Document text
FM-CW Based Miniature SAR Systems for Small UAVs
Prof. Peter Hoogeboom1'2, Jacco de Wit2, Adriano Meta2, Jordi Figueras2
'TNO Physics and Electronics laboratory
PO-box 96864, 2509 JG The Hague, The Netherlands
Phone: +31 70 3740041, Fax: +31 70 3740654
and
international Research Centre on Telecommunications-transmission and Radar
Delft University of Technology, Department of Electrical Engineering
E-mail: [email protected]
ABSTRACT
In some earth obsen’ation applications there is a requirement for low cost, high performance imaging
radar systems small enough to be operated from small, even unmanned, aircraft. The existing pulse radars
are non-cost effective complex systems. Therefore they are not suited for these applications. Frequency
Modulated Continuous Wave (FM-CW) radar systems are on the other hand generally very > compact and
relatively cheap to purchase and to use. IRCTR, together with TNO, is investigating the feasibility of
combining FM-CW technology > and high resolution Synthetic Aperture Radar (SAR) techniques. This
research has led to the development of two different systems.
An FM-CW millimetre wave SAR sensor (35 GHz) was built and is being tested. The system is battery
powered and uses a PC to control the radar and store the data, including GPS and data from motion
sensors. The system is mounted in a pod attached to a light motor glider. Furthermore, a detailed system
model has been developed in order to estimate and analyze the performance of the demonstrator system.
The measurements made with the demonstrator system are evaluated and the findings are used to improve
the system model. This improved model will be a strong aid in the design of future higher performance
FM-CW SAR systems.
The second development regards a miniature P-band (450 MHz) polarimetric FM-CW SAR. This system is
intended for use with an ultra-light aircraft. It is designed for tropical forest monitoring studies but could
be used in foliage penetration applications as well. This system will be battery’ powered and uses a small
memory stick to store the raw data. Data processing is now foreseen as an off-line activity. The design of
the system is still under study and the realization is expected to start in the second half of 2005. This
system will employ a Direct Digital Synthesizer to generate the waveform that is transmitted at P-band.
Two separate dual polarized patch antennas will be used for transmit and receive.
1.0 INTRODUCTION
Some airborne earth observation applications require low cost, small imaging radar systems of high
performance. Such systems should be suited for operation from very small, possibly even unmanned,
aircraft. Coherent pulse radars are usually complex systems being neither compact nor cost effective.
FM-CW radar systems are on the other hand generally very compact and relatively cheap to purchase and
to use. IRCTR investigates together with TNO the feasibility of FM-CW SAR in the field of airborne earth
observation. The combination of FM-CW technology and high resolution SAR techniques has led to the
development of several small, cost effective SAR systems that can be used on small platforms. SAR is a
novel application for FM-CW radar systems. Two developments are ongoing:
Hoogeboom, P.; de Wit, J.; Meta, A.; Figueras, J. (2005) FM-CW Based Miniature SAR Systems for Small UAVs. In Advanced Sensor
Payloads for UAV ( pp. 26-1 - 26-12). Meeting Proceedings RTO-MP-SET-092, Paper 26. Neuilly-sur-Seine, France: RTO. Available
from: http://www.rto.nato.int/abstracts.asp.
RTO-MP-SET-092
26-1
UNCLASSIFIED/UNLIMITED
Report Documentation Page
Form Approved
OMB No. 0704-0188
Public reporting burden for the collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and
maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information,
including suggestions for reducing this burden, to Washington Headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson Davis Highway, Suite 1204, Arlington
VA 22202-4302. Respondents should be aware that notwithstanding any other provision of law, no person shall be subject to a penalty for failing to comply with a collection of information if it
does not display a currently valid OMB control number.
1 . REPORT DATE 2. REPORT TYPE
01 MAY 2005 N/A
3. DATES COVERED
4. TITLE AND SUBTITLE
FM-CW Based Miniature SAR Systems for Small UAYs
5a. CONTRACT NUMBER
5b. GRANT NUMBER
5c. PROGRAM ELEMENT NUMBER
6. AUTHOR(S)
5d. PROJECT NUMBER
5e. TASK NUMBER
5f. WORK UNIT NUMBER
7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES)
TNO Physics and Electronics laboratory PO-box 96864, 2509 JG The
Hague, The Netherlands
8. PERFORMING ORGANIZATION
REPORT NUMBER
9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES)
10. SPONSOR/MONITOR'S ACRONYM(S)
11. SPONSOR/MONITOR'S REPORT
NUMBER(S)
12. DISTRIBUTION/AVAILABILITY STATEMENT
Approved for public release, distribution unlimited
13. SUPPLEMENTARY NOTES
See also ADM202032., The original document contains color images.
