[26] have filed a number The authors of this article, based in the School of
of patents for a microwave-based radar system for Electrical and Electronic Engineering at Manchester
concealed threat detection, and sell such a system for Metropolitan University, have developed a W-band
the detection of concealed threats. The Safe Zone Sys- (75–110 GHz) millimeter wave, polarimetric radar sys-
tems apparatus illuminates the target with a microwave tem and have recently been granted a U.K. patent for
pulse and uses the natural resonances of the object to this invention [29]. The authors’ system shares some
enable an autonomous threat-or-no-threat decision to common features with earlier work in the area [23]–
be made. The Science, Engineering, and Technology [28], but also has features which make it distinct from
(SET) Corporation has developed the CounterBomber other millimeter-wave radar devices. Direct compari-
system [27]–[28], which is currently deployed by U.S. son of the performance of the authors’ system with sim-
forces in Iraq and Afghanistan and is probably the most ilar systems has not been made although it is clear
advanced of the current commercially available tech- that the CounterBomber system is commercially at a
nologies. This system has recently attracted very sig- more advanced stage and has almost certainly under-
nificant funding, with a US$ 50 million contract being gone more rigorous field trials. The equipment has
awarded from the U.S. Army to detect suicide bombers been developed for the detection of threat objects,
at long range (~100 m). CounterBomber uses hetero- particularly PBIED, concealed under clothing upon
dyne millimeter-wave receivers to obtain polarimetric the human body at stand-off ranges of up to 25 m;
a system that operates at longer range is planned. In
operation, the system implements swept frequency
continuous wave radar with low-cost components to
deliver a compact, ultrawide bandwidth (UWB), high
resolution ( around ~ 10 mm ) radar system capable
of detecting, discriminating, and identifying a wide

400
350
Beam FWHM (mm)

300
250
200
150
100
50
0
5 10 15 20 25 30
Range-m

Figure 3. Measured FWHM (3 dB points) of the

Figure 2. Photograph of an early prototype of a radar-based Manchester Metropolitan University concealed threat radar
concealed threat detection system developed at Manchester system’s beam at 94 GHz for various ranges are shown
Metropolitan University; signal processing and autonomous by circular marker points; the theoretical beam width is
threat detection is carried out in real time on a remote PC. represented by the continuous line.

January/February 2012 163

 spectrum of concealed threat items. Although not There are two key radar principles that we make use
required for PBIED fragmentation device detection, of for threat detection: polarization changes induced
where a simple co- and cross-polarized radar return by the concealed object, and the axial structure profile
amplitude ratio is all that is required, threat detection information of the concealed object. Polarimetric infor-
can be rendered autonomously by application of a neu- mation is realized by transmitting millimeter radia-
ral network to the scattered polarimetric, time domain tion with electric field aligned horizontally (across
radar return. Such a system may be taught to alarm or the body), to minimize the radar cross section of an
reject certain classes of objects, allowing both highly upright human body, and measuring the scattered
specific and broad spectrum threat detection [16]–[20]. field both horizontally and vertically (along the body)
The overall cost of the system described here is kept low using two receiver channels (see Figure 4). Reorienta-
by use of direct detection receivers, which are consid- tion of a fraction of the incident horizontally polarized
erably less expensive than implementing a heterodyne millimeter-waves into the vertical polarization state is
receiver operating over the same wide frequency band. predominantly due to multiple reflections in the case
The millimeter-wave radar system described is by no of fragmentation within a PBIED, whereas for single
means a foolproof method of detecting and identifying solid items, which do not give rise to multiple reflec-
objects concealed on the human body, and false posi- tions, the reorientation is due to stronger reflection of
tive and false negative detection events are certainly electric field components in certain directions than
observed using this system. False alarms are gener- in others, and is thus a function of the object’s shape
ally more prevalent when attempting to detect smaller and symmetry.
objects, and objects that present a shape that is easily Axial structure profile is measured by transmission
mimicked by benign objects, e.g., a metal cased foun- of UWB millimeter-wave radar waveforms to spatially
tain pen and a small knife. Larger objects and those that resolve scattering surfaces that are sufficiently sepa-
have a unique shape, e.g., a large handgun, are easier to rated. The achievable axial resolution is
detect and classify [16]. Object orientation on the body
c
also contributes to the ease or difficulty of detection DR $ ,
2BW
and identification of concealed objects. For example,
a handgun presented side on is more easily detected where BW is the bandwidth scanned. By scanning
than the same weapon in a frontal aspect. Such effects over the entire W-band (~ 35 GHz), features separated
mean that detection and identification is best carried by only ~ 10 mm are resolvable, allowing the pres-
out while the person being screened is walking and ence of objects in front of the body, usually under lay-
presenting an ever-changing body aspect with respect ers of clothing, to be ascertained (see Figure 5). This
to the millimeter radar system. high-range resolution is a key factor, which helps
distinguish the Manchester Metropolitan University
system from similar systems disclosed in patents [24],
[26], [28]–[29].
The radar front-end is based on monolithic micro-
Body with No Objects wave integrated circuit (MMIC) direct-detection
0.25 Body with PBIED
Crosspolar Amplitude

