193
1 INTRODUCTION
InitialworkforFOliage PENetration (FOPEN) radar
systemsdatedbacktothelate‐1960tomid‐1970with
meagre results due to foliage attenuation that limit
thesystemstoshort‐to‐medium‐rangeoperation and
mannedaircraftcouldnotbeadequatelyprotectedat
those ranges. Later the development
of wideband
data links would enable significant processing and
imageinterpretationonthegroundtillthelate1980s
when the image collection community determined
that Synthetic Aperture Radar (SAR) could provide
acceptable and useful detection and characterization
offorestedregions,[1].
So far studies as been performed mainly using
aerial
platformequippedwithSARbutnowthefocus
is on ground based sensor systems able to detect
walking personnel and moving vehicles with very
low false alarm probability even in presence of
unfavourableweatherconditions(rain,wind)and/or
localseasonalfauna.FOPENsensorarecharacterized
byusinglowfrequencies,generallyU
‐VHFbands(30‐
1000 MHz) suitable for radar operation in dense
foliage environment where –on the contrary‐
traditional microwave radars in X, Ku bands
(normally used for border control) suffer strongly
form foliage attenuation and backscatter, [1] – [2].
FOPEN ground radar can be efficiently utilized in
nonheterogeneousenvironment
too.Forinstancefor
coastsurveillance in presence of vegetation near the
water.
2 FOPENPHENOMENOLOGY
FOPENapplicationinvolvesafundamentaltrade‐off
betweenresolutionandfoliagepenetrationcapability:
Surveillance Unattended Foliage Penetrating Radar
for Border Control and Homeland Protection
F.Amato,A.Farina,M.Fiorini&S.Gallone
SelexES–AFinmeccanicaCompany,Rome,Italy
ABSTRACT:Theincreasingrequestforsafety,securityandenvironmentprotectionatlocalandnationallevel
revealthedeficiencyofthetraditionalsurveillanceandcontrolcenterstosatisfytheneedsandrequirementsof
modernbordercontrolsystemsforhomelandprotectionwherelandborderisexpected
tobemonitoredaswell
asthemaritimeone.Thisis,forinstance,thecaseofanylandborderaffectedbyhiddenimmigrationand/or
illegaltrafficsaswellasanysmallareassuchascriticalinfrastructures ormilitary/civilianpostsinforestor
jungleenvironmentcharacterizedbyvegetation.Insuch
challengingenvironment,logisticsconstraintsstrongly
recommend to have very low power devices able to operate months or years without maintenance. A such
scenarioshouldbetheperfectplaceforimplementinganUnattendedGroundSensors(UGS)networkmaking
useFOliagePENetration(FOPEN)radarforbordercontrol.Thepaperaimstopresentthe
basiccharacteristics
andpreliminaryresultsofaSurveillanceUnattendedFOPEN(SUF)radarsuitablefordetectingmovingtargets,
peopleorvehicles,indenseforestenvironment.
http://www.transnav.eu
the International Journal
on Marine Navigation
and Safety of Sea Transportation
Volume 7
Number 2
June 2013
DOI:10.12716/1001.07.02.05
194
high‐resolutiondemandsahighcenterfrequency,but
penetration of foliage demands a wavelength long
enough to propagate through tree cover. FOPEN
radars have typically operated at VHF or UHF. At
higher frequencies, propagation through foliage
introducesmoresevereimagingeffects.Theeffectsof
propagationthroughfoliagecanbebroadly
separated
intothreecategories:phaseshift,thebackscatterand
theattenuation.Eachoftheseeffectscontributestoa
limitation in the radar ability to detect objects in
foliageenvironment.Manystudieshavebeencarried
outforSARapplication,[3]–[6].
Randomphaseshift:phasevariationistherandom
variationinthesignalphasearisingfrompropagation
through a distributed, non‐uniform medium (i.e., a
foliage canopy). Phase shift would de‐correlate the
radar returns, the more the phase is corrupted, the
lessisthecoherentprocessinggainandthereforethe
probability of detecting a surface target. An
additional motivation
for lower‐frequency operation
isrelatedtotheimpactofphasevariationthatismore
marked at higher frequencies (Phase shift (f) =
Frequency(MHz)*0.133).