14. ABSTRACT
15. SUBJECT TERMS
16. SECURITY CLASSIFICATION OF: 17. LIMITATION OF
18. NUMBER 19a. NAME OF
a. REPORT b. ABSTRACT c. THIS PAGE |J|J
unclassified unclassified unclassified
12
Standard Form 298 (Rev. 8-98)
Prescribed by ANSI Std Z39-18
UNCLASSIFIED/UNLIMITED
FM-CW Based Miniature SAR Systems for Small UAVs
ORGANIZATION
An FM-CW millimetre wave SAR sensor (35 GFIz) was built and is being tested. Within the frame of the
project a fully operational airborne demonstrator has been developed. This system is mounted in a pod,
is battery powered and uses a PC to control the radar and store the data, including GPS and data from
motion sensors. The radar unit was originally designed for collision avoidance radars and is very small and
of light weight, less than 1 kg. The range is limited to less than 1 km, which is sufficient for the
demonstration of the concept. A Stemme light motor glider is available as a low cost test platform.
Furthermore, a detailed system model has been developed in order to estimate and analyze the
performance of the demonstrator system. The measurements made with the demonstrator system proved
the validity of the FM-CW SAR algorithms although the poor performance of the sensor due to the non¬
linearity of the sweep and the high noise level prevented obtaining a good SAR image.
Based on the experience and know-how obtained with the FM-CW millimetre wave SAR sensor a second
development is currently ongoing in IRCTR. The second development regards a miniature P-band
polarimetric FM-CW SAR. This system is intended for use with an ultra-light aircraft. It is designed for
tropical forest monitoring studies carried out by the Wageningen University (WUR) in sites managed by
the Borneo Orangutan Survival Foundation (BOS) in Indonesia but could be used in foliage penetration
applications as well. This system will be battery powered and uses a small memory stick like data
recording device to store the raw data. Data processing is now foreseen as an off-line activity. The design
of the system is still under study and the realization is expected to start in the second half of 2005.
This system will employ a Direct Digital Synthesizer to generate the waveform that is transmitted at P-
band. Two separate dual polarized patch antennas will be used for transmit and receive. In the paper the
design will be discussed in detail.
Section 2 of this paper exposes the principles of FM-CW radars. Section 3 gives an overview of the SAR
principles and discusses the differences between pulse and FM-CW SAR signal processing. In section 4
the platforms used in both High resolution FM-CW SAR and P-SAR projects are described. Section 5
describes the high resolution FM-CW SAR project. A description of the system is provided and the results
of some airborne tests done so far are commented. Section 6 describes the P-SAR project. Section 6.1
includes an overview of several applications that P-band systems have and the requirements for the actual
application of the system; tropical forest monitoring. In section 6.2 the system is described while in
section 6.3 the time schedule of the complete P-SAR project is commented.
2.0 FM-CW RADAR PRINCIPLE
In FM-CW radar the transmitted frequency is a function of time, widely used modulation schemes are
sawtooth and triangular modulations. Assuming the transmitted frequency to be sawtooth modulated,
as shown in Figure la, the ideal transmitted signal can be written as:
where f, is the carrier frequency a = A f /At is the chirp rate. At is the sweep repetition interval and Af is
the frequency sweep. tn = nAt is the slow time variable, as opposed to the fast time variable t. Amplitude
values are neglected. The received signal from a point scatterer is the transmitted signal delayed in time by
the round trip propagation time to the scatterer and back, r = 2 R/c , where c is the speed of light and R is
the distance between the radar and the scatterer. After mixing, the intermediate signal is obtained:
26-2
RTO-MP-SET-092
UNCLASSIFIED/UNLIMITED
UNCLASSIFIED/UNLIMITED
FM-CW Based Miniature SAR Systems for Small UAVs
j2x\ f„T- a i \-az{l t )
(t) = e ^ 2 ^ rect
At
(2)
As can be seen in (2) the resultant beat frequency fb = az is proportional to range. It is assumed here that
the target is stationary. If the target is moving, an additional Doppler frequency shift / is superimposed on
the beat frequency.
