IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, VOL. 63, 2025 5209816
Accele a ed Algo i hm o Back-P ojec ion
Fully Focused SAR Al ime y
Se gi He nández-Bu gos , Fe an Gibe , Membe , IEEE, An oni B oque as , Li e Membe , IEEE,
Pol Villal illa, Alejand o Egido, Go ka Moyano, Ad ián Flo es de la C uz , and Ma co Fo na i
Abs ac — Fully ocused (FF) syn he ic ape u e ada (SAR)
back-p ojec ion algo i hms o ada al ime y ha e gained
a en ion in he al ime y communi y due o hei imp o ed
azimu h esolu ion compa ed o adi ional me hods, such as
he delay/Dopple algo i hm. Howe e , he signi ican com-
pu a ional cos o back-p ojec ion has limi ed i s use in
ope a ional applica ions. This a icle p esen s an accele a ed
back-p ojec ion algo i hm ha educes he un ime o he classic
back-p ojec ion by a ac o o 28 on a CPU-based a chi ec u e.
When implemen ed on a GPU-based a chi ec u e, he accele a ed
back-p ojec ion achie es a speedup anging om 8 o 13 imes
compa ed o he classic back-p ojec ion on he same GPU.
In compa ison o he classic back-p ojec ion unning on a
CPU, he combina ion o he accele a ed back-p ojec ion and
GPU p ocessing esul s in an o e all imp o emen o up o
1570 imes. The accele a ed algo i hm main ains he accu acy
o classic ime-domain algo i hms and allows o he p ocessing
o da a in he o de o he sensing ime unde a GPU-based
a chi ec u e. The algo i hm has been ex ensi ely alida ed using
ansponde and open ocean da a. These esul s indica e ha
ope a ional back-p ojec ion algo i hms a e easible o cu en
ada al ime y missions, such as Sen inel-6, and upcoming
missions like CRISTAL and Sen inel-3 Nex Gene a ion.
Index Te ms— Accele a ed algo i hm, accele a ed back-
p ojec ion, back-p ojec ion, delay/Dopple , ully ocused (FF),
ully ocused syn he ic ape u e ada (FF-SAR), GPU accele a-
ion, high- esolu ion ada al ime y, omega-K, open ocean, ada
al ime y, eal ime al ime y, syn he ic ape u e ada (SAR),
ansponde .
I. INTRODUCTION
AT THE beginning o space-based al ime y, ada al ime-
e s missions such as he ERS-1 (1991) [1], Jason-1
(2001) [2], and ENVISAT (2002) [3] used con en ional
Recei ed 18 No embe 2024; e ised 21 Feb ua y 2025; accep ed
26 Ma ch 2025. Da e o publica ion 31 Ma ch 2025; da e o cu -
en e sion 29 Ap il 2025. This wo k was suppo ed in pa by
he Indus ial Doc o a es Plan om he Sec e a y o Uni e si ies and
Resea ch o he Depa men o Business and Knowledge o he Go -
e nmen o Ca alonia h ough P ojec SA.45217 wi h Expedien Numbe
094 unde G an Doc o a s Indus ials 2020 and in pa by Eu opean
Space Agency (ESA) h ough he Poseidon-4 G ound P ocesso P o o ype
P ojec unde ESA Con ac 4000112536/14/NL/BJ. (Co esponding au ho :
Se gi He nández-Bu gos.)
Se gi He nández-Bu gos is wi h isa dSAT, S.L., 08005 Ba celona, Ca -
alonia, and also wi h he Signal Theo y and Communica ions Depa men ,
Uni e si a Poli ecnica de Ca alunya, 08034 Ba celona, Ca alonia (e-mail:
[email p o ec ed]).
Fe an Gibe , Pol Villal illa, Go ka Moyano, and Ad ián Flo es de la C uz
a e wi h isa dSAT, S.L., 08005 Ba celona, Ca alonia.
An oni B oque as is wi h he Signal Theo y and Communica ions Depa -
men , Uni e si a Poli ecnica de Ca alunya, 08034 Ba celona, Ca alonia.
