1. Lyles College of Engineering
Electrical & Computer Engineering Department
California State University, Fresno
ECE298 Project
Project Title: Simulation of Ultra Wide Band Signal Returned From Walking Human
Student Name: Jaideep Chowdhury
ID: 106906676
Semester: Spring 2013
Units: 3
Course# 34204
Instructor Name: Dr. Youngwook Kim
Instructor Approval:
Date: May 09, 2013
Presented in partial fulfillment of the Master of Science
Degree in Electrical Engineer
2. 1
Table of Contents
ACKNOWLEDGEMENT ...............................................................................................................................................3
LIST OF FIGURES ..........................................................................................................................................................4
LIST OF TABLES ............................................................................................................................................................6
ABSTRACT ......................................................................................................................................................................7
1. INTRODUCTION.............................................................................................................................................8
2. BACKGROUND ............................................................................................................................................. 10
2.1. RADAR...............................................................................................................................................................10
2.2. ULTRA WIDE BAND RADAR .........................................................................................................................13
3. DEVELOPMENT OF SIMULATOR............................................................................................................. 16
3.1. POSER SOFTWARE ..........................................................................................................................................16
3.2. BVH FORMAT...................................................................................................................................................17
3.3. POINT SCATTER MODEL OF HUMAN .........................................................................................................19
4. MEASUREMENTS......................................................................................................................................... 21
5. DATA ANALYSIS AND VERIFICATION.................................................................................................... 24
5.1. HUMAN MANNEQUIN STANDING SIDEWISE WITH NO HANDS IN ANECHOIC CHAMBER............24
5.2. HUMAN MANNEQUIN STANDING SIDEWISE WITH ALL HANDS IN ANECHOIC CHAMBER..........26
5.3. STANDING HUMAN MANNEQUIN FACING ANTENNA FRONT IN ANECHOIC CHAMBER..............27
5.4. DIFFERENT FACES OF A CYLINDER FIXED AT CERTAIN DISTANCE FROM THE RADAR..............28
5.5. TWO CYLINDERS AT SAME DISTANCE, ONE FIXED AND OTHER FACE CHANGES ........................30
5.6. TWO CYLINDERS, ONE FIXED AND OTHER APPROACH THE FIXED FROM BACK ..........................32
5.7. TWO CYLINDERS, ONE FIXED AND OTHER APPROACH THE FIXED ONE FROM BACK IN AN
ANECHOIC CHAMBER.................................................................................................................................................35
5.8. TWO CYLINDERS, ONE FIXED AND OTHER MOVING AT A FIXED RADIUS FROM RADAR...........37
5.9. TWO STANDING HUMANS, ONE FIXED AND OTHER APPROACH FIXED ONE FROM BACK .........39
5.10. TWO HUMANS, ONE FIXED AND OTHER MOVING AT A FIXED RADIUS FROM RADAR................42
6. CALCULATIONS .......................................................................................................................................... 44
6.1. PROPAGATION OF ELECTROMAGNETIC WAVE WITHIN THREE LAYERS OF HUMAN SKIN ........44
6.2. REFLECTIONS COEFFICIENT BETWEEN TWO DIELECTRIC MEDIA....................................................45
6.3. FOUR LAYER MODEL.....................................................................................................................................50
6.4. CALCULATION OF REFLECTION COEFFICIENT BETWEEN HUMAN SKIN AND AIR .......................53
6.5. CALCULATION OF REFLECTION COEFFICIENT CONSIDERING FRESNEL‟S ANGLE OF
INCIDENCE.....................................................................................................................................................................55
4. 3
ACKNOWLEDGEMENT
I would like to thank Dr. Youngwook Kim for his constant guidance throughout the project and
providing the UWB radar and other equipments for the project. Whenever I need any help or
running out of ideas, Dr. Kim was always there to give new ideas of approach.
I would like to thank Abhishek Parmar, Majid Ehtesham Borourjerdi, graduate students at Fresno
State who helped in getting the data for human. They stand for different positions in front of radar
for data collection.
In addition, I would like to thank my family, without their help I would not have accomplished my
goal. Also, I would like to thank the other ECE faculty of Lyles College of Engineering for their
guidance and facilities they provide to work in the lab environment.
5. 4
LIST OF FIGURES
Figure 1: Mono-static radar [13].................................................................................................................10
Figure 2: Bi-static radar [13] .......................................................................................................................11
Figure 3: Classification of Radar..................................................................................................................12
Figure 4: UWB Signal [16] ...........................................................................................................................13
Figure 5: Power Spectrum of UWB [17].......................................................................................................14
Figure 6: Human Walking Style in Poser......................................................................................................16
Figure 7: Header Section.............................................................................................................................18
Figure 8: Data Section.................................................................................................................................18
Figure 9: Point Scatter Model of Human.....................................................................................................19
Figure 10: Radar Configuration...................................................................................................................22
Figure 11: Example of UWB return from a target........................................................................................23
Figure 12: Human mannequin with no hands..............................................................................................24
Figure 13: Human mannequin with left hand..............................................................................................24
Figure 14: Range profile of human mannequin without any hands and with left hand only.........................25
Figure 15: Human mannequin sidewise with all arms .................................................................................26
Figure 16: Human mannequin sidewise with right hand..............................................................................26
Figure 17: Range profile of human mannequin with all hands and right hand only......................................26
Figure 18: Front human standing................................................................................................................27
Figure 19: Front human standing with one leg forward...............................................................................27
Figure 20: Range profile of human standing normally and one leg forward.................................................27
Figure 21: UWB radar with single target (cylinder)......................................................................................28
Figure 22: UWB return signal from one cylinder .........................................................................................28
Figure 23: UWB radar with two targets (One fixed and one rotating cylinder).............................................30
Figure 24: UWB return signal from two targets (one fixed and one rotating cylinder) .................................30
Figure 25: Moving cylinder furthest from fixed cylinder..............................................................................32
Figure 26: Moving cylinder little closer from fixed cylinder .........................................................................32
Figure 27: Moving cylinder closer from fixed cylinder .................................................................................32
Figure 28: Moving cylinder by side of fixed cylinder....................................................................................32
Figure 29: Moving cylinder furthest from fixed cylinder and fixed cylinder removed thereafter..................33
Figure 30: Moving cylinder little closer from fixed cylinder and fixed cylinder removed thereafter .............33
Figure 31: Moving cylinder closer from fixed cylinder and fixed cylinder removed thereafter .....................34
Figure 32: Moving cylinder furthest from fixed cylinder..............................................................................35
Figure 33: Moving cylinder little closer from fixed cylinder .........................................................................35
Figure 34: Moving cylinder closer from fixed cylinder .................................................................................35
Figure 35: Moving cylinder side of fixed cylinder ........................................................................................35
Figure 36: Return signal from fixed and moving cylinder in an anechoic chamber.......................................36
Figure 37: Two cylinders (one fixed and other moving in same radius from the radar)................................37
Figure 38: UWB return signal from two targets (one fixed and another moving in same radius)..................37
Figure 39: Moving human furthest from fixed human.................................................................................39
Figure 40: Moving human little closer from fixed human ............................................................................39
