PAPER

Performance Evaluation of SeaSonde

High-Frequency Radar for Vessel Detection
AUTHORS

ABSTRACT

Hugh J. Roarty
Erick Rivera Lemus
Ethan Handel
Scott M. Glenn
Coastal Ocean Observation
Laboratory, Rutgers University

High-frequency (HF) surface wave radar has been identied to be a gap-lling

technology for Maritime Domain Awareness. Present SeaSonde HF radars have
been designed to map surface currents but are able to track surface vessels in a
dual-use mode. Rutgers and CODAR Ocean Sensors, Ltd., have collaborated on
the development of vessel detection and tracking capabilities from compact HF radars, demonstrating that ships can be detected and tracked by multistatic HF radar
in a multiship environment while simultaneously mapping ocean currents. Furthermore, the same vessel is seen simultaneously by the radar based on different processing parameters, mitigating the need to preselect a xed set and thereby
improving detection performance.
Keywords: radar, detections, maritime domain awareness, dual-use

Donald E. Barrick
James Isaacson
CODAR Ocean Sensors, Ltd.,
Mountain View, CA

Introduction

he U.S. Integrated Ocean Observing System (IOOS ) led by the

National Oceanic and Atmospheric
Administration (NOAA) has designed
(Interagency Working Group on
Ocean Observation, 2009), is constructing, and has recently begun operating the more advanced portions of a
national high-frequency (HF) radar
network focused on the real-time mapping of surface currents. The primary
operational users of the resulting surface current maps are the U.S. Coast
Guard for Search and Rescue and the
NOAA HazMat team for ocean spill
response. The IOOS Mid-Atlantic Regions CODAR SeaSonde HF Radar
Network, led by Rutgers University,
is the rst region in the United States
to achieve operational status by constructing and operating the endto-end system that produces and
links validated real-time surface current maps to the Coast Guards Search
and Rescue Optimal Planning System
(Roarty et al., 2010b).

14

The Department of Homeland

Security has called for the development
of tools to provide wide-area surveillance from the coast to extend beyond
the horizon (Department of Homeland
Security Science and Technology,
2009). Rutgers and CODAR Ocean
Sensors, an academic-industry partnership established in 1997, have worked
together for over a decade to expand
the capabilities of compact CODAR
HF radars to include the dual-use application of detecting and tracking ships
without compromising the networks
ability to map surface currents. Initial
development focused on the demonstration and evaluation of a non-realtime end-to-end system for dual-use
vessel tracking in the New York Bight
multifrequency HF radar testbed
(Roarty et al., 2010a). Technology
demonstrations determined (a) that
vessels could be detected, (b) that
multilook detections could be associated with a known ship, and (c) that
the associated detections could then be
input to a range of tracking algorithms

Marine Technology Society Journal

whose output produced tracks and predicted trajectories on a computer screen,

providing useful information to operators. Radar hardware development
focused on developing network exibility beyond monostatic backscatter
operations, demonstrating (a) that bistatic and multistatic operations were
possible with a shore-based network
and (b) that buoy-based bistatic transmitters can be operated at all three of
the commonly used HF radar frequencies (5-6, 12-13, and 24-25 MHz).
The greatest challenge in developing
a robust ship surveillance capability
for any HF radar is the development of
the initial vessel detection algorithm.
This conclusion focused the initial research on the mathematical problem
of identifying and extracting the radar
return of a surface vessel hidden within
a highly variable and noisy background, requiring additional detection
algorithm development, testing, and
sensitivity analysis in a variety of environments with different noise characteristics. It is the aim of this paper to

analyze the parameters and settings of

the vessel detection algorithm that are
optimal for nding those ship echoes
among the other signals that are sent
back towards the radar. In Methodology, we describe the HF radar network, the SeaSonde HF radar used in
this test, the Automatic Identication
System (AIS) network used to ground
truth the radar detections, and the ship
detection software used to process the
radar data for ships. In Results, we discuss the results of the ship detection test
that was conducted. Lastly, the performances of the various ship detection
processing methods against the available targets are discussed in Discussion.