(a)
(b)
Figure 1: (a) The Instantaneous Frequency of the Transmitted and Received Sawtooth Modulated
Signals, (b) The Resultant Beat Frequency after Mixing the Transmitted and Received Signals.
Fig. lb shows the beat frequency corresponding to the transmitted and received signals of Fig. la. From
the figure it can be seen that the beat frequency is positive on one part of the modulation cycle and
negative on the other part. It is common practice to use only the positive beat frequencies to obtain the
range information. Note that the range resolution will deteriorate if only part of the sweep is used to
determine the range. However, generally the part of the sweep with negative beat frequencies is very
small.
A well-known problem in FM-CW radars is the leakage signal arriving at the receiver via the direct
coupling between the transmitting and receiving antennas. The noise side bands of the leakage signal, the
phase noise, could mask echo signals, if the antenna isolation is insufficient, or the phase noise of the local
oscillator is too high. The slope of the phase noise can be approximated by 1//1], where /is the offset
frequency from the carrier. In FM-CW radars this offset frequency is proportional to range therefore the
phase noise is proportional to HR. The received signal, on the other hand, is proportional to HR3 [2].
Therefore, the phase noise from a strong echo at short range can mask long range weak scattering.
3.0 FM-CW SAR PRINCIPLE
Opposite to pulse radars, where the raw signals are given in the time domain, the beat frequency is the
starting point of the following processing algorithms and therefore signals are given in the frequency
domain. The SAR signal processing starts from the mixed signal given by equation (2). In pulse radars a
short pulse is transmitted. Therefore the movement of the aircraft within the transmission of the pulse is
negligible. In this case the stop-and-go approximation, i.e. the assumption that the platform stops,
RTO-MP-SET-092
26-3
UNCLASSIFIED/UNLIMITED
UNCLASSIFIED/UNLIMITED
FM-CW Based Miniature SAR Systems for Small UAVs
ORGANIZATION
the radar transmits and receives a pulse, and the platform moves to the next position can be applied. Since
in general in FM-CW radars the duration of the sweep is much larger this assumption is not valid
anymore, especially for high speed platforms.
Assuming that the antenna is pointing perpendicular to the flight path, the distance between the antenna
and a scatterer placed at an azimuth position x and range position r is a function of time and is given by:
R(t’t«) = 'lr2+(v{tn+t) ~x)~ (3)
where v is the velocity of the platform. Considering, for simplicity, that x=0 and assuming that the range
position is much further than the azimuth position, x, the round trip delay between the sensor and the
scatterer, can be written as:
_ 2r | v2(tn+t)2
c rc
In order to gives and insight on the phase terms conforming the received signal (4) is substituted in (2).
The term r is usually negligible or can be corrected by residual video phase remove [3]. If it is suppressed
in (2), the following expression is obtained:
jin
2 r f„V
-rJ+ 2
JoV
t+2-
fo V2
(5)
The objective of the SAR processing is to compress the scatterer response in both azimuth and range
directions. To do so a two-dimensional Fourier analysis with phase correction must be carried out. The
information about range position is carried by the phase term 2 ratjc , while the time variation of the
Doppler frequency given by f0v2t2jrc provides the azimuth position. Note that the fact that the stop-and-
go approximation is not valid for FM-CW SAR systems introduces more phase terms that have to be
corrected.
After the SAR processing the azimuth resolution is given by [3]:
Ax =
(6)
where fomax is the maximum Doppler frequency of the scatterer. The maximum Doppler frequency is given
by:
(7)
where 0az is the angular azimuth beamwidth and is the wavelength. If (7) is substituted in (6) it follows:
Ax =
(8)
This result leads to two interesting conclusions about synthetic aperture radars. The first one is that the
resolution is inversely proportional to the azimuth beamwidth. A wider beamwidth (and therefore a
26-4
RTO-MP-SET-092
UNCLASSIFIED/UNLIMITED
UNCLASSIFIED/UNLIMITED
FM-CW Based Miniature SAR Systems for Small UAVs
smaller antenna) allows a better azimuth resolution. The second one is that the resolution is not anymore
dependant on the range. That is the reason why synthetic aperture techniques are increasingly popular in
imaging radar applications.