6. 5
Figure 41: Moving human closer from fixed human....................................................................................39
Figure 42: Moving human side of fixed human ...........................................................................................39
Figure 43: Moving human furthest from fixed human and then fixed human removed...............................40
Figure 44: Moving human little closer from fixed human and then fixed human removed ..........................40
Figure 45: Moving human closer from fixed human and fixed human removed thereafter..........................41
Figure 46: Moving human by side of fixed human and fixed human removed thereafter ............................41
Figure 47: Two humans standing one fixed and other moving.....................................................................42
Figure 48: Return signal from two humans .................................................................................................42
Figure 49 : Attenuation of electromagnetic wave within different layers of human skin .............................44
Figure 50: Wave reflection and transmission between two media for normal incidence .............................46
Figure 51: Perpendicular polarized wave incident at oblique angle of incidence between two media .........48
Figure 52 : Four layer model.......................................................................................................................50
Figure 53 : Net reflection coefficient between skin and fat .........................................................................53
Figure 54 : Net reflection coefficient between air and skin .........................................................................54
Figure 55 : Oblique Incidence between Air and Human body layers............................................................55
Figure 56 : Reflection from each point scatterer .........................................................................................56
Figure 57: Radar Cross Section [19].............................................................................................................57
Figure 58: Human standing model in Feko ..................................................................................................60
Figure 59: Human Head in Feko..................................................................................................................60
Figure 60: Human head with shoulder model in Feko .................................................................................61
Figure 61: Human model in POSTFEKO .......................................................................................................61
Figure 62: RCS of standing human ..............................................................................................................61
Figure 63: Human Walking..........................................................................................................................64
Figure 64: Reflected signal from human walking.........................................................................................64
Figure 65: Variation of Relative Permittivity with frequency [19] ................................................................68
Figure 66: Variation of Loss Tangent with frequency [19]............................................................................68
7. 6
LIST OF TABLES
Table 1: Variations of Intrinsic Impedance Parameters with Frequency ......................................................52
Table 2: Variations of Reflection coefficient with Frequency.......................................................................52
Table 3: Human Physical Dimensions..........................................................................................................63
8. 7
ABSTRACT
Detection and tracking of human subjects have many applications in the field of security and
surveillance. The applications include law enforcements, search-and –rescue missions of natural
calamities and military operations. Many types of surveillance systems, such as thermal, video,
seismic sensors are available, but radar offers a unique advantage compared to other methods of
surveillance. Radar uses electromagnetic waves that can penetrate through obstacles and detect
targets in all weather conditions.
This project involves the simulation of Ultra Wide Band (UWB) signal reflected from a
human subject for the purpose of detection and monitoring of its motion. The short pulse length of
UWB radars has a high range resolution that can resolve parts of a human body. A point scatter
model is used to decompose the human body. In order to investigate the effect of multiple
reflections from cylindrical human body parts, the real radar measurements of cylinders and
humans are performed on a lawn and within an anechoic chamber. Reflection coefficients from skin
and Radar Cross Section (RCS) of human that are the essence of the development of simulator were
derived analytically. From the human motion detailed in BioVision Hierarchical (BVH) data file,
the time domain return signal is simulated using MATLAB. The results show the simulated UWB
signal from an animated walking human.
9. 8
1. INTRODUCTION
Human detection or tracking has many applications in the fields of security and surveillance
for military, intruder detection, or human rescue purposes. The military could use this for hostage
rescue when detection of humans inside building becomes critical. In addition, human detection has
applications for rescue operations during disaster search such as earthquake, flood or fire. There are
a number of technologies that are used to detect and/or track human. Among them, the most
important type is video surveillance. However, computer vision has limited ability to detect humans
in poor visibility conditions. This type of surveillance fails when the targets are inside buildings or
behind the walls.
Acoustic, seismic, infrared and electromagnetic sensors have been used in solving human
detection system problems to track and monitor human activities. For instance, Sarkar [1] used
seismic and infrared sensors for target detection. Zoubir[2] proposed to use cheap acoustic sensors
to reliably detect humans. Varshney and Iyengar [3] solved the problem of human detection using
footstep signals from seismic and acoustic sensors. Also, Watanabe [4] proposed a reliable method
of human detection for visual surveillance systems to distinguish humans from other objects.
Research on human motion is a topic of recent interest due to its potential applications in the
field of security. Early research in this area applied fusion techniques to use the strengths of several
sensors in a complimentary fashion. Milch [5] used video and radar for determining human targets.
Many sensor fusion systems with radar were used to determine a target as a human. Augmentation
of radar with other types of sensors in certain applications prevents the advantage of radar like its
ability to operate far away from targets and see through walls and other obstacles.
The human detection problem with radar can be divided into two main tasks: detection of
target and determine of whether the detected target is human or not. Radar based human detection
10. 9
systems need signal processing to distinguish human targets from other objects. There has been
research to detect human targets with Doppler radar. Micro-Doppler radar signatures from a human
have been researched to identify and analyze human motion through simulations and measurements
[6-8].In [9], a Doppler radar was used to the extract human walking parameters and proposed a
method for estimating human walking parameters from radar measurements.
The high resolution Ultra-Wide Band (UWB) radar is an emerging technique for target
detection. It provides a complimentary technology for human detection and tracking in through-
wall or poor visibility conditions. Yarovoy [10] used UWB radar to find humans trapped in
buildings by sensing respiratory motions. The spectral variations of the radar return were analyzed
to distinguish breathing and non-breathing targets. The spectral response of breathing targets has
peaks and contains frequencies corresponding to the frequency of respiration. Thus, it has been
applied for human detection problems. Moreover, research has been done with UWB radar to track
multiple human with CLEAN Algorithm [11] and classification of common shuman activities [12].
In this project, a UWB simulator was developed to emulate human walking. In [8], Dorp and
Groen presented a human walking model to calculate the radar response. Based on this model, a
point scatter model of human was developed in this project. Human walking styles have been
developed in Poser software and converted into BVH data format, which was used to animate
realistic human motion using MATLAB. The point scatter model was used to segment the human
skeleton. A UWB radar was used to collect data for different scenarios with human mannequin to
validate the point scatter model. Once the point scatter model was validated, effect of multiple
reflections was investigated with cylinders and real human. The data was analyzed and the results
are presented in this report. Parameters used for calculation of reflection coefficient of human skin
are presented in Appendix A.
11. 10
2. BACKGROUND
2.1. RADAR
The term „RADAR‟ was developed in 1940 as an acronym for RAdio Detection And Ranging.
The applications of radar are detection, tracking and determining the range, altitude and direction of
objects. Today, usage of radar is diverse. For example, it is commonly used for air traffic control, as
well as ocean and outer space surveillance systems.
Radar is an electromagnetic system used to detect and locate objects or target by analyzing the
echo signal reflected off the target. The distance of a target from the radar can be estimated by
noting the time taken by the signal to travel to the target and return. Radar which has both the
transmitter and receiver at the same place is known as mono-static radar, whereas, if the transmitter
and receiver are placed at different locations, it is called bi-static radar. Figure 1 and Figure 2
present the two types of radar system.
Figure 1: Mono-static radar [13]
12. 11
Figure 2: Bi-static radar [13]
The reflected signal is different from the transmitted waveform in three ways: amplitude,
frequency and time delay. Attenuation of the radar signal happens because of space losses, partial
reflection of the transmitted energy, and atmospheric effects. Radar uses frequency shift and time
delay of the reflected signal to calculate the radial velocity and distance of the target.