receivers connected to orthogonally orientated, hori-

Body with Clutter
(Arbitrary Units)

0.2 zontal and vertical gain horns and a voltage controlled

oscillator (VCO) and multiplier to provide the swept
0.15 frequency output. The transmitter is a gain-horn located
0.1 at approximately the focal length of a 300 mm diameter
polyethylene lens (see Figure 6). A reasonable estimate
0.05 of the upper usable beam width of 350 mm, commen-
surate with the typical width of the human body, sets
0
0 0.2 0.4 0.6 0.8 1 the maximum effective range of the system to ~ 25 m
Copolar Amplitude (Arbitrary Units) (see Figure 3). Range-to-target information is provided
by a conventional radar range-finding technology.
Figure 4. Measured radar amplitude returns from the Both copolarized and cross-polarized radar return
Manchester Metropolitan University concealed threat radar signals are transformed from the frequency domain to
system showing polarimetric discrimination of a person
the time domain and scaled according to the range at
with (circle symbols) and without a PBIED (diamond
which the target is measured to be. This accounts for
symbols) and with concealed nonthreat items (plus
symbols) at a range of 25 m. Multiple scans are taken and the wide variation in signal strengths that the system
the amplitude of the radar return in both copolarized and encounters over the variable stand-off operating range
cross-polarized channels are plotted, in this case there is at which screening can take place, currently from a few
very obvious clustering such that the case where a PBIED meters out to 25 m. The scaled temporal data is then
is present is clearly differentiable. fed to a neural network classifier which has a priori

164 January/February 2012

 0 0
20 Body Alone 20 Body with PBIED
40 40
Number of Sweeps

Number of Sweeps
60 60
80 80
100 100
120 120
140 140
160 160
180 180
200 200
2,900 2,950 3,000 3,050 3,100 3,150 3,200 3,250 3,300 2,900 2,950 3,000 3,050 3,100 3,150 3,200 3,250 3,300
Target Depth (mm) Target Depth (mm)
(a) (b)

Figure 5. Axial depth profiles for a person at an indicated range of ~ 3 m. The plots show the strength of return (red being
highest and blue zero) at different ranges (horizontal axis) as a function of increasing numbers of frequency scans of target
(vertical axis). (a) is for a body with no explosive device while (b) is for a body with a PBIED without any fragmentation; the
device clearly shows up as a area of significant radar return immediately in front of the body response.

trained networks for items of interest; these training not reproduced by any other commonly carried benign
sets can include a broad spectrum of threat objects or objects. The exact type of fragmentation used, includ-
can instead focus on very specific items. Detection and ing nails, nuts and bolts, and ball bearings, did not
discrimination of PBIED containing shrapnel compo- greatly affect performance. Other, smaller, threat items,
nents (see Figures 7 and 8) is particularly effective, such as handguns, can be detected and differentiated
with the system being able to differentiate successfully with some success, but at lesser standoff ranges, where
between a person carrying such a device and a per- the beam size is smaller and more suited to the detec-
son without an explosive device but carrying the usual tion of these items. By taking multiple scans, a more
accoutrements of modern living such as a mobile robust detection of the threat level can be made. The
phone, iPod, car keys, wallet, etc, at ranges up to 25 m effect of multiple scans on the discrimination of PBIED
(see Figure 4). by both polarimetric and axial profile determination is
Fragmentation components of a simulated PBIED shown in Figures 4 and 5. In walking tests, where the
present such a good target because the small fragments subject walks towards the radar system as readings are
whose size is around one centimeter to provide a strong taken, the range-scaled polarimetric data clearly allows
cross-polarized return from a large body area, which is differentiation between a person carrying a simulated

Radar Ranger
Lens

Multiplier to
Microwave
W-Band
Source
Frequencies
Control,
Processing, and
Threat
Evaluation
Single Receiver Channel

Analog to Zero
Low-Noise
Digital Bias
Amplifier
Converter Detector

Figure 6. Schematic of Manchester Metropolitan University concealed threat detection radar system showing major system
blocks.