Methodology

HF Radar Network
These experiments were conducted
within the New Jersey Shelf Observing
System (Glenn & Schoeld, 2002). A
major component of this observing
system is an HF radar network. The
network was created in 1998 with the
placement of two 25-MHz systems on
the southern coastline of New Jersey
(Kohut & Glenn, 2003). The network
was then expanded with the placement
of four 5-MHz systems spanning the
New Jersey coastline (Gong et al.,
2009). The 25-MHz network was
moved north in 2003 in support of the
Lagrangian Transport and Transformation Experiment (Chant et al., 2008).
The work discussed here utilizes the latest addition to the network: the placement of a 13-MHz system outside the
entrance to New York Harbor in Sea
Bright, New Jersey. This network also
contributes to the Mid-Atlantic Regional Coastal Ocean Observing System
(MARCOOS), which has a total of
30 radars from Cape Hatteras to Cape
Cod that are operated by eight universities (Roarty et al., 2010b).

13-MHz SeaSonde System

FIGURE 1

The radar was deployed in Sea

Bright, New Jersey, 40 km south of
the Battery in New York City. The
radar was a direction-nding type
radar, SeaSonde Remote Unit SSRS100, manufactured by CODAR Ocean
Sensors and was installed in October
2008. The radars primary function
was the measurement of surface currents, which are provided in real time
to the NOAA National HF Radar
Network (Temll et al., 2006). The
radar also has the dual-use capability
to detect the location of ships at sea.
The radar consists of a compact receive antenna with three elements: two
directional crossed loops and an omnidirectional monopole, a monopole
transmit antenna, and a hardware
housed within a climate-controlled enclosure (Figure 1). The radar transmits
a radio wave with a center frequency
of 13.46 MHz and a bandwidth of
50 kHz. The bandwidth of the radar
sets the spatial range resolution of the
system, which was about 3 km for this
particular bandwidth. The details of
the waveform are given in Table 1.
Separate transmit and receive antennas
were used for this study spaced at least
one wavelength apart, which is approximately 23 m at the 13-MHz radio
band. A ship with a vertical structure
of a quarter wavelength (6 m) is the
minimum-sized optimal reector
(Ruck et al., 1970).
Figure 2 shows the spatial and temporal radial vector coverage for ocean
currents of the radar over a 1-week
period, which coincided with the ship
detection exercise. The radar collected
range data, which are a time series of
the complex echo signal voltages
before Doppler processing, from
00:00 GMT to 01:00 GMT on February 26, 2009. These range les are the
result of the rst fast Fourier transform

Picture of the (A) transmit antenna, (B) receive

antenna, and (C) equipment enclosure for the
SeaSonde 13-MHz radar.

TABLE 1
Characteristics of the radar waveform used in
the study.
Waveform Characteristic

Value

Center frequency

13.46 MHz

Bandwidth

50 kHz

Blank

668.8 s

Blank delay

8.55 s

Sweep rate

2 Hz

Pulse shaping

On

May/June 2011

Volume 45

Number 3

15

FIGURE 2
Radial coverage map of the SeaSonde at Sea Bright, New Jersey, over a 1-week period. The color
map illustrates the temporal coverage along the radial grid (black = 75%, red = 50%, pink = 25%).
(Color versions of gures available online at: http://www.ingentaconnect.com/content/mts/mtsj/
2011/00000045/00000003.)

Figure 3. The AIS transmissions are

used as ground truth for the HF radar
ship detections. The range of the
shore-received AIS signal is typically
30 nautical miles, but under certain
atmospheric conditions, range can be
upwards of hundreds of nautical
miles (International Maritime Organization, 2006). Data from the individual AIS receivers were sent back
to the Rutgers Coastal Ocean Observation Laboratory (Glenn & Schoeld,
2009), where it was archived using
the Coast Guard s oftware AIS
Source. The data were then time
stamped using the clock on the computer. The computer kept time with
a software tool to synchronize with
the National Institute of Standards
and Technology Internet Time Service. Figure 4 shows the tracks of vessels sent via AIS over a 3-week period,
indicating that this is a target-rich environment and that vessels do not
always stay in the shipping lanes.

Detection Software
(FFT) of the frequency-modulated
continuous wave received signal. The
range data were collected using
an FFT length of 512 points. With a
2-Hz sweep of the radar, each range
le encompasses 256 s of coherent integration time. There were a total of
15 range les over the hour-long period. The time on the computer and
all the subsequent les it generates are
synchronized to atomic time via GPS
by the Macintosh operating system.