Radar has advantages over the other types of sensors. The transmitted radar signal is not
affected by light and dark and thus, radar is used where optical methods of surveillance are limited
due to environmental hazards such as haze, fog or at night. Smoke and fog reduce the radar signal
minimally. Diverse technologies have been introduced like computer vision, infrared detectors,
Laser Detection and Ranging (LADAR) and radar to monitor human subjects. Among them, radar
has the unique advantage of penetrating through obstacles and detecting in all weather conditions.
Hence, radar can be used in situations where sensors have low performance. If radar can provide
more information about the human activities, then it can increase situational awareness in border
patrols to optimize usage of weapons and law enforcements. Radar only has a disadvantage in that
its measurements cannot be directly displayed.
13. 12
The radars can be classified on the basis of the signal transmitted. Continuous-wave (CW)
radar, the simplest type of radar, transmits a sine wave at a fixed frequency. Hand-held device used
by police to determine speed of cars is an ideal example of CW radar, whereas, a pulsed radar
transmits a chirp signal for a certain duration followed by no transmission for an interval of time. In
short, radar can be classified as:
Figure 3: Classification of Radar
14. 13
2.2. ULTRA WIDE BAND RADAR
Ultra Wide Band (UWB) technology is an emerging technique with unique features that make it
attractive for use in many fields such as radar applications, transmission and positioning systems.
Many studies have recommended the deployment of UWB radar in medical imaging [14]. UWB
radar has high spatial resolution as a result of employment of narrow pulses. The informational
content of the UWB radar increases because of the smaller pulse volume of the signal. Figure 4
shows an example of a reflected UWB signal from a target with multiple scattering centers. The
spectrum occupancy of a signal transmitter by UWB radar is at least 25% of the center frequency.
In 1990, Defense Advanced Research Projects Agency (DARPA) under US Department of Defense
defined the fractional frequency band [15]
Fractional Bandwidth
lowerupper
lowerupper
ff
ff
2 Equation 1
Figure 4: UWB Signal [16]
15. 14
Figure 5: Power Spectrum of UWB [17]
In 2002, FCC in the United States allocated 3.1 to 10.6 GHz spectrum for the use of UWB. The
reduced signal length of UWB radar offers the following advantages:
Range measurement accuracy of the detected target. This improves the radar resolution.
Identification of target classes and types because the received signal carries the information
about the target as a whole and also about its separate elements.
Reduction of passive interference created by rain, mist, or aerosols because the scattering
cross section of interference source within a small pulse volume is reduced relative to the
target scattering cross section.
Increase stability and probability of target detection because oscillations reflected from the
individual parts of the target do not interfere and cancel, which provides a more uniform
radar cross section.
16. 15
Improved stability when observing targets at low elevation angles because the main return
signal from a target, and any ground return signal, arrives at the antenna at different times,
which enables to separate.
Low transmission energy as result of short pulse.
Low probability of intercept because UWB signals are transmitted below noise level and
virtually hard to detect.
Increased radar‟s robustness from external electromagnetic radiation effects and noise.
UWB radar provides a complementary human detection and tracking technology whose
performance is better than other sensing technologies. It provides range information of a target
within the limits of frequency allocation, because it transmits a low-powered pulse. In addition, its
short pulse can contain information about a human posture.
17. 16
3. DEVELOPMENT of SIMULATOR
3.1. Poser Software
Poser is a Computer Graphics Interface for animation and rendering of 3D polymesh human
figure. Poser comes with a large library of pre-built human figures including poses, facial
expressions, and walking styles that one needs to create 3D animation. The Poser Library includes
animated and static poses of daily human activities like walking, dancing, standing and sitting. The
content of the human walking model in 3D format was generated in a .bvh file and the information
from the file was imported in MATLAB to design the realistic human walking model. This model
was developed to get the detailed human hierarchical skeleton information. Figure 6 shows a human
walking style that was developed in Poser.
Figure 6: Human Walking Style in Poser
18. 17
3.2. BVH Format
The BVH file provides information of the motion data in addition to the skeleton
hierarchical information. The name „BVH‟ stands for BioVision Hierarchical data. The BVH file
format was developed by a motion capture services company, called Biovision,. The BVH file
consists of two sections:
1. Header section – Provides information about the hierarchy and initial pose of the skeleton
2. Data section – Provides the motion data
The header section begins with the keyword „HIERARCHY‟. The following line starts with
the keyword „ROOT‟ followed by the root segment name of the hierarchy that needs to be defined.
A BVH file can contain any number of skeleton hierarchies. Each segment of the hierarchy contains
some data relevant to that segment and thereafter it recursively defines its subsequent sections. The
information of a segment is the offset of that segment from its parent or in the case of the root
object the offset is considered to be zero. The offset is denoted by the keyword „OFFSET‟ followed
by the X, Y and Z offset of the segment from its parent.
The line following the offset contains header information for the channels. These are
represented by the keyword „CHANNELS‟ followed by a number and a list of labels indicating the
number of channels and type of each channel respectively.
There can be either of the two keywords „JOINT‟ or „ENDSITE‟ on the line of data
following channel specification. A joint definition is the same as the root definition except for the
number of channels. The end site information indicates the current segment has no children. The
following line „}‟ denotes the end of any joint or end site [18]. Figure 7 shows a data section of a
BVH file.
19. 18
The second section of the BVH file i.e. the motion section starts on a separate line with the
keyword „MOTION‟. The subsequent lines with keyword „FRAMES: (a number)‟ and „FRAME
TIME‟ indicates the number of frames or motion samples in the file and the sampling rate of the
data respectively. Figure 8 shows a sample of motion section of a .bvh file.
Figure 7: Header Section
Figure 8: Data Section
20. 19
3.3. Point Scatter Model of Human
The human walking model has been particularly developed for realistic animations in virtual
reality. Computation of UWB return signal on the basis of animated human walking is the goal of
this project. This requires detailed information about the time-varying human body parts.
The human skeleton is segmented into different body parts. In Figure 9 these body parts are
denoted by the space between the two blue points. Each of these parts is considered to be
cylindrical. The cylinder has uniform shape and the reflection is considered from the center of each
of the cylindrical body parts and denoted by green points in Figure 9. The human body parts are
connected to each other by means of time-dependent translations and rotations.
Figure 9: Point Scatter Model of Human
21. 20
In developing the simulator, twelve main human body parts are considered, such as: right
and left upper arm, right and left lower arm, right and left upper leg, right and left lower leg, right
and left feet, abdomen and head. The reflection is assumed to happen from the center of the
cylinder. Hence, the full human can be modelled as twelve point scatterers. The return UWB signal
is calculated from each of these point scatterers. It is assumed that there are no multiple reflections
between the point scatterers. To validate the point scatter model and investigate the effect of
multiple reflections from the different scatterers, various real world scenarios with UWB radar were
developed, which are discussed in later sections.
22. 21
4. MEASUREMENTS
In order to validate the human point scatter model and the effect of multiple reflections from
different human body parts, data were collected for various real world scenarios and processed with
MATLAB. The different scenarios are as follows:
Fixed radar with human mannequin in an anechoic chamber
Fixed radar with one cylinder
Fixed radar with two cylinders
Fixed radar with two standing humans
Measurements were performed using the P220 UWB radar manufactured by Time Domain Co.