January/February 2012 165

 50 80

40
60

30
Counts

Counts
40
20

20
10

0 0
0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3
Copolar Amplitude X-Polar Amplitude
(a)

50 20

40
15

30
Counts

Counts
10
20

5
10

0 0
0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3
Copolar Amplitude X-Polar Amplitude
(b)

Figure 7. Amplitude histograms of radar return in copolarized and cross-polarized (X-polar) channels for (a) a person
without a PBIED and (b) with PBIED. The data was taken for a person walking toward the radar system, each data point has
been range scaled (to remove effects of range on amplitude), and normalized with reference to a test target.

PBIED, as shown in Figure 9, and a person without

such a device as in Figure 7.
Although millimeter-wave imaging is a useful tech-
nology for security screening of people, the technique
currently has limitations which prevent the effective

Figure 8. A representative example of a piece of a simulated

PBIED used in system testing at Manchester Metropolitan
University; shrapnel is embedded in wax to mimic a suicide
vest. Other samples were made using nails and ball-bearings Figure 9. Simulated PBIED suicide vest used in system
embedded in wax and results were similar. testing at Manchester Metropolitan University.

166 January/February 2012

detection of threat items, and in particular PBIED, [18] D. A. Andrews, N. Rezgui, S. E. Smith, N. J. Bowring, M. South-
at useful standoff ranges (. 20 m). UWB millimeter- gate, and J. G. Baker, “Detection of concealed explosives at stand-
off distances using wide band swept millimetre waves,” in Proc.
wave radar systems can detect the presence or absence SPIE., 2008, vol. 7117.
of certain concealed threat objects at ranges in excess of [19] D. A. Andrews, N. J. Bowring, N. D. Rezgui, M. Southgate, E.
20 m by reliance on information carried in the scattered Guest, S. W. Harmer, and A. Atiah, “A multifaceted active swept
radar return from the person. Consequently, millime- millimetre-wave approach to the detection of concealed weapons,”
in Proc. SPI, 2008, vol. 7117.
ter radar offers a viable alternative, and a potentially [20] N. J. Bowring, J. G. Baker, N. Rezgui, M. Southgate, and J. F. Alder,
complementary screening method to conventional “Active millimetre wave detection of concealed layers of dielectric
millimeter-wave imaging, and will undoubtedly find material,” in Proc. SPIE, 2007, vol. 6540.
increasing use in concealed threat detection. [21] L. Yong, Y. T. Gui, N. J. Bowring, and N. Rezgui, “A microwave
measurement system for metallic object detection using swept fre-
quency radar,” in Proc. SPIE, 2008, vol. 7117.
References [22] A. Agurto, Y. Li, G. Y. Tian, N. Bowring, and S. Lockwood, “A re-
[1] R. Appleby and R. N. Anderton, “Millimeter-wave and submilime- view of concealed weapon detection and research on perspective,”
ter-wave imaging for security surveillance,” Proc. IEEE, vol. 95, no. in Proc. IEEE ICNSC, 2007, pp. 443–448.
2, pp. 1683–1690, Aug. 2007. [23] [Online]. Available: http://arielrnd.com/
[2] S. Oka, H. Togo, N. Kukutsu, and T. Nagatsuma, “Latest trends in [24] B. Kapilevich and A. Lipeles, “Hand-held device and method for
millimeter-wave imaging technology,” Progress Electromagn. Res. Lett., detecting concealed weapons and hidden objects,” International
vol. 1, pp. 197–204, 2008. Application 148327, June 18, 2007.