Forest Backscatter: backscatter is the reflection of
transmittedenergybacktothesensorbyinteractions
with single or multiple foliage elements or by
interaction between these elements and the ground.
Consider a radar spatial resolution cell containing
windblown trees, such a cell contains both fixed
scatterers (ground, rocks, tree trunks) and moving
scatterers (leaves, branches). The returned signal
correspondinglycontains bothaconstant(or steady)
and a varying component. The steady component
givesrisetoa DCorzero‐Dopplerterminthepower
spectrum of the returned signal, and the varying
componentgivesrisetoanACterminthespectrum.
Thusasuitablegeneralanalyticrepresentation
forthe
total spectral power density Ptot(v) in the Doppler‐
velocity power spectrum from a cell containing
windblownvegetationisprovidedby:
  
1
,
11
tot ac
r
Pv v Pv v
rr


wherevisDopplervelocityinm/s,ristheratioofdc
powertoacpowerinthespectrum,δ(v)istheDirac
deltafunction,whichproperlyrepresentstheshapeof
the dc component in the spectrum, and P
ac(v)
represents the shape of the ac component of the
spectrum proven [5] to decay at rates close to
exponential:


exp ,
2
ac
Pv v v

whereβistheexponentialshapeparameterthatisa
function of wind conditions and is largely
independentofradarcarrierfrequencyovertherange
from VHF to X‐band. Foliage backscattering is more
pronounced at high depression angles.
Thefixedclutterreturnscanhaveadccomponent
(zeroDoppler)raisingupto60‐70dBabovethenoise
level.
Doppler spectra, in order to perform efficient
clutterrejection,twovaluesofthresholdscanbeused:
i.e.1 m/s in case of light air (wind speed: 1‐7 mph),
2m/sincaseofwindy/gale(windspeed:15‐60mph).
Foliage Attenuation: propagation through foliage
leads to attenuation of the radar signal in part by
absorptionandinpartbythescatteringoftransmitted
energy away from the target and sensor. Foliage
attenuation increases significantly with frequency.
Two‐wayHH polarizationsignalattenuation
reportedin[6],[7]fora30°depressionangleincreases
from5.5dB
atUHFto17.0dBatLbandandto33.6
dBatCband.Foliage attenuation tend to be more severe
at smaller depression angles. This is due primarily to the
increase in foliage path length as depression angle
decreases [6], [7]. Foliage attenuation exhibits a slight
dependence on polarization. In particular, attenuation tends
to be slightly larger for VV polarization than for HH
polarization .This is especially noticeable at lower
frequencies [7], at which attenuation is primarily driven not
by leaves and branches, but instead by tree trunks, most of
which are vertically oriented.
3 GROUNDSURVEILLANCEUNATTENDED
FOPENSYSTEM
3.1 SYSTEMdescription
The surveillance of critical perimeters is one of the
most important issues in Homeland security and
protection systems. Ground surveillance needs
are
relevant across multiple scales, from border
protection applications (hidden immigration, illegal
traffic,narcotraffic) tosmallareasprotection (critical
infrastructure, military/civilian posts). Furthermore,
thisinfras tructurecanbefixedormobile.Thesecurity
and protection systems must be able to provide full
coverage continuously in a variety of cluttered
environments, such as
forest or jungle domains.
However,mostexistingsystemshavebeendeveloped
using airborne SAR and are not suitable for 24h
operations.
In this article we propone a land surveillance
system based on ground sensors, eventually
interoperatingwithairbornesensors,withcapabilities
todetectwalkingpersonnelandmovingvehicles.The
proposed
system architecture provides the
capabilitiesof:
detection, localization, tracking and recognition
of people and vehicles irregularly entering in a
forestedareaoflandborders.
adaption of the system configuration and
deployment to optimize performance in response
tochangingenvironmentalconditions.
multilayer data fusion and system operating
capability
by providing the situation awareness
and control to prevent and manage suspicious
behavior.
EasytouseandlowcostsolutionswithLPI(Low
Probability Intercept) capabilities to not be
detectedbeforethetargetisdetected.
Multi‐scale Common Picture capability which
will provide different pictures of the region of
interest with different fields of view at different
resolutionsandtimescales.