Ltd provided by Dr. Kim. The radar system consisted of a P220 UWB radar, two horn antennas, a
LAN cable and a computer to record the data. The radar operated in the monostatic mode where the
waveform pulses were transmitted from a single horn antenna and the scattered waveforms were
received by a collocated horn antenna. Horn antennas (A-info, JXTXLB1080-M and
JXTXLB10125), which have high signal power and directivity, were used to transmit and receive
signals from the illuminated targets. The two ports on the P220 UWB radar were used to connect
transmitting and receiving antennas. The gain of the antennas is approximately 10dBi at 4.7 GHz
that is the center frequency of the radar. The radar was controlled by a computer using application
software provided with the equipment and an Ethernet cable was connected to the computer to
record the data digitally. The radar configuration used for measurements is shown in Figure 10.
23. 22
Figure 10: Radar Configuration
For measurements, two cardboard cylinders wrapped with aluminum foil were used as
targets to verify the effect of multiple reflections between two cylinders. In addition, radar return
signal was investigated with two humans standing in front of the UWB radar. The whole setup was
placed in an open space so that there would not be any effects of clutter or other scatterers within
the test area. However, a small amount of reflection from the ground is unavoidable. Some clutter
from ground reflection and direct coupling between the two antennas was visible in the sample
range profile of a received radar signal in Figure 11 which shows the relative amplitude of the
signal with time.
Horn Antenna P220 UWB radar
24. 23
Figure 11: Example of UWB return from a target
The returning waveform consists of complex pulse due to the shape of target. The shape of
the target return waveform depends on the many scattering points of the target that reflect the
transmitted signal with different reflection coefficients from different distances.
The radar return signal can be represented by the following equation:
)(*)(
1
nn
n
TtuCtx
Equation 2
where, x(t) is the received signal, u(t) is the template signal, Cn is the amplitude, Tn is the time
delay, N is the number of the scattering paths. The above equation creates an unique return signal
shape because Cn and Tn depends on the position and configuration of the target.
The data were saved in .scn format on a computer and later read and analyzed with
MATLAB. For each real world scenario, 15 data files were recorded. Within each of these data
files, there were 1,952 samples. A MATLAB code was written to read the data from each .scn file
and plot the return signal from the targets for further investigation.
Clutter Target
25. 24
5. DATA ANALYSIS and VERIFICATION
5.1. Human mannequin standing sidewise with no hands in anechoic chamber
In order to validate the point scatter model, a human mannequin was placed in an anechoic
chamber and wrapped with aluminum foil to get maximum reflection from the target. The first
scenario was tested with human mannequin no hands shown in figure 12. The subsequent scenario
was created with human mannequin left hand only demonstrated figure 13.
Figure 12: Human mannequin with no hands Figure 13: Human mannequin with left hand
The collected data was processed in MATLAB to find the range profile of the target. Figure
14 shows the range profile of the human mannequin for the above two scenarios, where the red
denotes the mannequin without hands and black denotes the mannequin with left hand only.
26. 25
Figure 14: Range profile of human mannequin without any hands and with left hand only
The range profile, demonstrated in figure 14, illustrates that the left hand of the human
mannequin can be visible distinctly. Further investigations were done with human mannequin in
anechoic chamber for other different scenarios.
Left
Hand
27. 26
5.2. Human mannequin standing sidewise with all hands in anechoic chamber
The two scenarios of the human mannequin with all hands and with right hand only are
demonstrated in figure 15 and figure 16.
Figure 15: Human mannequin sidewise with all arms Figure 16: Human mannequin sidewise with right hand
Figure 17: Range profile of human mannequin with all hands and right hand only
From the range profile in figure 17, it was concluded that the reflection from the right hand of
human mannequin was not visible clearly because it was hidden partially by the full human.
28. 27
5.3. Standing human mannequin facing antenna front in anechoic chamber
The two scenarios of human mannequin standing normally and with right leg forward are
demonstrated in figure 18 and figure 19 respectively.
Figure 18: Front human standing Figure 19: Front human standing with one leg forward
Figure 20: Range profile of human standing normally and one leg forward
From the range profile of the above two scenarios in figure 20, it can be observed that the
reflection from the right leg is visible whereas that from the left leg is difficult to identify.
Right
Leg
29. 28
5.4. Different faces of a cylinder fixed at certain distance from the radar
One cylinder wrapped with aluminum foil was placed in an open lawn to ensure that there
would not be any effect from other targets. This scenario is demonstrated in Figure 21
Figure 21: UWB radar with single target (cylinder)
Figure 22: UWB return signal from one cylinder
Radar System
Cylinder
Cylinder
30. 29
The face of the fixed cylinder directed towards the radar was changed four different times.
In Figure 22, the return waveform of four scenarios were represented using different colours. It was
observed that the shape and amplitude of the UWB return waveform remained almost the same
although there was a minute shift in the return signal with each facing of the cylinder. This was
because UWB signals are very sensitive to any slight changes in position and when the facing of the
cylinder was changed there might be a very minute change of distance between the radar and
target.
Thus, it was concluded that the cylinder wrapped with aluminum foil was a good choice of
target for multiple reflections investigation because of its uniform surface.
31. 30
5.5. Two cylinders at same distance, one fixed and other face changes
Two cylinders were placed on a lawn in front of the UWB radar system to ensure that there
would not be any effect from other targets. The scenario is demonstrated in figure 23
Figure 23: UWB radar with two targets (One fixed and one rotating cylinder)
Figure 24: UWB return signal from two targets (one fixed and one rotating cylinder)
Radar System
Fixed Cylinder
Rotating Cylinder
Cylinders
32. 31
The fixed rotating cylinder was placed to the side of fixed static cylinder and rotated four
times at the same distance from the radar. In figure 24, the four different scenarios are shown with
different colors revealing the shape of the return UWB signal remained almost the same but there
was slight difference in the amplitude. There were some effects of multiple reflections but it was
difficult to establish these conclusively. Hence, more scenarios are presented in the following
sections.
In addition, it was observed that although the rotating cylinder was rotated at the same
position, there is a slight shift in the return signal. This may have been because when the cylinder
was rotated at the fixed distance from the radar, there was a very minute change in the distance of
the target from radar‟s point of view and UWB radar is sensitive to this very minute change of the
target distance.
33. 32
5.6. Two cylinders, one fixed and other approach the fixed from back
Figure 25: Moving cylinder furthest from fixed cylinder Figure 26: Moving cylinder little closer from fixed cylinder
Figure 27: Moving cylinder closer from fixed cylinder Figure 28: Moving cylinder by side of fixed cylinder
Fixed CylinderFixed Cylinder
Fixed Cylinder
Fixed Cylinder
Moving Cylinder
Moving Cylinder
Moving Cylinder Moving Cylinder
34. 33
Investigations were done with two cylinders and removing the fixed cylinder for each of the
scenarios shown in figures 25,26,27, and 28. The return signal for the two cylinders is denoted by
red and that with fixed cylinder removed is denoted by blue in figures 29, 30, 31 and 32.