AIS Receivers
Rutgers also operates an AIS receiver
network, which was utilized in this
study; this allows transponders on

16

vessels to broadcast the ships position

and identication. When earlier work
was performed (Roarty et al., 2010a),
the authors were limited to verifying
detections of vessels where a selfrecording GPS could be placed on a
vessel by the researchers or when the
GPS information could be provided
by other researchers (Rossby &
Gottlieb, 1998). The ability of the researchers to obtain AIS position data
on the vast majority of ships at sea
has greatly accelerated the research.
The Rutgers AIS network has receivers, which are manufactured by Shine
Micro, Inc., located at its eld station in Tuckerton, Sandy Hook, and
Loveladies, New Jersey, as shown in

Marine Technology Society Journal

The ship detection algorithm is explained in Roarty et al. (2010a). The

ship detection code is written in the
MATLAB programming language
and is designed to run ofine in a
batch-processing mode. The range
data that were collected by the radar
were read by the software to process
for the hard targets. The ship detection
code utilizes a constant false alarm rate
(CFAR) to nd targets. A signal that is
above the background by some threshold on the monopole and at least one
of the two loops is considered a possible detection. Figure 5 shows the
spectra from the monopole of the Sea
Bright radar site at 00:15:21 GMT
on February 26, 2009. The x axis denotes Doppler shift, the y axis shows
signal strength, and the z axis denotes

FIGURE 3
Location of the three AIS receivers operated by Rutgers University shown as green circles.
(Color versions of gures available online at: http://www.ingentaconnect.com/content/mts/
mtsj/2011/00000045/00000003.)

FIGURE 4
Map of the study area showing tracks of vessels (red lines) sent via AIS over a 3-week period.
The Nantucket, Hudson Canyon and Barnegat (clockwise from right) shipping lanes are shown
in the bottom right.

the radar range cell, which was 3 km

for these spectra. The Bragg peaks
from which surface currents are
derived (Barrick, 1972) are shown
0.4 Hz. The large signals at zero
Doppler are signals returned to the
radar from stationary objects. Ship
signals can be seen between the
Bragg peaks and the zero Doppler
signals, with positive Doppler shift
denoting a ship moving towards the
radar and the corollary with a signal
on the negative Doppler measuring a
ship moving away from the radar.
The ship detection code is designed
to identify these signals. The code utilizes two schemes as the basis for
thresholding in its CFAR peakpicking, where one averages in time,
innite impulse response (IIR), and
the other averages in Doppler and
range space using a median to create
the signal background. In its current
form, the code is able to process the
data using three combinations of
threshold and integration time for
each background simultaneously.
This results in a set of six detection
packages after each software run. The
detection code performs a sliding
FFT on the range data so a new detection le is output every 32 s. This setting is adjustable so that the user can
input the desired update rate for the
detections. The output of the detection code is a series of les that contain
range, range rate, and bearing of possible detections from the radar. The les
also include the uncertainties in the
above quantities, the signal-to-noise
ratio for each antenna, and an estimate
of the radar cross section of the possible
target.

Results
The range data that were collected
at the radar site were transported back
May/June 2011

Volume 45

Number 3

17

FIGURE 5
Picture of power spectra for Antenna 3 of the SeaSonde at 00:15 GMT on February 26, 2009. The
x axis is Doppler shift (Hz), the y axis is signal strength (dB), and the z axis denotes the range bin from
the radar (scalar). Vessel echoes are observed between the two sea-echo Bragg peaks at approximately 0.4 Hz.

to the laboratory for ofine processing.

The user sets three combinations of
integration time and threshold for
each of the two backgrounds. Hence,
the output of one run of the range
data through the ship detection code
results in six concurrent data streams
of possible ship detections. Because
the ship detection code only outputs
six packages (three for the IIR method
and three for the Median method) for
each software run, the software is run
several times to ll out our desired
processing matrix. The threshold
(dB)/FFT points that were initially
tried using the IIR background were
6/16, 7/32, 8/64, 9/128, 10/256,
and 11/512. The threshold (dB)/FFT
points that were tried using the
Median background were 8/32, 9/64,
10/128, 11/256, 12/512, and 13/1024.
The threshold was increased because
the average of signal to noise increased
by 1 dB for each doubling of the FFT
length (Roarty et al., 2010a). A plot of
range (km), range rate (m/s), and bearing (degrees clockwise from north) of
possible detections using the IIR background, 256-point FFT, and 10-dB
threshold is shown in Figure 6. The
trails of vessels can be seen in the
range and range rate subplots. There
are also false positives in the data

18

stream as noted by the single detections with no adjacent detections in

space or time.