195
The proposed system architecture is depicted in
Figure1.Theproposedsystemiscomposedofasetof
subnetsthataregeographicallydistributedalongthe
boundariesofawideareatobecontrolled,suchasthe
border of a nation. The subnets are composed of
homogeneous sensors connected with wireless
or
wirelinks.Eachsubnetensurestheexchangeofdata
betweenlocalclustersofsensors.TheFigure1shows
thesubnetscomposedoftwo typesofFopenradars:
Unattended Ground Sensor (UGS) and Fixed Tower
Ground Radar (FTGR) that will be described in the
section 3.3. However, the type and
configuration of
thesensorstobeemployedintheotherssubnetscan
be selected on the basis of the characteristics of the
siteunderconsideration.Thesystemarchitecturehas
theadvantagetobemodular andscalable anditcan
be organized with different level C2 centers (local,
regional,national),dependingalso
onthesizeofthe
consideredboundaries.
UGSsubnet
FTGRsubnet
Othersensors
subnets
...
Figure1. System architecture of GROUND
SURVEILLANCEUNATTENDEDFOPENSYSTEM
Figure2. Multi‐scale approach for hierarchical architecture
ofSystemtoprovideCommonOperationalpicture.
Thesizeoftheregion,thenatureoftheborderand
thecomplexityofthescenariorequiretheprovisionof
differentpicturesoftheregionwithdifferentfieldof
view at different resolution and time scales,
suggesting a multi‐sensor/multi‐scale approach
integratedinahierarchicalarchitectureofthe whole
system,anexampleisshowninFigure2.
Typicallyaglobalfieldofviewofthewholeregion
isnecessaryatthehigherCommandandControl(C2)
leveltocapturetheoverallsituation. Ahigherlevelof
resolution and refresh rate is necessary at the lower
andlocallevel to analyze
and control indepth each
singlezoneofaregion.
Therefore the surveillance segment may be
structured according to a multilayer architecture
where layers realize different trade‐offs in terms of
field of view and granularity and refresh time. All
data collected by the sensors are exploited by the
fusion
engine, [8]. It is responsible to track and
classifyrelevantentitiespresentinthescenarioandto
provideahighqualityrepresentationofthesituation.
Cameras can also be used to this end as they are
usually fully integrated with the rest of the system
andcouldbepresentedontop
ofthecartographyon
theoperatorconsole,cf.Figure3.
Figure3. Cameras presentation on operator console, live
data.
3.2 SubnetDescription
Each subnet ensures the exchange of data between
localclustersofsensors.In0thearchitectureofaUGS
subnetisshown.EachnodeisaFOPENradarsensor
with a very small coverage region (purple cyrcles).
The typical detection range of a single sensor is 100
meters.The
surveillanceperimeterofasubnetcanbe
extended up to several kilometers by deploying a
fixednumberofsensors(eg.50‐150).Thetargetcanbe
detectedbymorethanonesensor,inordertoprovide
themultistaticcoordinates.
Figure4.UGSsubnetconcept
196
Adjacentsensornodesareconnectedtogetherviaa
low power RF link (blue arrows). Each sensor
forwards the information to the nearest sensors (to
assurealternativepathsinthecaseoffault)andinthe
endtheinformationissenttoamasterstation,viathe
short range radio
link. The master station performs
data fusion and medium range connection with the
othermasterstations,ortheC2centre.
3.3 FOPENRADARDESCRIPTION
Logisticsconstraintsdrivethetechnologytoverylow
power devices, that are able to operate for several
months or years, without maintenance. Another
importantissueis,together
withagoodprobabilityof
detection, the low false alarm probability that is
requestedtobeloweredupto1falsealarmperday,
orlower,eveninpresenceofbadweatherconditions
(rain,wind)and/orlocalseasonalfauna.
The main requirements/constraints addressed are
therangeof the detections, which
isreducedby the
attenuationduetofoliageandthelowantennaheight,
that is usually limited to 1‐2 meters for logistic
purposes. Moreover, logistic constraints drive the
technology to very low power devices. Considering
thatphotovoltaiccellsarenotsuitableforinstallation
onthegroundintheforestand
thattheradarsmust
be able to operate for several months or years,
without maintenance, power consumption must be
keptatminimumlevel,andtheemittedpowermust
be kept at a level of several mW. Camouflage and
anti‐tamperareoftenrequired,and,sincethenumber
of displayed sensors
can be high (50‐150 for each
subnet)verylowcostisamandatoryrequirement.