Figure 29: Moving cylinder furthest from fixed cylinder and fixed cylinder removed thereafter
Figure 30: Moving cylinder little closer from fixed cylinder and fixed cylinder removed thereafter
Fixed Cylinder Moving Cylinder
Fixed Cylinder
Moving Cylinder
35. 34
Figure 31: Moving cylinder closer from fixed cylinder and fixed cylinder removed thereafter
Figure 32: Moving cylinder by side of fixed cylinder and fixed cylinder removed thereafter
It was observed that the shape of the return signal from the moving cylinder remained
almost the same when the fixed cylinder was removed. Also, as the cylinder moved closer, it was
not clear whether the delayed returned signal was due to multiple reflections or resulting from other
clutters. For further verifications, more investigations were carried out with cylinders in an
anechoic chamber.
Fixed Cylinder
Moving Cylinder
Fixed Cylinder
Moving Cylinder
36. 35
5.7. Two cylinders, one fixed and other approach the fixed one from back in an
anechoic chamber
Figure 32: Moving cylinder furthest from fixed cylinder Figure 33: Moving cylinder little closer from fixed cylinder
Figure 34: Moving cylinder closer from fixed cylinder Figure 35: Moving cylinder side of fixed cylinder
Fixed Cylinder
Fixed Cylinder
Fixed Cylinder Fixed Cylinder
Moving Cylinder
Moving Cylinder
Moving Cylinder
Moving Cylinder
37. 36
Figure 36: Return signal from fixed and moving cylinder in an anechoic chamber
Figure 36 shows the return signal of the four scenarios shown in figure 32, 33, 34 and 35
denoted by red, blue, green and yellow colors respectively. It was observed that as the moving
cylinder approaches the fixed cylinder, there were effects of multiple reflections. For further
verification, measurements were taken on a lawn with another scenario.
Fixed Cylinder Moving Cylinder
38. 37
5.8. Two cylinders, one fixed and other moving at a fixed radius from radar
Two cylinders were placed at a certain distance from the radar. One cylinder is fixed and the
other cylinder is moved away from fixed cylinder in the same radius from the radar system.
Figure 37: Two cylinders (one fixed and other moving in same radius from the radar)
Figure 38: UWB return signal from two targets (one fixed and another moving in same radius)
Fixed CylinderMoving Cylinder
Radar System
39. 38
Figure 37 shows one of the four scenarios of fixed and moving cylinders (with fixed radial
distance from radar). From figure 38, for the four different scenarios, it was observed that there
were multiple reflections from the two cylinders. The delayed return signal traversed an extra
distance (i.e. the distance between the two cylinders). The distance between the two cylinders were
calculated as the moving cylinder moved away from the fixed cylinder. This distance confirmed by
the extra distance that the return signal travelled as shown in figure 34. This concludes that multiple
reflections happen between the two cylinders in the return signal.
40. 39
5.9. Two standing humans, one fixed and other approach fixed one from back
Figure 39: Moving human furthest from fixed human Figure 40: Moving human little closer from fixed human
Figure 41: Moving human closer from fixed human Figure 42: Moving human side of fixed human
Fixed Human
Fixed Human
Fixed Human Fixed Human
Moving Human
Moving Human
Moving Human Moving Human
41. 40
Figure 43: Moving human furthest from fixed human and then fixed human removed
Figure 44: Moving human little closer from fixed human and then fixed human removed
42. 41
Figure 45: Moving human closer from fixed human and fixed human removed thereafter
Figure 46: Moving human by side of fixed human and fixed human removed thereafter
From the above investigation, it was difficult to conclude the multiple reflections from the real
human. Also, from figure 43, 44, 45, and 46, it was observed that there was a little shift in the return
signal of moving human. This was because due to the fact that the real human is always moving.
43. 42
5.10. Two humans, one fixed and other moving at a fixed radius from radar
Two humans were standing at a certain distance from the radar system, one fixed and other
human moving away from the fixed human in a radial distance from the radar.
Figure 47: Two humans standing one fixed and other moving
Figure 48: Return signal from two humans
Moving Human Fixed Human
Radar System
44. 43
Figure 47 shows one of the four scenarios of fixed and moving humans. In this case, one
human is moving away from fixed human in a radial distance from the radar. In figure 48, for the
four different scenarios, it is observable that there was not much effect of multiple reflections for
the two humans case. Hence, it can be concluded that there would not be any effect of multiple
reflections between the point scatterers of the human model.
Generally, reflection coefficient of human skin is very low. Hence, energy of the reflected
wave from the human skin is small, which reduces the effect multiple reflection for real
human.There was not much effect of multiple reflection because when the electromagnetic wave
incident on the human skin, it penetrates through different layers and attenuates substantially.
Hence, the reflection from skin is low and ignorable and that reduces the possiblities of multiple
reflection for real human case.
45. 44
6. CALCULATIONS
In order to make the simulator realistic, reflection coefficient of human skin and radar cross
section of human were calculated and incorporated in the developed simulator.
6.1. Propagation of Electromagnetic Wave within three layers of human skin
It is necessary to study how the incident electromagnetic wave decays while propagating
through the different human body layers. Many research studies have modelled human body layers
with a good dielectric which has a frequency dependent electric characteristic. The permittivity and
conductivity of the human body tissue depends on the composition and physiological functions of
the tissue. As the electromagnetic wave penetrates through the human skin, it attenuates and
reflection from the inner layer (bone) is small. Figure 49 shows the simulation result of how the
electromagnetic wave decays within 14cm thickness of human body layers (skin, fat, and bone).
Here, the thicknesses of skin, fat and bone are considered to be 4cm, 4cm, and 6cm respectively.
Figure 49 : Attenuation of electromagnetic wave within different layers of human skin
SKIN FAT BONE
46. 45
6.2. Reflections Coefficient between two dielectric media
The propagation of electromagnetic waves is affected by the electromagnetic parameters of
a material like permittivity, magnetic permeability, and conductivity. Reflection coefficient is a
function of the constitutive parameters of the two media, the direction of wave travel (angle of
incidence), and the direction of the electric and magnetic fields (wave polarization). It is a complex
quantity, which is represented by the impedance of two media separated by a interface:
IR
IR
ZZ
ZZ
Rho
Equation 3
where, ZR represents the intrinsic impendence of the second medium and ZI represents the intrinsic
impedance of the first medium.
CASE I: NORMAL INCIDENCE
The wave travel is perpendicular (normal incidence) to the planar interface formed by two
semi-infinite lossless media, each having parameters of ε1,µ1 and ε2,µ2 shown in Figure 50. When
the incident wave encounters the interface, a fraction of the wave intensity will be reflected into
medium 1 and part will be transmitted into medium 2.