The AIS data were then used to see

how the radar was performing when it
came to detecting the speed and location of vessels at sea. The AIS data were
rst ltered by time to coincide with
the measurements, then geographic
proximity to the radar site (a 60-km
threshold was used) and then binned
by ship identication. Eighteen ships
passed this rst stage of ltering.
Then any ship that was located on
the bay side of the radar or had zero
radial velocity was removed. This left
four ships for possible detection, one
tug boat (the Dolphin), two cargo containers (the Maas Trader and the MOL
Efciency), and a tanker (the Joelmare).
The latitude, longitude, and time

FIGURE 6
Plot of target detections from 00:10 to 00:55 GMT on February 26, 2009. Panels from top to bottom are range, range rate, and bearing. The yellow horizontal lines in the middle are the expected
positions of the very strong Bragg sea clutter echoes that would mask vessel detection.

Marine Technology Society Journal

reports of these four ships were used to

calculate the ship range, radial velocity,
and bearing relative to the radar at Sea
Bright, New Jersey. A text le of ship
position and time was processed using
the program GPSTracker, which is
part of the SeaSonde detection software
package. This software is normally
used to perform the same task when
measuring the antenna pattern of the
receive antenna with a transponder
on a vessel. This software generates a
le for each vessel that contained the
range, radial velocity, and bearing
of that particular vessel as shown in
Figure 7. These data were then used
for comparison with the range, radial
velocity, and bearing calculations
from the radar.
The next step was to compare the
data from the radar with data obtained
via AIS. If the calculated range is
within half the width of a range bin
(1.5 km in this case) and within two
Doppler bins (varied between 0.02
and 1.4 m/s, depending on the length
of the FFT window) of the actual vessel, then the detection is considered a
hit. A detection rate is then calculated
as the number of times the radar detected the target divided by the total
number of time sample possibilities,
which was every 32 s for this experiment. The detection rates for Joelmare
using the IIR and Median background
are shown in Tables 2 and 3, respectively. The detection rates for Maas
Trader using the IIR and Median
background are shown in Tables 4
and 5, respectively. The detection
rates for the Dolphin using the IIR
and Median background are given in
Tables 6 and 7, respectively. In the
case of the Maas Trader, once the highest detection rate was found for the
initial runs, then the thresholds were
varied along the FFT length with the
highest detection rate. The detection

FIGURE 7
Range, range rate, and bearing plot of four vessels in the vicinity of the Sea Bright HF radar
station on February 26, 2009 as measured by AIS.

TABLE 2
Detection rate in percent for the Joelmare with different combinations of threshold (columns) and
FFT points (rows) using the IIR background.
6 dB
16

7 dB

8 dB

9 dB

10 dB

11 dB

NSD

32

NSD

64

0.6

128

15.6

256

13.2

512

10.8

NSD stands for no ship detected.

TABLE 3
Detection rate for the Joelmare with different combinations of threshold (columns) and FFT
points (rows) using the Median background.
8 dB
32
64
128
256

9 dB

10 dB

11 dB

12 dB

13 dB

0.2
3.3
34.0
33.3

512

12.5

1024

7.6

May/June 2011

Volume 45

Number 3

19

TABLE 4
Detection rate in percent for the Maas Trader with different combinations of threshold (columns) and
FFT points (rows) using the IIR background.
6 dB
16

7 dB

8 dB

10 dB

11 dB

64.9

64.9

NSD

32

NSD

64

NSD

128
256

9 dB

58.7
70.3

70.3

68.9

66.2

512

38.5

NSD stands for no ship detected.

TABLE 5
Detection rate for the Maas Trader with different combinations of threshold (columns) and FFT
points (rows) using the Median background.
8 dB
32

9 dB

11 dB

12 dB

13 dB

61.3

58.8

58.8

13.8

64

23.6

128
256

10 dB

55.0
65.0

65.0

62.5

512

46.5

1024

30.8

Detection rate for the Tugboat Dolphin with different combinations of threshold (columns) and
FFT points (rows) using the IIR background.
6 dB

32

7 dB

where is the range, range rate, or

bearing bin size and SNR 3 is the
signal-to-noise ratio on Antenna 3
(the monopole). This gure is characteristic of the uncertainty provided
by the SeaSonde HF radar that determines bearing using direction nding,
i.e., low uncertainty in range and range
rate measurements and higher uncertainty for the bearing estimate.

