International Journal
on Marine Navigation
and Safety of Sea Transportation
Volume 3
Number 3
September 2009
311
1 INTRODUCTION
The CRM-203 type Coastal Surveillance Radar is
solid state Frequency Modulated Continuous Wave
(FMCW) sensor with low transmission power. One
of the most significant parameters of this coastal ap-
plication is small targets detection possibility in
heavy sea clutter conditions and high range resolu-
tion. The requirements perform FMCW technology,
which is rapidly advancing recently. Fully solid-state
transmitter design (due to the low radiated power)
ensures excellent Mean Time Between Failure
(MTBF) and practical without service continuous
operation.
FMCW transmitter produces a constant amplitude
linear frequency modulated signal. The principle of
FMCW radar is presented on Figure 1.
Figure 1. Principle of FMCW radar
Radar signal is transmitted, reflected by the sur-
face of the target and then received after a delay
time τ:
τ =2R / c (1)
where: c = speed of light; and R = distance.
The difference between the transmitting and re-
ceiving frequency f
R
is directly proportional to the
distance and is used to further FFT processing
(Wawruch & Stupak 2008):
f
R
= 2Rf / cT (2)
where: f = frequency deviation; and T = modula-
tion period.
Frequency Modulated Continuous Wave technol-
ogy offers low probability of intercept feature be-
cause of the low peak power and frequency modula-
tion.
2 RADAR GENERAL DESCRIPTION
The prime function of CRM-203 is detection and es-
timation of planar co-ordinates for sea surface tar-
gets and automated tracking the selected ones to per-
form the coastal surveillance tasks. The radar sensor
gives a presentation of the current sea situation and
calculates the future situation to accomplish the au-
tomated radar plotting aids.
Functional diagram of described radar is present-
ed on Figure 2. Radar sensor basically includes the
antennas integrated with FMCW transceiver, anten-
nas motor drive and Signal Processing & Control
Unit (SPCU) also including local interface and con-
CRM-203 Type Frequency Modulated
Continuous Wave (FM CW) Radar
S. Plata
Telecommunication Research Institute Ltd., Gdansk, Poland
R. Wawruch
Gdynia Maritime University, Gdynia, Poland
ABSTRACT: Paper presents description of the principle of work, structure and basic technical parameters of
the Maritime Coastal Surveillance Frequency Modulated Continuous Wave (FMCW) Radar CRM-203 type
constructed by Telecommunication Research Institute Ltd. in Gdańsk. Results of its tests in real conditions
and comparison with pulse ship radars with scanners installed in the same place will be presented during the
conference.
transmitter
receiver
f
min
f
max
τ
T-
τ
-T/2
T/2
0
f(t)
312
trol circuitry. Each radar transceiver is controlled by
the SPCU which is connected to the Operations Cen-
tre (OC).
The main functions performed by radar are:
selection of the operative mode (local or remote);
reception from the OC of all controls/ commands
and selections needed for complete operation ca-
pability; in case of control line failure all the con-
trols are automatically put in a default condition -
in order to guarantee the antenna rotation, the ra-
dar emission and the automatic acquisition and
tracking of targets;
processing of the radar video; compression and
transmission to the OC of the digitised video;
automatic or on-demand acquisition of targets
falling inside predefined automatic acquisition
zones or acquired by the operator;
automatic tracking of targets falling inside prede-
fined automatic tracking areas; and
transmission to the OC, once per antenna revolu-
tion, of status and alarms from the sensor (BITE),
track data (position, speed/course) of targets un-
der tracking.
Figure 2. CRM-203 functional diagram
3 ANTENNAS
CRM-203 in the coastal surveillance application has
typical requirements as small target detection in
weather and sea clutters and high angular resolution.
To provide good angular resolution a narrow azi-
muth beam is required. A narrow azimuth beam is
desirable to reduce resolution cell size for three main
reasons:
to provide accurate bearing information on the
target;
to differentiate between targets which are close
together; and
to reduce clutter returns.
The used system features 12 feet, X-band anten-
nas with horizontal polarisation and the following
electrical parameters:
3dB horizontal beam width = 0.7°;
3dB vertical beam width = 22°; and
gain = 32 dB.
Each antennas group consists of a pedestal sup-
porting the rotating unit. The pedestal contains the
drive mechanism, the rotary joint and an 4096 pulses
encoder for transmission of antenna position data.
The power rating for motor controller is 1.5 kW. An-
tennas rotating speed is selectable between 12 and
30 rpm. The antennas group is designed to withstand
severe marine environmental conditions such as salt
spray, sun light, sand, etc.