47. 46
Figure 50: Wave reflection and transmission between two media for normal incidence
The following expressions for incident, reflected, and transmitted electric field components
can be represented, assuming the incident electric field of amplitude E0is polarized in the x direction
respectively, as
)**exp(**ˆ 10 zjEaE xi Equation 4
)**exp(***ˆ 10 zjEaE b
xr Equation 5
)**exp(***ˆ 20 zjETaE b
xt Equation 6
where, Ӷb
and Tb
represent reflection and transmission coefficients at the interface. Using
Maxwell‟s equations, the magnetic field components corresponding to the above equations are
represented as
)**exp(**ˆ 1
1
zj
E
aH
y
yi
Equation 7
)**exp(**ˆ 1
1
0
zj
E
aH b
yr
Equation 8
)**exp(**ˆ 2
2
0
zj
E
TaH b
yt
Equation 9
48. 47
The reflection and transmission coefficients are determined by enforcing continuity of the
tangential components of the electric and magnetic fields across the interface (z=0) by using the
above equations
bb
T1 Equation 10
21
)1(
bb
T
Equation 11
The above two equations are simplified to
i
r
i
rb
H
H
E
E
12
12
Equation 12
i
tbb
H
H
T
1
2
12
2
1
*2
Equation 13
From the above two equations, it is clear that the reflection and transmission coefficients of
a planar interface for normal incidence are functions of the constitutive properties of the two media
and is independent of the polarization of the incident wave. This is because the electric and
magnetic fields of a normally incident plane wave are always tangential to the boundary
irrespective of the polarization of wave.
49. 48
CASE II: OBLIQUE INCIDENCE
The electric field of the uniform plane wave incident on a planar interface at a nonzero angle
of incidence is oriented perpendicularly to the plane of incidence. This is referred to as
perpendicular polarization as shown in Figure 51.
Figure 51: Perpendicular polarized wave incident at oblique angle of incidence between two media
The incident electric and magnetic fields for the oblique incidence can be represented by the
following equations:
))cos*sin*(*exp(**ˆ 10 iiy
i
zxjEaE Equation 14
))cos*sin*(**exp()sin*ˆcos*ˆ( 1
1
0
iiizix
i
zxj
E
aaH
Equation 15
))cos*sin*(**exp(*0*ˆ 1 rr
b
y
r
zxjETaE Equation 16
))cos*sin*(**exp()sin*ˆcos*ˆ( 1
1
0
rr
b
rzrx
r
zxj
E
aaH
Equation 17
50. 49
))cos*sin*(**exp(***ˆ 20 tt
b
y
t
zxjETaE Equation 18
))cos*sin*(**exp()sin*ˆcos*ˆ( 2
2
0
tt
b
tztx
t
zxj
ET
aaH
Equation 19
By applying the boundary conditions on the continuity of the tangential components of the
electric and magnetic fields and Snell‟s law of reflection and refraction the above equations reduce
to
bb
T 1 Equation 20
btbi
T
21
cos
)1(
cos
Equation 21
Solving for b
and b
T , the above equations reduce to
ti
ti
i
r
b
E
E
cos*cos*
cos*cos*
12
12
Equation 22
ti
i
i
t
b
E
E
T
cos*cos*
cos**2
12
2
Equation 23
The above two terms are usually referred to as the plane wave Fresnel reflection and
transmission coefficients for perpendicular polarization.
51. 50
6.3. Four Layer Model
There has been considerable research on the interaction between the incident
electromagnetic pulse and human body layers. Electric parameters of each body layer play an
important role in the propagation of the electromagnetic waves through human body layers. From
the previous section, it was concluded that the electromagnetic wave decays almost completely
when it propagates through each layer of skin, fat and bone. Hence, the main internal structure of
human body layers is modelled as three layers: skin, fat and bone. Figure 52 shows how the
incident electromagnetic waves reflect and transmit through different layers of skin, fat, and bone.
Figure 52 : Four layer model
Thus, a four layer model of air, skin, fat and bone was used to calculate the net reflection
coefficient between human skin and air.
52. 51
In addition, the permittivity of each human body layer depends on its molecular structure
and is a complex quantity which is expressed as:
'''
* j Equation 24
where '
is the relative permittivity of the biological tissue and ''
is the loss factor denoted as
*0
''
Equation 25
where, is the conductivity , ω is frequency. The conductivity of body layer tissue depends on its
composition and physiological functions.
Koohestani, Pires, Skrivervik, and Moreira show frequency dependent description of
permittivity and loss tangent of each human body layers (skin, fat, bone), which were implemented
in the simulation to calculate net reflection coefficient from human skin in the frequency range of 3
to 12 GHz. The relative permittivity and loss tangent of the materials that have been used to model
human body layers are presented in Appendix A. [19]
Table 1 shows the relative permittivity and loss tangent of skin, fat, and bone from the graph
shown in Appendix A. From this information, wave impedance of each layer was evaluated to
calculate the net reflection coefficient of human skin. Reflection coefficient from human skin is
shown in Table 2.
The impedance of a dielectric medium is calculated by the following expression:
**
**
wj
wj
=
j
w
j
*
*
Equation 26
where,
*w
is the loss tangent and µ is the permeability of the dielectric medium.
53. 52
The following TABLE 1 shows the relative permittivity and loss tangent of skin, fat and bone.
Table 1: Variations of Intrinsic Impedance Parameters with Frequency
FREQUENCY SKIN FAT BONE
(GHz) RELATIVE
PERMITTIVITY
LOSS
TANGENT
RELATIVE
PERMITTIVITY
LOSS
TANGENT
RELATIVE
PERMITTIVITY
LOSS
TANGENT
3 38 0e 0.28 5 0e 0.15 19 0e 0.33
4 37 0e 0.29 5 0e 0.16 18 0e 0.36
5 36 0e 0.30 5 0e 0.17 17 0e 0.39
6 35 0e 0.32 5 0e 0.18 16 0e 0.42
7 34 0e 0.35 5 0e 0,19 15 0e 0.45
8 33.5 0e 0.38 5 0e 0.20 14 0e 0.47
9 32.5 0e 0.42 5 0e 0.21 13 0e 0.50
10 31.5 0e 0.45 5 0e 0.22 12 0e 0.53
11 30.5 0e 0.49 5 0e 0.23 11 0e 0.56
12 30 0e 0.53 5 0e 0.24 10.5 0e 0.58
0e is the permittivity of free space which is equal to 8.85e-12 F.m-1
Table 2: Variations of Reflection coefficient with Frequency
FREQUENCY (GHz) REFLECTION
COEFFICIENT
3 -0.9061-0.0259i
4 -0.9020-0.0283i
5 -0.8976- 0.0308i
6 -0.8933- 0.0344i
7 -0.8893- 0.0391i
8 -0.8861- 0.0431i
9 -0.8822- 0.0492i
10 -0.8770- 0.0550i
11 -0.8721- 0.0620i
12 -0.8711- 0.0668i
54. 53
6.4. Calculation of Reflection Coefficient between Human Skin and Air
Figure 53 : Net reflection coefficient between skin and fat
Ӷ1= partial reflection coefficient of a wave incident on fat, from skin
Ӷ2= partial reflection coefficient of a wave incident on skin, from fat
Ӷ3= partial reflection coefficient of a wave incident on bone, from fat
T1= partial transmission coefficient of a wave incident on skin into fat
T2= partial transmission coefficient of a wave incident on fat into skin
The coefficients can be expressed as
SF
SF
ZZ
ZZ
1 ;
FS
FS
ZZ
ZZ
2 ;
FB
FB
ZZ
ZZ
3 ;
SF
F
ZZ
Z
*2
T1 ;
SF
S
ZZ
Z
*2
T1
The total reflection coefficient from the skin can be expressed as
……………+**T*T-**T*T+*T*T-= 3
3
2
221
2
32
2
213211s
n
n
TT )*(*** 32
0
3211
55. 54
Since |Ӷ3| < 1 and |Ӷ2| < 1, the infinite series can be summed using geometric series result
1||,
1
1
0
xfor
x
xn
n
to give
32
3213211
32
321
1
*1
****
*1
**
TTTT
S Equation 27
Figure 54 : Net reflection coefficient between air and skin
Following the same procedure as in the previous model, the net reflection coefficient
between skin and air is calculated. From the previous model, the reflection coefficient between skin
and fat is already calculated. Thus,
0
0
4
ZZ
ZZ
S
S
;
S
S
ZZ
ZZ
0
0
5 ;
S
s
ZZ
Z
0
3
*2
T ;
SZZ
Z
0
0
4
*2
T
Hence, the net reflection coefficient from human skin is given by the following equation:
S
SS
S
S TTTT
*1
****
*1
**
5
54544
5
43
4 Equation 28
56. 55
6.5. Calculation of Reflection Coefficient considering Fresnel’s Angle of Incidence
For the oblique angle of incidence, the reflection coefficient between air and skin is different
from that in the normal incidence case. Hence, incident and transmitted angles are calculated for
each of the twelve scattering points of human. Figure 55 shows an oblique angle of incidence on the
human skin from air and how the electromagnetic wave gets reflected and transmitted through
different layers.