In this analysis differences between pulse and FM-CW SAR processing for the ideal case where the
platform is moving at a steady velocity and the antenna is pointing perpendicular to the flight path has
been considered. However, undesired movements of the platform that can lead to a change in the
trajectory or the appearance of a squint angle in the pointing of the antenna have also to be corrected. In
[4] more details on the signal processing algorithms can be found.
4.0 AIRBORNE PLATFORMS USED IN THE PROJECT
Two different platforms are being used in the on-going projects. For the 35 GHz High resolution FM-CW
SAR system the Stemme S10 motor glider is being used. The 450 MHz P-SAR project will be mounted in
an Edge X ultralight aircraft. This section provides some details on the aircrafts.
(a) (b)
Figure 2: The Platforms used: (a) The Steme S10 Motor Glider and (b) The Edge X Ultralight Aircraft.
4.1 The Stemme S10 Glider
The Stemme S 10 is a twin seat, light surveillance motor glider that can take off unassisted. It has a stalling
speed (minimum velocity to maintain flight level) of 78 km/h. When powered it has a maximum cruising
speed of 248 km/h. It can fly at altitudes up to 9140 m. It has 2 fuel tanks of 45 1 each, which allows a
range of about 1290 km. The glider also offers the possibility of flying unpowered, which minimizes the
vibrations of the aircraft. In this case the best glide ratio (distance travelled divided by the decrease in
altitude) is 50 and it is achieved at 106 km/h.
Two standardized pods can be mounted under the wings. These pods have a diameter of about 35 cm and
they are 80 cm long excluding the aerodynamic fairings. The maximum payload is 50 kg per pod.
4.2 The Edge X Ultralight Aircraft
The Edge X is a two-seat microlight aircraft from Airborne Australia. It is equipped with Streak Wings.
The main advantages of this kind of aircraft are their portability, ease of set-up and maintenance and
mechanical simplicity. They can land and take-off without requiring a conventional airstrip or specialized
fuel. The manufacturer claims that it is very stable even in adverse conditions and that the roll and pitch is
very light and predictable but this aspect has to be analyzed.
RTO-MP-SET-092
26-5
UNCLASSIFIED/UNLIMITED
UNCLASSIFIED/UNLIMITED
FM-CW Based Miniature SAR Systems for Small UAVs
ORGANIZATION
The dimensions from wheel to wheel are 1.89 meters. The aircraft total empty weight is a minimum of
167 kg and the maximum take off weight is 401 kg. If we assume that it can carry two persons, on average
160 kg, this leaves 74 kg as a total payload. However, considering that it may also carry other equipment
and that with lower weight it has better manoeuvrability, it is advisable not to exceed a 50 Kg payload.
The trim speed (the band of velocities where the glider flies without manoeuvring) is between 74 and
92 Km/h (20.5 to 25.5 m/s). The maximum level speed (maximum speed at a certain level) is 138 km/h at
minimum weight and 148 km/h at maximum weight. It can fly at altitudes varying from 300 to 1000 m.
It has a 44 1 fuel tank, which leads to a range of about 300 km or 3 hours duration, depending on weather
conditions and loading.
5.0 35 GHZ SAR DEVELOPMENT
5.1 System Design
The work on the demonstrator system started in 2001. In order to speed up the development and to show
that an FMCW SAR system can be relatively cheap, it was decided to use off-the-shelf components as
much as possible.
The demonstrator system operates in stripmap mode. The resolution in range as well as in azimuth
direction has primarily been chosen to be 1 m. During the project, the resolution will be gradually
enhanced to 30 cm in both range and azimuth direction. Further specifications are listed in table I.