4 TRANSCEIVER
Functional diagram of the transceiver unit is shown
on Figure 3. Direct Digital Synthesizer (DDS) pro-
duces a synthesized “chirp” (linear frequency modu-
lated) signal. Next this signal is up-converted by
multiplier. The last stage of transmitter circuit is sol-
id state power amplifier (PA), which feeds X-band /
10W FMCW signal to antenna. The receiver consists
of a low noise amplifier (LNA), image rejection
mixer and intermediate frequency amplifier (IFA).
Figure 3. Functional diagram of transceiver unit
The DDS advantages include very fast switching
(typically sub microseconds), excellent phase noise,
transient-free (phase continuous) frequency changes,
extraordinary flexibility as a modulator, and small
size, among others. Frequency changes look like
those of a Voltage Controlled Oscillator (VCO)
smooth and without phase discontinuity sweep
across a defined frequency range with synthesizer
accuracy, but without the glitches and transient pro-
duced by any other synthesizer technique. Because
of the synthesis techniques, this characteristic is
unique to the DDS and enables it to produce a syn-
thesized “chirp”. It is very important in FMCW ap-
plications because frequency modulation accuracy is
313
directly influencing on accuracy of distance meas-
urements and frequency modulation non-linearity
decreases targets detection.
The proposed transceivers have some additional
features which make them specially suited for
coastal applications:
sector blanking: emission can be inhibited within
an adjustable sector, so as to avoid undesired re-
turns (e.g. land clutter);
12 dB/okt. frequency curve slope of IFA amplifi-
er ensures equal intermediate frequency (IF) out-
put signals for targets in different ranges; and
digital automatic receiver gain control function
ensures optimal IF signal output level, inde-
pendently of under detection target radar cross
section.
5 SIGNAL PROCESSING AND CONTROL UNIT
5.1 Functional diagram
Functional diagram of the signal processing is
shown in Figure 4.
Figure 4. Functional diagram of signal processing
The frequency measurement performed to obtain
the range measurement is made digitally using the
Fast Fourier Transformation (FFT). So the IF signal
is digitised and sent to the spectrum analyser that
performs FFT. On the input of the signal processing
an analog-to-digital converter samples the IF signal
with 8 MHz frequency and 12-bit resolution. Next
the spectrum analysis of the digitised IF signal is
performed on the base of 8192-point FFT. At the
output of the spectrum analyser a periodogramm
presenting 4096 range cells is obtained. Range cell
size is 5.6 m for radar scale range 12 NM.
The signal after the frequency analysis can be
best referred as the video signal. The signal is indeed
an exact analogue of the video of pulse radar. The
range data output from the spectrum analyser is fur-
ther processed like in pulse radar: CFAR (Constant
False Alarm Ratio) thresholding and binary integra-
tion during the dwell time on a target are performed.
Spectrum analysis
The IF signal is analysed using FFT transform. The
analysis is carried out in real time. The analysed sig-
nal can be modelled as a sum of sinusoids embedded
in noise and clutter. In FMCW processing scatters at
different ranges appear as different constant fre-
quency components at the IF output. The FFT re-
sponse to a sinusoidal input reveals a main lobe and
side lobes. The width of the main lobe indicates
Fourier Domain Resolution, which for CRM-203
application is very narrow and equal 1 kHz. This
Fourier Domain Resolution or differently bandwidth
of FFT frequency cell is very important parameter of
FMCW radar, because of detection performance.
Probability of detection depends on the ratio of the
target received signal level to the sum of clutter and
noise. FMCW transceiver noise power N
i
is function
of the FFT frequency cell bandwidth:
N
i
= kT
e
B
FFT
(3)
where: k = Boltzman’s constant; T
e
= effective noise
temperature; and B
FFT
= FFT frequency cell band-
width.
This relationship explains excellent CRM-203 ra-
dar noise properties allowing low transmitter power.
5.2 CFAR thresholding
The radar must detect a target against a changing
background of clutter and noise. The clutter reflec-
tivity and statistics will generally vary with range
and direction. The problem is how to set a threshold
to provide an acceptable probability of false alarm
P
fd
whilst maximising the probability of detection
P
d
. Standard detection strategy is to fix the P
fd
. In
CRM-203 application an automatic CFAR detector
is used. To control of the false alarms, the detector
must be able to estimate the parameters of the prob-
ability density function of the clutter and noise re-
turns. A well known method of estimating the clutter
mean level is the cell-averaging CFAR circuit. The
mean level of the cell under test is obtained from the
average of a number of surrounding clutter cells. A
gap between the cell under test and the surrounding
cells is method to ensure that a distant relative strong
target does not contaminate the clutter estimates. In
CRM-203 radar smallest off CFAR window is taken
to calculate threshold. This strategy helps to detect
small targets in neighbourhood of strong clutter re-
gion. Size of CFAR window is small so a threshold
can follow the local clutter mean and can give a
much better performance in detection in our case.