Alejand o Egido is wi h Eu opean Space Agency, Eu opean Space Resea ch
and Technology Cen e, 2201 AZ Noo dwijk, The Ne he lands.
Ma co Fo na i is wi h he S a ion G oup o Eu opean Space Agency
(ESA)/ESTEC, 2201 AZ Noo dwijk, The Ne he lands.
Digi al Objec Iden i ie 10.1109/TGRS.2025.3556544
pulse ansmission a low pulse epe i ion equency (PRF),
enabling he obse a ion o Ea h su ace wi h an azimu h
esolu ion limi ed by he an enna oo p in , ypically in he
o de o kilome e s. La e on, ada al ime y missions such
as C yoSa -2 (2010) [4] and Sen inel-3 (2016) [5] we e
based on he ansmission o sho bu s s o pulses a e y
high PRF, app oxima ely 19 kHz in a closed-bu s mode o
64 pulses, assu ing cohe ence among hem and inc easing
he numbe o independen looks ob ained om a single
sca e e . The high PRF oge he wi h he pulse cohe ence
allowed o use o he delay/Dopple algo i hm [6], imp o ing
he azimu h esolu ion up o 300 m. Newe missions, such
as Sen inel-6 [7], employ nea ly con inuous pulse ansmis-
sion a ound 9–10 kHz by in e lea ing he ansmission and
ecep ion o pulses, maximizing he measu emen p ecision
while ensu ing a connec ion be ween da a om p io al ime e
missions, such as Jason se ies [8]. Cu en ly, high- esolu ion
ope a ional p ocesso s a e based on he delay/Dopple
algo i hm.
Egido and Smi h [9] in oduced he ully ocused syn he ic
ape u e ada (FF-SAR) back-p ojec ion algo i hm, which
u he imp o ed he azimu h esolu ion o he heo e ical
maximum, in he o de o subme e . This enhancemen in
azimu h esolu ion has allowed o a educ ion in noise in open
ocean e acked pa ame e s such as sea su ace heigh (SSH)
and signi ican wa e heigh (SWH) [10]. FF-SAR algo i hms
ha e also acili a ed mo e p ecise iden i ica ion o leads and
icebe gs [11], he s udy o swells om al ime y da a [12],
and he es ima ion o inland wa e pa ame e s such as wa e
ele a ion [13] and wa e ex en [14].
Howe e , he signi ican high-compu a ional e o equi ed
by he classic FF-SAR back-p ojec ion algo i hm o apply all
he necessa y phase co ec ions o each along- ack poin
on he su ace has so a impeded i s implemen a ion as
an ope a ional p ocesso . Consequen ly, al e na i e and mo e
e icien FF-SAR algo i hms ha e been de eloped. In 2019,
a nume ical 2-D equency domain algo i hm was p esen ed
o he C yoSa -2 mission [15]. In 2024, a closed- o m 2-D
equency domain algo i hm was in oduced o he Sen inel-
6 mission [16]. None heless, he equency-domain algo i hms
equi e speci ic local condi ions o he app oxima ions o be
e ec i e, such as cons an sa elli e eloci y and cons an PRF.
Mo eo e , hese app oxima ions can educe he quali y pe o -
mance o he esul s compa ed o ime-domain algo i hms such
as he FF SAR back-p ojec ion. These limi a ions mo i a e
he need o in es iga e ways o accele a e he back-p ojec ion
algo i hm.
© 2025 The Au ho s. This wo k is licensed unde a C ea i e Commons A ibu ion 4.0 License.