Table I: 35 GHz FM-CW SAR System Parameters (High Resolution Mode)
Carrier frequency
35 GHz
Antenna Gain
24 dB
Frequency sweep
500 MHz
Antenna Isolation
52 dB
PRF
1 kHz
Beamwidth Az/El
6728°
Modulation
Sawtooth
Platform velocity
25 m/s
IF band
Dc to 2.5 MHz
Altitude
150 m
Transmitted power
18 dBm
Max. Range
730 m
The core of the demonstrator system is a PXI chassis manufactured by National Instruments. The chassis
includes a 1.26 GHz Pentium III controller, a 10 MHz, 12-bit A/D board to sample the radar data,
a 100 kHz, 16-bit A/D board to sample the motion data, and a 40 MHz, 12-bit D/A board to control the
frequency modulation. The radar data are sampled at 5 MHz; resulting in a continuous data rate of
approximately 9.5 Mbyte/s. The 35 GHz FM-CW front-end is manufactured by Epsilon-Lambda
Electronics. In addition, the demonstrator system is supplied with gyroscopes, accelerometers and a GPS
receiver to be able to determine the position and the attitude of the system. Finally, a digital camera has
been added to supply optical images of the imaged area.
During the flights, the demonstrator system can be controlled and monitored from the cockpit with the aid
of a pocket PC. To this end, some cables can be pulled from the pod to the cockpit through a tube in the
wing. Otherwise, the system is self supporting; it is fed by a battery which is also installed in the under
wing pod. The system can run for approximately 2.5 hours on a fully charged battery.
26-6
RTO-MP-SET-092
UNCLASSIFIED/UNLIMITED
UNCLASSIFIED/UNLIMITED
FM-CW Based Miniature SAR Systems for Small UAVs
Figure 3: FM-CW Radar Mounted in the Pod Attached to the Plane Wing.
5.2 Measurement Results
An airborne campaign was organized at the Strausberg airfield, Germany, on June 22nd and 23rd, 2004.
Four comer reflectors were placed in a grass field, see Fig. 4a, and the GPS coordinates of their positions
were measured. Two small 33 RCS dBm2 comer reflectors (1 and 2) and two large 40 RCS dBm2 comer
reflectors (3 and 4) were set up. Moreover, a GPS ground station was set up in the middle of the scene.
Several runs were flown along the comer reflectors at 100, 150, and 300 m altitude. Additionally, some
flights were made at an altitude of 200 m, during which the engine of the motor glider was switched off.
The weather was very turbulent and the wind was directed almost perpendicular to the runway. A squint
angle due to the aircraft yaw was therefore present during most measurements. The data presented in the
paper are obtained from a run flown at 1 00 m altitude.
The slant range to the middle of the scene is around 240 m. In this case, the demonstrator system was
operating in low resolution mode transmitting a bandwidth of 200 MHz, leading to a theoretical range
resolution of 75 cm. The azimuth compressed data are shown in Fig. 4b. Multilook with a reduction factor
equal to 16 has been used to reduce the speckle and so enhance the contrast of the image; of course also
the azimuth resolution has increased with the same factor. The geometry of the responses corresponds
very well to the geometry shown in Fig.4a. The practical range resolution is about 2 m, which is almost
three times the theoretical resolution. The spreading is mostly due to the residual frequency sweep non¬
linearity. This residual non-linearity may be further compensated with auxiliary data processing.
The resolution in azimuth direction is around 50 cm (in the multilook image), which is already very good.
Table II: Calibration Results Obtained by Processing the
Images Collected during the Airborne Campaign
Comer 1 RCS
32.3 dBm2
Comer 2 RCS
33.2 dBm2
Comer 3 RCS
37.8 dBm2
Comer 4 RCS
40.0 dBm2
Image clutter c0
19.9 dB
Noise o0
17.9 dB
RTO-MP-SET-092
26-7
UNCLASSIFIED/UNLIMITED
UNCLASSIFIED/UNLIMITED
FM-CW Based Miniature SAR Systems for Small UAVs
ORGANIZATION
The peak level of the response of the 40 dBm2 comer reflectors is about 18 dB above the noise level.
In order to investigate the performance a calibration process was carried out from the data collected. From
the measured image response of the known reference the backscatter coefficient for uniform areas in the
image can be derived, and by using an integral calibration approach the method is independent of the
azimuth or range resolution and it does not involve the scene, radar, or processor partial coherence [5].
The use of multilook images does not influence the integral calibration algorithm.