8 dB

9 dB

10 dB

11 dB

We will now discuss detection results of the radar with four vessels
that passed in front of the radar during
the test period.

NSD

Cargo Container MOL Efciency

0.5

64

5.4

128

92.4

256

73.1

512

60.0

NSD stands for no ship detected.

20

pffiffiffiffiffiffiffiffiffiffiffi ;
SNR 3

Discussion

TABLE 6

16

rate for the MOL Efciency was not calculated due to a short AIS record, but
there are results that will be discussed
in the next section.
An example of the association of
the detections with the ground truth
data is shown in Figure 8. This gure
shows the detections made by the radar
associated with the GPS position of
the cargo ship the Maas Trader from
Figure 6. The detections are shown
as blue diamonds with error boxes
around the detection, with half the
height signifying the uncertainty of
the measurement and the width of
the box denoting the length of the
FFT window. The uncertainty is
given by the equation

Marine Technology Society Journal

The MOL Efciency, International

Maritime Organization (IMO) Ship
No. 9251365, is a cargo container
with a length of 294 m and a beam
of 32 m. The vessel was exiting New
York Harbor on a southeast course

TABLE 7
Detection rate for the Tugboat Dolphin with different combinations of threshold (columns) and
FFT points (rows) using the Median background.
8 dB
32
64
128

9 dB

10 dB

11 dB

12 dB

13 dB

10.8
31.0
93.2

256
512
1024

78.6
56.2

Did the operators of the vessel stop

transmitting the signal, or were the receivers unable to record the signal? The
radar was able to make detections on
the vessel coincidental with the AIS
data. If we assume that the vessel maintained its course and the radial velocity
maintained its rate of change, then the
radar did indeed make the additional
detections of the vessel as clearly
shown in Figure 9.

32.8

Tanker Joelmare
and was approximately 15 km from
the radar. The AIS record of the
MOL Efciency only spans from

00:00 to 00:23 GMT on February

26, 2009. It is unclear to the authors
as to why the AIS record terminated.

FIGURE 8
Plot of target detections (blue dots) and corresponding uncertainty values (blue squares) associated with GPS track of the Maas Trader (solid aqua line). The panels from top to bottom are
range (km), radial velocity (m/s), and bearing (degrees clockwise north). The uncertainty values
for each measurement are shown as the height of each blue box; the length of the FFT is the
width of the box. (Color versions of gures available online at: http://www.ingentaconnect.com/
content/mts/mtsj/2011/00000045/00000003.)

The Joelmare, IMO Ship No.

9288019, is a tanker with a length of
228 m and a beam of 32 m. The tanker
Joelmare was exiting New York Harbor
but then turned around and headed
back into the harbor. The radar was
able to detect the vessel 34% of the
time with the Median background
and 16% of the time using the IIR
background. We propose two explanations as to why the radar did
not detect this vessel very well. First,
from Figure 7: The radial velocity of
the vessel was noisiest of all the vessels.
This lack of a constant radial velocity
would spread the energy of the returned signal over several Doppler
bins. This would cause the signal to
not be detected, because its amplitude
falls below the threshold set in the software. The second reason as to why the
radar did not detect this vessel was
that the vessel was north-northeast of
the radar site, and the signal had to
propagate over large sections of land
that attenuated the signal in those
directions.

Cargo Container Maas Trader

The Maas Trader, IMO Ship No.
9308625, has a length of 139 m, a
beam of 23 m, and a gross tonnage of
9981. More particulars on this vessel as
well as the Dolphin are given in Table 8.
May/June 2011

Volume 45

Number 3

21

FIGURE 9
Plot of target detections for range, range rate, and bearing. The solid aqua line from 00:10 to
00:23 GMT shows the path of the MOL Efciency from the AIS signal. There are additional detections
past 00:23 on the gure, but the AIS data were not available to compare with the radar detections.
(Color versions of gures available online at: http://www.ingentaconnect.com/content/mts/mtsj/
2011/00000045/00000003.)

TABLE 8
Characteristics of vessels detected in this study.