[3] M. C. Kemp, A. Glauser, and C. Baker, “Recent developments [25] [Online]. Available: http://www.detectanddefeat.com/
in people screening using terahertz technology—Seeing the [26] J. Hausner and J. M. West, “Object detection method and appara-
world through terahertz eyes,” in Proc. SPIE, 2006, vol. 6212, tus,” U.S. Patent 0 052 576, Mar. 8, 2007.
pp. 27–34. [27] [Online]. Available: http://www.setcorp.com/
[4] R. Appleby and H. B. Wallace, “Standoff detection of weapons and [28] R. J. Douglass, J. D. Gorman, and T. J. Burns, “System and method
contraband in the 100 GHz to 1 THz region,” IEEE Trans. Antennas for standoff detection of human carried explosives,” U.S. Patent 7
Propagat., vol. 55, no. 11, pp. 2944–2956, Nov. 2007. 800 527, Oct. 11, 2005.
[5] J. F. Federici, B. Schulkin, F. Huang, G. Dale, R. Barat, F. Oliveira, [29] N. Bowring, D. Andrews, N. D. Rezgui, and S. Harmer, “Remote
and D. Zimdars, “THz imaging and sensing for security applica- detection and measurement of objects,” U.K. Patent 2 458 764,
tions—Explosives, weapons and drugs,” Semiconduct. Sci. Technol., Mar. 18, 2009.
vol. 20, no. 7, pp. S266–S280, July 2005.
[6] D. Sheen, D. McMakin, and T. E. Hall, “Three-dimensional milli-
meter-wave imaging for concealed weapon detection,” IEEE Trans.
Microwave Theory Tech., vol. 49, no. 9, pp. 1581–1592, 2001.
[7] L. V. Volkov, A. I. Voronko, N. L. Volkova, and A. R. Karapetyan,
“Active MMW imaging technique for contraband detection,” in
Proc. 33rd European Microwave Conf., 2003, pp. 531–534.
[8] S. E. Clark, J. A. Lovberg, C. A. Martin, and V. Kolinko, “Passive
millimeter-wave imaging for airborne and security applications,”
in Proc. SPIE, 2003, vol. 5077, pp. 16–21.
[9] P. F. Goldsmith, C. T. Hsieh, G. R. Huguenin, J. Kapitzky, and E.
L. Moore, “Focal plane imaging systems for millimetre wave-
lengths,” IEEE Trans. Microwave Theory Tech., vol. 41, no. 10, pp.
1664–1675, Oct. 1993.
[10] R. Zouaoui, R. Czarny, F. Diaz, A. Khy, and T. Lamarque, “Multi-
sensor millimeter-wave system for hidden objects detection by col-
laborative screening,” in Proc. SPIE, 2011, vol. 8022.
[11] D. M. Sheen, T. E. Hall, R. H. Severtsen, D. L. McMakin, B. K.
Hatchell, and P. L. J. Valdez, “Standoff concealed weapon detec-
tion using a 350-GHz radar imaging system,” in Proc. SPIE, 2010,
vol. 7670 , pp. 117–126.
[12] D. M. Sheen, D. L. McMakin, W. M. Lechelt, and J. W. Griffin, “Cir-
cularly polarized millimeter-wave imaging for personnel screen-
ing,” in Proc. SPIE, 2005, vol. 5789.
[13] D. M. Sheen, D. L. McMakin, T. E. Hall, and R. H. Severtsen, “Ac-
tive millimeter-wave standoff and portal imaging techniques for
personnel screening,” in Proc. IEEE Conf. Technologies for Homeland
Security, 2009, pp. 440–447.
[14] A. S. Ibrahim, K.J. R. Liu, D. Novak, and R. B. Waterhouse, “A
subspace signal processing technique for concealed weapons de-
tection,” in Proc. IEEE ICASSP, 2007, pp. 401–404.
[15] J. Hausner and N. West, “Radar based concealed threat detector,”
in Proc. IEEE MTT-S Int. Microwave Symp., 2007, pp. 765–768.
[16] N. Rezgui, D. Andrews, N. Bowring, S. Harmer, and M. Southgate,
“Standoff detection of concealed handguns,” in Proc. SPIE., 2008,
vol. 6948.
[17] D. A. Andrews, S. E. Smith, N. Rezgui, N. J. Bowring, M. South-
gate, and S. W. Harmer, “A swept millimetre-wave technique for
the detection of concealed weapons and thin layers of dielectric
material with or without fragmentation,” in Proc. SPIE, 2009, vol.
7309.

January/February 2012 167