Comer reflector 4 (see again Fig. 4a) has been chosen as the known reference reflector and its RCS has
been set to a value of 40 dBm2, its theoretical value. The backscatter coefficient of the clutter and of the
other comer reflectors has been calculated from this reference reflector. The same procedure has been
repeated using also other images of the same area in order to collect a valid data set and then averaging the
results in a root mean square way. For the calculation of the noise level, a file has been collected with the
radar pointing at the sky. Raw data have been processed with the SAR processor, so that this noise data
have passed through the same processing chain as the aircraft data. Results are shown in table II. From
these two conclusions can be drawn:
• the RCSs of the comer reflectors are quite consistent with the theoretical values: 33 dBm2 for
comer 1 and 2, and 40 dBm2 for comer 3 and 4;
• the backscatter coefficient of the clutter is close to the equivalent backscatter coefficient of only
noise.
The value of the backscatter coefficient for grass is around -12 dB [6], so between the expected one and the
calculated one there is a difference of more than 30 dB. Some of this discrepancy can be explained by the
fact that the comer reflectors could not be exactly pointing at the radar, but still the difference between the
theoretical and the calculated backscatter coefficient is quite high. What has been measured as clutter is
therefore only noise. When the noise data have been collected the set up of the radar was a little bit
different than during the airborne measurement: the pod cover was not used, and the power supply for the
motion sensors was disassembled. It has been seen in other tests that both of them have some influence,
specially the power supply. This could explain the difference of 2 dB between the noise level in aircraft
images and the level in the noise file. The results from the airborne campaign are promising even if the
noise level [7], is the major limitation. However the corner reflector responses are enough to validate the
signal processing. In [8] to [10] details on other tests made can be seen.
Figure 4: (a) Airborne Picture of the Area of Measurements with the Corner Reflectors
(numbered 1 to 4) and (b) SAR Image of the Four Reflectors after the Squint Angle Correction.
26-8
RTO-MP-SET-092
UNCLASSIFIED/UNLIMITED
UNCLASSIFIED/UNLIMITED
FM-CW Based Miniature SAR Systems for Small UAVs
6.0 450 MHZ P-BAND POLARIMETRIC SAR DEVELOPMENT
6.1 Application and Requirements
P-band offers unique capabilities that have not been fully exploited yet. Its large wavelength allows
penetrating surfaces like forest canopy, ice sheets and certain kinds of soil. Thus, very useful data, that
could not be obtained otherwise, can be retrieved. Indeed this frequency range has revealed such a
potential that there are on-going discussions in the International Telecommunications Union (ITU) on the
convenience of allocating a specific band for earth observation systems in the 420-470 MHz range and
therefore diminish the problems of interference that the existing systems currently face. Applications from
space are enabled this way.
The capability of penetrating the sub-surface of P-band has revealed this frequency to be the most
adequate in a wide range of applications. As an example there have been studies on the application in the
determination of soil moisture and dielectric properties. Other studies suggest the suitability of P-band to
detect man-made structures like archaeological sites, and recent experiments have proved it adequate to
detect water in dry areas. P-band also has applications to ice sheet studies. Data from P-band imaging
radars can help estimating the internal structure of ice sheets and therefore provide a vision of the
evolution of the ice sheet in the poles which could help determine the impact of climate change.
The ability to penetrate dense forest canopy makes P-band adequate for topographical studies in these
areas.
Another important commercial application is its use in sea bottom topography. Under favourable
meteorological conditions the bottom topography of seas can be mapped with airborne radars.
The interaction between marine currents and the bottom topography causes variations in the surface
current velocity. These variations give rise to differences in the spectrum of the wind-generated waves,
which can be observed by the radar. Sea bottom topography using airborne radar avoids or reduces the use
of expensive bathymetric mapping from ships.
The P-SAR system designed will be used in tropical forest monitoring activities in Indonesia. Within the
ESA Earth Explorer Opportunity Mission (EEOM), the programme BIOMASCA has been proposed for
worldwide forest biomass monitoring using polarimetric P-band SAR. Plans are going on to study the
practicability of a satellite based P-band SAR. However, the poor availability of P-band data has prevented
development of robust biomass estimation algorithms. Especially the lack of data in the tropics (where
biomass levels can be much higher), and the general lack of multi- temporal data are considered as a major
problem. The variety of existing biomass levels over the test sites in Indonesia (of the BOS organisation)
will give insight into the possibilities and limitations of long wavelength SAR for direct biomass
estimation. BOS and WUR intend to acquire observations over long time periods to capture seasonal
variations. Thus, a unique and very meaningful scientific data set will be collected.