After CFAR the binary integrator is used with the
“M-of-N” rule in accordance with formulas valid for
Gaussian noise.
314
The Signal Processing & Control Unit includes a
local display facility, in order to allow local mainte-
nance and set-up operations. All radar controls are
available on the local panel. Moreover SPCU feeds
video signal to Radar Display Unit, which accom-
plish:
video acquisition and processing;
plot extraction; and
tracking.
Track data are sent to the OC for further pro-
cessing. Also plot information can be routed through
the same communications channel. All the radar
controls, including those available at the local panel,
can be also sent by the OC via the remote interface.
The radar continuously sends the status information
to the OC, together with target data.
6 TECHNICAL DATA
Basic technical are presented in Tables 1-7.
Table 1. Transmitter
__________________________________________________
Parameter Value
__________________________________________________
Output power 1mW-2W (switched)
Carrier frequency 9.39.5 GHz
Frequency deviation switched according to the required
scale range:
54 MHz at 6 NM
27 MHz at 12 NM
13.5 MHz at 24 NM
Range scales 0.25 NM 48NM
Modulation DDS based linear FMCW
Sweep repetition period 1 ms
__________________________________________________
Table 2. Receiver
__________________________________________________
Parameter Value
__________________________________________________
IF bandwidth 4 MHz
Noise factor 2 Db
Maximum gain 120 Db
Frequency curve slope of IF amplifier 6 dB/oct; 12 dB/oct; 18
dB/oct.
__________________________________________________
Table 3. Antennas
__________________________________________________
Parameter Value
__________________________________________________
Antenna length 3.6 m
Beamwidth (3 dB) horizontal/vertical 0.70°/22°
Polarisation Horizontal
Gain 32 dBi
Rotation speed min/max. 12/30 rpm
Drive motor 1.5 kW
__________________________________________________
Table 4. Signal processing
__________________________________________________
Parameter Value
__________________________________________________
FFT signal processing 8192-points FFT
Sampling frequency 8 MHz
Number of range cells 4096
Signal thresholding CFAR
Signal integration binary, number of detections de-
pendent on antenna rotation speed
Sea clutter reduction signal correlation from 2 antenna
rotations.
__________________________________________________
Table 5. Display unit
__________________________________________________
Parameter Value
__________________________________________________
Display size 22 inch
Resolution 1280 × 1024 pixels
Acquisition automatic up to 100 targets
Tracking automatic of all acquired targets
Zones 2 guard zones
Target information target number, target range and bearing
from radar position, target course
Options ARPA anti-collision functions
__________________________________________________
Table 6. Range and angle measurements
__________________________________________________
Parameter Value
__________________________________________________
Scale range [NM] 12 / 24 / 48
Range cell size [m] 5.6 / 11 / 22
Range measurement accuracy 1% of selected range or 50 m
(whichever is greater)
Angle resolution 0.1°
Bearing accuracy 0.7°
__________________________________________________
Table 7. Environmental conditions
__________________________________________________
Parameter Value
__________________________________________________
Wind operational 30 m/s
Wind survival 50 m/s
Humidity 98 %, 25 °C
Temperature operational from -10 °C to +50 °C (inside operat-
ing room) and from -30 °C to +50 °C
(outside operating room)
Temperature survival from -40 °C to +65 °C
__________________________________________________
7 RECAPITULATION
Described radar was installed in the radar laboratory
of the Gdynia Maritime University this year. Its an-
tenna is located on the roof of the university building
nearby the south entrance to the Gdynia Harbour.
Operational tests of the radar will be conducted in
December 2008. Its detection possibilities, accura-
cies and clutters resistance will be checked during
measurements in real hydro-meteorological condi-
tions. Results will be compared with information
about positions, courses and speeds received from
Automatic Identification Systems (AIS) installed on
board detected and tracked objects and data about
these objects received at the same time from four
different ship pulse radars installed in the same la-
boratory. Results of these tests will be presented on
the conference.
REFERENCES
Wawruch R. & Stupak T. 2008. Charakterystyka radaru na falę
ciągłą. Prace Wydziału Nawigacyjnego Akademii Morskiej
w Gdyni No. 21, p.120-130.