Figure 55 : Oblique Incidence between Air and Human body layers
For calculation of reflection coefficient using Fresnel‟s equation, it is required to calculate
the total intrinsic impedance of the three layers (skin, fat and bone). The combined intrinsic
impedance of the three layers (skin, fat, and bone) is calculated using Equation 12, where is the
net reflection coefficient between human skin and air calculated in the previous section by Equation
28, 1 and 2 denotes the intrinsic impedance of air and combined layers (skin, fat, and bone)
respectively.
57. 56
Figure 56 shows reflections from each of the twelve scattering points of the human model.
Figure 56 : Reflection from each point scatterer
The following steps were used to calculate reflection coefficient for oblique incidence:
1. Angle of incidence was calculated from each of the twelve scattering points from the fixed
radar position.
2. Snell‟s Law was applied to calculate the transmitted angle between air and three layers
(skin, fat, and bone).
3. Using equation 22, reflection coefficients from each of the twelve scattering points were
calculated.
58. 57
6.6. Radar Cross Section (RCS)
Electromagnetic waves are diffracted or scattered in all directions when incident on a target.
The scattered waves are of two parts: the first part is made of waves that have same polarization as
the receiving antenna and the other portion of the scattered waves have a different polarization to
which the antenna does not respond. Radar cross section is the measure of a target‟s ability to
reflect radar signals towards the direction of radar receiver. The intensity of the back scattered
energy that has the same polarization as the radar‟s receiving antenna is used to define the target
RCS. Target RCS is a measure of the ratio of backscatter power per unit solid angle in the direction
of the radar from the target to the power density that is intercepted by the target. A target‟s RCS (σ)
is easily defined as the product of three factors:
𝑹𝑪𝑺 𝝈 = 𝐏𝐫𝐨𝐣𝐞𝐜𝐭𝐞𝐝𝐜𝐫𝐨𝐬𝐬 𝐬𝐞𝐜𝐭𝐢𝐨𝐧 ∗ 𝐑𝐞𝐟𝐥𝐞𝐜𝐭𝐢𝐯𝐢𝐭𝐲 ∗ 𝐃𝐢𝐫𝐞𝐜𝐭𝐢𝐯𝐢𝐭𝐲 Equation 29
where reflectivity is the percent of intercepted power scattered by the target, directivity is the ratio
of the back scattered power in the direction of radar to the backscattered power of an isotropic
scatterer.
Figure 57: Radar Cross Section [19]
59. 58
Assuming PDi is the power density of an incident wave located at a range R away from the
radar, PDr is power density of the scattered wave at the receiving antenna. The amount of reflected
power from the target is
Dir PP * Equation 30
where denotes the target cross section.
Also,
)**4/( 2
RPP rDr Equation 31
Combining the above equations, we get
)(***4 2
Di
Dr
P
P
R Equation 32
In order to ensure that the radar receiving antenna is in the far field (i.e., scattered waves
received by the antenna are planar), the above equation is modified to give monostatic RCS:
)(lim***4 2
Di
Dr
R P
P
R
Equation 33
60. 59
6.7. FEKO Software
Humans RCS is difficult to ascertain as it depends on person‟s posture, relative orientation,
and incident angle of the wave. Even the type of person‟s clothing can affect the reflectivity of the
human body parts and observed RCS. FEKO is electromagnetic simulation software that is used for
scattering analysis of metallic and dielectric bodies of different size for calculating radar cross
section. It is a Method of Moments (MOM) tool that is used to calculate the radiation pattern,
impedance and gain of an antenna while mounted on some defined geometry. In addition, it can
calculate the near field around an antenna and the electric currents that flow on an antenna or the
surrounding structure.
The multiple solution techniques available within Feko make it applicable to a wide range of
applications that include:
1. Analysis of multiple dielectric layers in a large structure
2. Scattering problems: RCS analysis of large and small structures.
3. Antenna placement: analysis of antenna radiation patterns, radiation hazard zones with an
antenna placed on a large structure, e.g. ship, aircraft, armoured car
The basic flow of performing a FEKO analysis consists of:
1. Building a geometry, for the antenna in CADFEKO
2. Building a geometry to represent surrounding geometry in CADFEKO
3. Meshing the created antenna and surrounding geometries.
4. Requesting solution types and setting solution parameters.
5. Running the FEKO solver.
6. Read in and interpret results using POSTFEKO
61. 60
The FEKO library provides a full human standing model, a model of the human head, a model
of human head with shoulder as shown in Figure 58, Figure 59 and Figure 60. These models was be
used for calculation of human RCS.
Figure 58: Human standing model in Feko
Figure 59: Human Head in Feko
62. 61
Figure 60: Human head with shoulder model in Feko
After setting the desired solution parameters, the FEKO solver was run. However, in the
desired frequency range, CADFEKO failed to run FEKO solver completely because of limited
available memory. It requires huge memory allocation for solving in high frequency range. After
many trials, it was found that FEKO can solve up to 500MHz frequency for the available memory
in computer used for simulation.
Figure 61: Human model in POSTFEKO Figure 62: RCS of standing human
63. 62
6.8.Calculation of Human RCS
Humans are complicated targets because of the intricate motion of body parts moving along
different trajectories at different speeds. Over the years research has been done to analyze and
mathematically model human walking [10]. In 2007, researchers from the Sensors and Electronics
Directorate of the Army Research Laboratory simulated human body radar signature and analyzed
the human RCS for different configurations as frequency, functions of aspect angle, and
polarization [21]. Some of the important conclusions based on their simulations are:
There is a strong return from the back, as the body torso is approximately flat, creating a
larger RCS compared to other curved surfaces of the body.
Angular RCS variation is more for higher transmitted radar frequency.
Body shape does not have a major influence on average RCS, although there is a
dependence on overall body size.