Moreover, most of the world’s tropical peat swamp forests are located in Indonesia (30 out of 50 million
ha). These areas are relatively small but contain enormous carbon stocks, approximately equal to the
carbon stored in all temperate forests of the world together. Because of deforestation and land use change
these areas are at risk and are already producing enormous carbon emissions (through oxidation and fire).
These emissions are a major threat to the world’s climate. Peat land restoration is considered urgent and
very important. A key element of peat land restoration is the ability to get information on hydrology and
flooding under the forest canopy. P-band radar is the best suitable instrument for this purpose. The use of
P-band in peat swamp forests will be studied by WUR and BOS and a link will be made to temporal
dynamics of flooding. For this purpose one of the peat swamp areas under management by BOS is
equipped with permanent hydrological sensors measuring water levels along a large 23 km transect across
a peat dome.
RTO-MP-SET-092
26-9
UNCLASSIFIED/UNLIMITED
UNCLASSIFIED/UNLIMITED
FM-CW Based Miniature SAR Systems for Small UAVs
ORGANIZATION
The radar should be fully polarimetric for better land use classification capabilities. Since the area under
observation is quite homogeneous and the measurements are meant for investigation of future satellite
missions, a high resolution will not be necessary. A maximum of 30 meters in ground range resolution is
considered sufficient, but 15 meters is desired. Multiple looks will be used in order to reduce speckle
noise.
Considering the increasing interest in developing a spacebome P-band SAR, small angles of incidence are
preferred. The six major types of forest to be observed are palm oil, rubber, mangrove, secondary tropical
forest, primary tropical forest and primary tropical forest burnt. The reflectivity of these areas varies from
-5 dBm2 to -35 dBm2. The dynamic range between the cross-polarized and co-polarized channels
is 15 dB. The desired clutter-to-noise ratio (CNR) after the signal processing is a minimum of 15 dB.
Table III summarizes the requirements of the system.
Table III: Specifications of the System
Ground range resolution
<15 m (<30 m max)
Azimuth resolution
<1 m (<15 m multi look)
Number of looks
>15
Swath width
800 m
Angle of incidence
25° to 65°
CnRsar
>15 dB
Altitude of the platform
500 to 600 m
Velocity of the platform
75 km/hr
6.2 System Design
The core of the system is the DDS. This device synthesizes the chirp signal. Different modulations can be
performed by simply programming several registers. The device can synthesize signals with a bandwidth
up to 1 00 MHz, which gives a great flexibility when performing the modulation. A clock signal at a rate
equal to the sweep repetition frequency (SRF), I/Oclock, is available for external use. The DDS is
controlled by a microcontroller.
The output signal of the DDS is then upconverted to the carrier frequency by using a 400 MHz low phase
noise crystal oscillator and an active mixer. After being filtered, the resultant signal is amplified and sent
to the antenna. A switch controlled by means of the I/O clock drives the signal sequentially to the vertical
or the horizontal polarization feeding points of the antenna.
An antenna similar to the one used in transmission receives the echo from the scatterer. The signals
received from the horizontal and vertical polarization feeding points are processed separately.
The received signal is first amplified by a low-noise amplifier. After amplification it is mixed with a
portion of the transmitted signal in an active mixer. The resultant low frequency signal is filtered and
amplified before being digitized and stored in a data acquisition system. The data acquisition system
consists of an analog to digital converter and a set of flash memories controlled by a microcontroller to
store the data. In order to increase simplicity the processing of the data is done off-line.
The microcontroller has a USB interface to facilitate the transfer of the stored data to a PC.
The whole system timing is based on the 400 MHz oscillator used to upconvert the signal. The sinusoidal
output of the oscillator feeds a comparator integrated in the DDS. The output of the comparator is divided
by 2 in order to supply the internal clock signal for the DDS. This internal clock signal is also used to
control the data acquisition system.
26-10
RTO-MP-SET-092
UNCLASSIFIED/UNLIMITED
UNCLASSIFIED/UNLIMITED
FM-CW Based Miniature SAR Systems for Small UAVs
6.3 Development Plan
Preparatory work on this project has already been performed. During the pre-project stage literature
research has been carried out, the required specifications of the system have been established, the main
aspects and challenges have been taken into consideration and an initial design of the system has been
made [11].