The authors were unable to nd the

height above the water for any of the
vessels detected. The cargo container
Maas Trader was heading south in
the Barnegat shipping lane for this
test. The radar at Sea Bright had the
best detection rate of 70% using the
IIR background and 65% with the Median background. Both of these cases
occurred with the lowest threshold
and yielded the highest number of
false positives. It would be up to a potential user of the detection les to determine where the threshold should be
set. If the detection les were to be
passed on to a tracker, the performance
of the data in the tracker could help
determine what the threshold level
should be. Another option is to run
all combinations of FFT length and
threshold and let a tracker determine
which detections are authentic and
which ones are false. The radar was
not able to detect the vessel at 00:40.
This was due to the fact that the vessel
was crossing through the zero Doppler
area, where there are large echoes from
stationary objects. A plot of the FFT
length and threshold combinations
versus detection rate, which is a summary of Tables 4 and 5 for the Maas
Trader test case, is given in Figure 10.
A peak in the detection rate is found
between the 128- and 256-point
FFT. With a 2-Hz sweep, this converts
to a 1- to 2-min averaging period as
optimal for the detection of these vessels with the HF radar.

Vessel Characteristic

Maas Trader

Dolphin

MOL Efciency

Joelmare

MMSI No.

237956000

366920980

351166000

477738400

IMO No.

9308625

7319010

9251365

9288019

Type

Cargo

Tug

Cargo

Tanker

Length (m)

139

41

294

228

Beam (m)

23

32

32

Hull type

Single

Single

NA

NA

Tugboat Dolphin

Gross tonnage

9981

198

NA

NA

Depth (m)

11.8

NA

NA

Draught (m)

4.25

12.1

6.6

Freeboard (m)

0.75

NA

NA

Hull material

Steel

Steel

NA

NA

The Dolphin, IMO Ship No.

7319010, has a length of 41 m, a
beam of 10 m, and a gross tonnage of
198. The tugboat Dolphin was heading
north into New York Harbor inside of
the Barnegat shipping lane. Figure 4
shows that this is a heavily transited

NA stands for not available.

22

Marine Technology Society Journal

Summary

FIGURE 10
Response of detection rate to the variance of FFT length and threshold level.

route as well. The Sea Bright radar had

the highest detection rate on this vessel
with a 92% using the IIR background
and 93% using the Median background. The vessel had a large

superstructure, which made it an

ideal target for the radar. The best
case detections from the Dolphin and
Maas Trader were placed on a map in
Figure 11.

FIGURE 11
Detections of the Maas Trader (IIR background, 256-point FFT, and 10-dB threshold) and
Dolphin (median background, 256-point FFT, and 10-dB threshold) overlaid on the path of
the vessel from GPS/AIS.

A case study has been performed

using a SeaSonde HF radar to detect
vessels at sea in a dual-use mode. The
selected 13-MHz HF radar that operates within the MARCOOS and
provides radial current data to the
NOAA National HF Radar Network
simultaneously detected the speed
and location of several ships at sea.
The detections made with the HF
radar were checked against the GPS
position of the target sent via the AIS
system. An optimal integration time
for this type of radar with this class
of vessel is between 1 and 2 min. The
detection rates for some vessels were
above 90%. Lower thresholds resulted
in higher detection rates but also led to
higher false alarm rates. Overall, the
median background performed better
than the IIR background, but there
were instances where the IIR background was the best.
One benet of HF systems for
vessel detection/tracking is that they
provide over-the-horizon detection
capability (Khan et al., 1994). The systems being developed and evaluated
within the 5-MHz and 13-MHz
bands regularly see vessels between
50 and 110 km. However, the focus
of this paper was on the optimal processing parameters that would enable
the best detections. Because of this,
the authors focused on vessels that
were close to shore and would provide
the most signal to test these parameters
without introducing other parameters
from the radar equation. The authors
will, in future work, use these optimal
parameters to test the range limits of
the vessel detection system as a separate study.
A signicant nding was that the
same targets were seen by the detection
algorithm simultaneously, at different

May/June 2011

Volume 45

Number 3

23

FFT/coherent-integration times, with

different thresholding and backgrounds. Thus, one is not forced to
preselect a xed processing parameter suite. A properly optimized association algorithm (which is under
development) would search for detections of the same vessel among
all of the output combinations and
thereby yield a much improved detection. This would increase the
detection rate seen by an individual
look of 90%, for example, to perhaps
98%, while reducing the overall false
alarm rate.
This offers the opportunity to convert the National HF Radar Network
into a dual-use system that would
provide surface currents to the U.S.
Coast Guard as well as provide wide
area surveillance in the maritime domain to the Department of Homeland
Security.