To date, researchers have used many probabilistic approximations to RCS. Geisheimer [22]
predicted the RCS of each body part according to a percentage of overall surface area. Depending
on size, shape, density and incident angle, the reflectivity changes. Hence, for our simulator, we
used the model developed by Van Dorp [8], which considers most of the parameters for RCS
calculations. Each body part is approximated as a sphere or a cylinder in shape. The analytical
expression for calculating RCS takes into account the radius and height of the cylindrical body
parts, broadside incident angle. For a cylinder with diameter Rc, height Hc, and angle of incidence
(θc), RCS (σ) is given by:
4
cos)
*2
()
*2
(
***)sin
**2
(sin*
**2
cos**
eee
H
c
R
Hi
ccRiri
c
cc
cc
Equation 34
64. 63
The following Table 3 shows the dimensions of each body part considered for the development of
the simulator. These dimensions were calculated from the human mannequin.
Table 3: Human Physical Dimensions
Human Body Parts Diameter (cm) Height (cm)
Right Feet 7.5 22
Left Feet 7.5 22
Upper Left Leg 12 42
Lower Left Leg 10 53
Upper Right Leg 12 42
Lower Right Leg 10 53
Left Upper Arm 10 30
Right Upper Arm 10 30
Left Lower Arm 8 42
Right Lower Arm 8 42
Chest 28 57
Head 14 25
66. 65
8. CONCLUSION
UWB radar simulator was developed to emulate the human walking. The developed point
scatter model of human was verified with the range profile. It can be concluded from the
measurement of range profile that reflections are from each human body parts. The multiple
reflections from different point scatterers of human were investigated through the measurements.
Initially, measurements were done with cylinders in lawn for a clutter free environment. From the
measurements, it was concluded that multiple reflections exists between cylinders when they are
located closely. On the other hand, multiple reflections from human when measured in lawn were
not observed because its magnitude is very small and ignorable. This was due to the material
property of human skin. The reflection coefficient from human skin was derived analytically using
a four layer model (air, skin, fat, and bone). In addition, human radar cross section was calculated
from the analytical radar cross section equation of cylinder. The physical dimensions required for
the calculation are obtained from the human mannequin. From the human motion detailed in BVH
file, the time domain reflected signal from the point scatterers was simulated from an animated
human walking using MATLAB.
Future Work
For the verification of the developed simulator, the statistical distribution of simulated UWB
signal can be investigated as a future research. It is difficult to get exactly the same return signal
from the real walking human and that from the animated human motion because the size and exact
posture of a real human are different from those of the animated human. As an alternative way, we
can compare the statistical distribution of UWB signal to investigate their own signatures for the
classification of human motions.
67. 66
REFERENCES
1. Sarkar, S.; Ray, A; Jin, X; Darmala, T.; , “Target Detection and Classification Using
Seismic and PIR Sensors,” Sensors Journal, IEEE, vol.12, No.6, June 2012
2. Zoubir,A. M; Moebus, M; Viberg, M.; “Parametrization of acoustic images for detection of
human presence by mobile platforms”, Acoustics Speech and Signal Processing, IEEE
International Conference 2010, pp. 3538-3541, 14-19 March 2010.
3. Iyengar, S.G; Varshney,P.K.; Darmala, T. ;, “On the Detection of Footsteps Based on
Acoustic and Seismic Sensing”, Signals, systems and Computers, ACSSC 2007 Conference
Record of the Forty-First Asilomar Conference on, pp. 2248-2252. 4-7 Nov. 2007.
4. Watanabe, T.; Shimosakoda, Y.; Kuno, Y.; Nakagawa, S.;, “Automated Detection of Human
for Visual Surveillance System,” Pattern Recognition, 1996., Proceedings of the 13th
International Conference on, Vol.3, no., pp.865-869 vol.3, 25-29 Aug 1996
5. Milch,S., and Behrens, M.,: “Pedestrian detection with radar and computer vision,”
Proceedings of Conference on Progress in Automobile Lighting, Darmstadt, Germany,
2001.
6. Geisheimer, J.L.; Grenekar, E.F.; Marshall, W.S.;, “High- Resolution Doppler model of the
human gait”. SPIE Proc. Radar Sensor Technology and Data Visualization, July 2002, Vol.
4774, pp. 8-18
7. Thayaparan, T.; Abrol, S.; Riseborough, E.; Stankovic, L.; Lamothe, D.; Duff,G.;, “Analysis
of radar micro-Dopller signatures from experimental helicopter and human data”, IET Radar
Sonar Navig., 2007,1, pp. 289-299
8. Anderson, M.; Rogers, R.; “Micro-Doppler analysis of multiple frequency continuous wave
radar signatures”. SPIE Proc. Radar Sensor Technology, May 2007, vol. 6547
9. Dorp, P. Van; Groen, F.C.A.; “Human Walking Estimation with Radar” , IEE, IEE
Proceedings online no. 20030568, March 2003
10. Yarovoy, A.G., Ligthart, L.P., Matuzas,J., and Levitas, B., “UWB radar for human being
detection”, IEEE Aerospace and Electronics Systems Magazine, Vol. 21, Issue 3, pp. 10-14,
March 2006
68. 67
11. Kim, Y.; Singh, J.; Kwon, J.; Lee, N.;, “Application of Ultra Wide Band Radar for Multiple
Human Tracking withCLEANAlgorithm”,http://ap-s.ei.tuat.ac.jp/isapx/2011/pdf/[ThG3-
]%20D03_1002.pdf
12. Bryan, J.; Kwon, J.; Lee, N.; Kim, Y.; “Applications for Ultra- Wide Band Radar for
classification of human activities”, IET Radar, Sonar and Navigation, September 2011
13. http://www.answers.com/topic/monostatic-radar
14. Taoufik, E.; Nabila, S.; Ridha, B.; “The Reflection of Electromagnetic Field by Body Tisse
in the UWB Frequency Range”, IEEE International Radar Conference, Mai, 2010, pp. 1403-
1407.
15. Ultra-Wideband Radio Systems. Their Peculiarities and Capabilities, Electromagnetic
Phenomena, V.7, •1 (18), 2007
16. Ultra Wide Band Radar Technology- James D. Taylor
17. Introduction to Ultra Wideband (UWB); Ou Yang, WCNG @ UR, 4/2/2007
18. M. Meredith, S. Maddock, “Motion Capture File Formats Explained,” Department of
Computer Science, University of Sheffield, Sheffield URL: http://www.dcs.shef.ac.uk/~
{mikem, steve}
19. Koohestani, M.; Pires, N.; Skrivervik, A.; Moreira, A., “Influence of the Human Body on a
New Coplanar-fed Ultra- Wideband Antenna”, Antennas and Propagation (EUCAP), 6th
European Conference, 2012, pp. 316-319
20. http://www.tscm.com/rcs.pdf
21. Dogaru, T. , Nguyen, L. and Le, C., “Computer Models of the Human Body Signature for Sensing
Through the Wall Radar Applications,” U.S. Army Research Laboratory Technical Report ARL-
TR-4290, September 2007
22. Geisheimer, J.L., Marshall, W.S., and Greneker, E., “A continuous-wave (CW) radar for gait
analysis,” Signals, Systems, and Computers, Vol. 1, pp. 834-838, 2001
69. 68
APPENDIX A
Figure 65: Variation of Relative Permittivity with frequency [19]
Figure 66: Variation of Loss Tangent with frequency [19]