Based on this pre-project work a four-year research should be conducted in order to realize and test the
system and to carry out FM-CW modulation studies. The system should be realized by the end of the first
year. The second year will be devoted to ground tests and flying tests in the Netherlands and by the third
year the system should be operational on site.
The adaptation of the SAR algorithms developed for the 35 GFIz high resolution FM-CW SAR system
will start during the first year and will be conducted in parallel to the construction of the system. The third
year will be devoted to test measurements and validation of the processing algorithms.
And initial study of FM-CW modulations with DDS based on literature research and simulation will start
during the second year. Once the system has been tested and the algorithms are validated, FM-CW
modulations will be studied from data obtained with the system.
7.0 CONCLUSIONS
In IRCTR, and founded by the Netherlands Science Foundation, STW, an investigation is ongoing on the
feasibility of FM-CW SAR systems mounted on small aircraft. The first system developed was a 35 GHz
high resolution FM-CW SAR. The limited performance of the radar sensor, which was adapted from an
existing automotive radar in order to speed up the investigation, has prevented obtaining a good quality
RTO-MP-SET-092
26 - 11
UNCLASSIFIED/UNLIMITED
UNCLASSIFIED/UNLIMITED
FM-CW Based Miniature SAR Systems for Small UAVs
ORGANIZATION
image. Nevertheless the measurements have allowed the validation of the FM-CW SAR signal processing
algorithms.
Based on the know-how obtained in the previous project currently a second system is being developed.
The second system is a P-band polarimetric FM-CW SAR which will be mounted in an ultralight aircraft.
The system will be used by WUR and BOS to carry out tropical forest monitoring in Indonesia but P-band
has a wide range of applications. Since the phase noise and the non-linearity of the sweep revealed to
be the major problems in obtaining a good-quality image a low phase noise oscillator and a DDS
(which provides a very linear modulation) will be used. This second system is expected to be operational
in 2 years. A funding proposal has been submitted to STW.
REFERENCES
[1] Wehner, D.R., “High-Resolution Radar”, Artech House, Boston, 1995.
[2] Skolnik M.L., “Introduction to Radar Systems”, McGraw-Hill, Inc. London, 1980.
[3] Carrara, W.G., Goodman, R.S., and Majewski, R.M., “Spotlight Synthetic Aperture Radar. Signal
Processing Algorithms”, Artech House, Boston, 1995.
[4] Meta, A. and Hoogeboom, P., “High Resolution Airborne FM-CW SAR: Preliminary Results”,
European Synthetic Aperture Radar Conference, EUSAR04, Ulm (Germany), 25-27 May 2004.
[5] Gray, A.L., Vachon, P.W., Livingstone, C.E. and Lukowski, T.I. “Synthetic Aperture Radar
Calibration Using Reference Reflectors”, IEEE Transactions on Geoscience and Remote Sensing,
Vol28,No3, May 1990.
[6] Ulaby, F.T. and Dobson, M.C., “Handbook of Radar Scattering Statistics for Terrain”, Artech House,
1989.
[7] de Wit, J.J.M. and Hoogeboom, P., “High Resolution FM-CW SAR Performance Analysis”,
International Geoscience and Remote Sensing Symposium, IGARSS03, Toulouse (France),
21-25 July 03.
[8] de Wit, J.J.M, Meta, A. and Hoogeboom, P., “First Airborne FM-CW SAR Campaign, Preliminary
results” International Radar Symposium IRS04, Warsaw (Poland), 19-21 May 2004.
[9] de Wit, J.J.M. and Hoogeboom, P., “High Resolution FM-CW SAR Experimental System: First
Results” European Synthetic Aperture Radar Conference EUSAR04, Ulm (Germany), 25-27 May
2004.
[10] Meta, A., de Wit, J.J.M. and Hoogeboom, P., “Development of a High Resolution Airborne
Millimeter Wave FM-CW SAR”, European Radar Conference, Amsterdam (The Netherlands),
14-15 October 2004.
[11] Figueras i Ventura, J. and Hoogeboom, P., “Design of a Small, Low Cost, P-band, Airborne
Polarimetric Synthetic Aperture Radar”, European Radar Conference, Amsterdam (The
Netherlands), 14-15 October 2004.
26-12
RTO-MP-SET-092
UNCLASSIFIED/UNLIMITED