Despitethelowcost,theperformanceoftheradars
mustbegoodenoughtodetectwithhighprobability
walkingpersonnelandmovingvehicles,withaLow
Probability of Intercept (LPI) and a low false alarm
probability,
less than 1 false alarm per day, even in
presence of bad weather conditions (rain, wind)
and/orlocalseasonalfauna.
Figure5.Range‐Dopplemeasurementinurbanscenario
Experimentalresultinurbanheavytrafficscenario
has been conducted as a preliminary analysis to
validate the prototype sensors. Results were
encouragingandhasbeendemonstratedresolutionof
6mforpersonandlessthen30mforcars,cf.Figure5.
Inthisarticleweproposeaninnovativegroundbased
FOPEN
UHF/VHF radar family composed by the
followingtypes:
UGR (Unattended Ground Radar): FMCW radar,
with an advanced digital processing that have a
LowProbabilityofIntercept(LPI)capabilitiesand
a minimum power consumption. The emitted
powerisofseveralmW.UGRaredeployed with
an antenna height that is
usually limited to 1‐2
meters for logistic purposes in a forested
environment.Camouflageandanti‐tampercanbe
satisfied. The typical detection range of a single
sensor is 50‐100 meters, depending on the
environment.
FTGR(FixedTowerGroundRadar):FMCWradar,
with an advanced digital processing. FTGR are
deployedonmediumheighttower(eg25m). The
emitted power is in the order of 1W. The FTGR
requiresalowpowersupplyandcanbepowered
by photovoltaic cells. The typical detection range
of a single sensor is 1‐5 km, depending on the
environment.
4 CONCLUSIONAND
FUTUREWORKS
Intheeraofbudgetconstraintsandtimepressurewe
are living nowadays the requests for low‐power,
unattended border control systems are increasing.
The technology progress make possible to integrate
UGS and FTGR in different system solutions and
combinations according to scenario and users needs
even in demanding
environmental conditions like a
forest.
U‐VHF radar sensors are under developing at
Selex’s premises following preliminary encouraging
resultspartiallypresentedinthiswork.
REFERENCES
[1M. E. Davis “Foliage Penetration Radar: Detection and
Characterization of Objects Under Trees” SciTech Pub.
Inc.,ISBN978‐1‐891121‐00‐5,NewYork,USA,2011
[2]B.T.Binder, M.F. Toups,S.Ayasli, andE.M. Adams,
SAR Foliage Penetration Phenomenology of Tropical Rain
Forest and Northern U.S.
Forest, Proc. IEEE Int. Radar
Conference,1995,pp158–163
[3]M.E. Davis,DevelopmentsInFoliagePenetrationRadar
RadarConference2010IEEE,10‐14May2010pp: 1233,
Washington.
[4]M. N. Roy, S. Swamp, R. K. Tewari, Radio Wave
Propagation Through Rain Forests of India, IEEE
TransactionsonAntennas
andPropagation,vol. 38,no.
4,April1990
[5]J.B.Billingsley,Low‐AngleRadarLandClutter,SciTech.
Pub.Inc.,2001,NewYork
[6]M. E. Davis, Foliage Penetration Radar: Detection and
Characterization of Objects Under Trees, SciTech. Pub.
Inc.,
[7]J. G. Fleischman, S. Ayasli, E. M. Adams, and
D. R.
Gosselin. Part I: foliage attenuation and backscatter
analysis of SAR imagery. IEEE Transactions on
Aerospace and Electronic Systems, 32(1):135‐144,
January1996
[8]Z.Pietrzykowski,J.Uriasz,“KnowledgeRepresentation
in a Ship’s Navigational Decision Support System”,
TransNav – Int. Journal on Marine Navigation and
SafetyofSeaTransportation,ISSN
2083‐6473,ISSN2083‐
6481(electronicversion), Vol.4, No.03‐2010pp. 265‐
270