Acknowledgment
This material is based on work
supported by the U.S. Department
of Homeland Security under Award
No. 2008-ST-061-ML0001.
The views and conclusions contained in this document are those of
the authors and should not be interpreted as necessarily representing the
ofcial policies, either expressed or
implied, of the U.S. Department of
Homeland Security.

Lead Author:
Hugh J. Roarty
Coastal Ocean
Observation Laboratory
Institute of Marine and
Coastal Sciences
Rutgers University
New Brunswick, NJ 08901
Email: [email protected]

24

References
Barrick, D. 1972. First-order theory and
analysis of MF/HF/VHF scatter from the sea.
IEEE Trans Antennas Propag. 20(1):2-10.
doi: 10.1109/TAP.1972.1140123.
Chant, R.J., Glenn, S.M., Hunter, E.,
Kohut, J., Chen, R.F., Houghton, R.W.,
Schoeld, O. 2008. Bulge formation of a
buoyant river outow. J Geophys Res.
113(C1):C01017.
Department of Homeland Security Science
and Technology. 2009. High-Priority Technology Needs. Washington, D.C. 24 pp.
Available at: http://www.dhs.gov/xlibrary/
assets/st_high_priority_technology_needs.pdf.
Glenn, S.M., & Schoeld, O. 2009. Growing
a distributed ocean observatory: Our view
from the COOL room. Oceanography.
22(2):128-45.
Glenn, S.M., & Schoeld, O.M.E. 2002.
The New Jersey shelf observing system.
OCEANS 02 MTS/IEEE. 3:1680-7.
Gong, D., Kohut, J.T., & Glenn, S.M.
2009. Seasonal climatology of wind-driven
circulation on the New Jersey shelf. J Geophys
Res. 109(C7):C07S07.
Interagency Working Group on Ocean
Observation. 2009. A Plan to Meet the
Nations Needs for Surface Current Mapping.
Washington, D.C. 64 pp.
International Maritime Organization. 2006.
Automatic Identication Systems. London:
IMO Publishing. 102 pp. Available at: http://
www.ioos.gov/library/surfacecurrentplan9_
3lowres.pdf.
Khan, R., Gamberg, B., Power, D., Walsh, J.,
Dawe, B., Pearson, W., & Millan, D. 1994.
Target detection and tracking with a high
frequency ground wave radar. Journal of
Oceanic Engineering. 19(4):540-8.
Kohut, J.T., & Glenn, S.M. 2003. Improving
HF radar surface current measurements
with measured antenna beam patterns.
J Atmos Oceanic Technol. 20(9):1303-16.
doi: 10.1175/1520-0426(2003)020<1303:
IHRSCM>2.0.CO;2.

Marine Technology Society Journal

Roarty, H.J., Barrick, D.E., Kohut, J.T., &

Glenn, S.M. 2010a. Dual-use compact HF
radars for the detection of mid- and largesize vessels. Turk J Electr Eng Comput Sci.
18(3):373-88.
Roarty, H.J., Glenn, S.M., Kohut, J.T.,
Gong, D., Handel, E., Rivera Lemus, E.,
Seim, H. 2010b. Operation and application
of a regional high-frequency radar network in
the Mid-Atlantic Bight. Mar Technol Soc J.
44(6):133-45.
Rossby, T., & Gottlieb, E. 1998. The
Oleander Project: Monitoring the variability
of the gulf stream and adjacent waters between New Jersey and Bermuda. Bull Am
Meteorol Soc. 79:5-18. doi: 10.1175/15200477(1998)079<0005:TOPMTV>2.0.CO;2.
Ruck, G.T., Barrick, D.E., Stuart, W.D.,
& Krichbaum, C.K. 1970. Radar Cross
Section Handbook. New York: Plenum Press.
949 pp.
Temll, E., Otero, M., Hazard, L., Conlee, D.,
Harlan, J., Kohut, J., Lindquist, K.
2006. Data management and real-time
distribution in the HF-radar national network.
In: OCEANS 2006 MTS/IEEE Proceedings,
pp. 1-6. Boston, MA. doi: 10.1109/
OCEANS.2006.306883.