Fo mo e in o ma ion, see h ps://c ea i ecommons.o g/licenses/by/4.0/
5209816 IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, VOL. 63, 2025
In his a icle, we p esen a a ian o he back-p ojec ion
algo i hm called accele a ed back-p ojec ion. The accele a ed
back-p ojec ion echnique educes he compu a ional cos o
he classic back-p ojec ion algo i hm h ough ce ain app oxi-
ma ions, wi hou signi ican ly comp omising he pe o mance
o e dis ibu ed a ge s. Fo ha , we assume ha he ada
geome y does no change signi ican ly wi h espec o sho
along- ack dis ances on he su ace, and we can app oxima e
his small change using a linea app oxima ion, educing
conside ably he numbe o ope a ions o ocus. This linea
app oxima ion is well known in he SAR ada al ime e com-
muni y, as i has been commonly employed in high- esolu ion
algo i hms such as he delay/Dopple me hod [6] o ob ain
a closed- o m exp ession o azimu h beam o ming a e a
Fou ie T ans o m. This same app oxima ion is b ie ly men-
ioned in [13, Sec. 3.2.1.6]in he con ex o he FF-SAR
back-p ojec ion algo i hm, whe e i is applied and no ed wi h-
ou u he analysis. Mo eo e , as explained in [17] o SAR
imaging, an addi ional speeding-up app oach can be achie ed
by spli ing and a e aging he syn he ic ape u e in small
subape u es, educing conside ably he numbe o ope a ions
needed o ocus he su ace.
Since main aining he accu acy o he classic back-
p ojec ion is essen ial, a alida ion analysis is conduc ed
wi h eal Sen inel-6 da a o unde s and he limi a ions o he
app oxima ions and o ensu e ha he quali y pe o mance o
ime-domain algo i hms is no comp omised: Fi s , poin a ge
scena ios such as ansponde s [18] and co ne e lec o s [19]
allow us o alida e he eal poin a ge esponse (PTR) and
compa e i o he heo e ical PTR. Also, hey allow us o
iden i y he eal limi s o he accele a ed back-p ojec ion. A e
ha , es s o e he open ocean ha e been conduc ed o e alua e
he accu acy and p ecision o he accele a ed algo i hm when
e ie ing SSH, SWH, and σ0. This e alua ion aims o ensu e
ha we imp o e conside ably he un ime pe o mance wi hou
sac i icing he accu acy wi h espec o he classic back-
p ojec ion algo i hm.
Mo eo e , we ha e conduc ed un ime pe o mance es s
using CPU-based p ocesso s and GPU-based p ocesso s. Ou
esul s show ha we can inc ease he un ime pe o mance o
he classic back-p ojec ion by ×28 using CPU, and by ×1570
using GPU a chi ec u e, all while assu ing good quali y pe o -
mance. The speeding ac o s achie ed wi hou comp omising
he geophysical e ie als pa e he way o he implemen a ion
o FF-SAR p oduc s in u u e ope a ional p ocesso s.
II. GEOMETRIC FOUNDATIONS OF THE ALGORITHM
The main accele a ion ac o o he p oposed algo i hm is
based on he s ong linea phase dependence ha is obse ed
along he illumina ion ime o e a b igh poin a ge when
ocusing no a he speci ic loca ion o he poin a ge bu a
a ce ain nea by su ace in he along- ack di ec ion, in he
o de o me e s.
Fig. 1shows he azimu h phase o all he pulses wi hin he
illumina ion ime o e he ansponde in Ga dos [20] o a
Sen inel-6 pass on Ma ch 15, 2024, when ocusing a di e en
dis ances along- ack om he eal loca ion o he poin a ge .
The azimu h phase is de e mined by ob aining he phase o
each ecei ed pulse wi hin he illumina ion ime, speci ically
Fig. 1. Sen inel-6 L1A azimu h phase ( ad) a he ange o maximum powe
du ing 3.4 s o syn he ic ape u e o di e en ocusing dis ances wi h espec
o he ansponde o Ga dos. (Top) Along- ack co ec ed phase. (Bo om)
De ended phase. The slope o he linea phase is a unc ion o he along- ack
dis ance be ween he poin a ge loca ion and he ocusing poin .
Fig. 2. Along- ack phase slope di e ence be ween he Ga dos ansponde
and he heo e ical model. The ansponde phase slope is de e mined h ough
linea i ing.
a he ange sample wi h he maximum powe , a e applying
all he classic back-p ojec ion co ec ions and p io o he
cohe en in eg a ion. As expec ed, when ocusing all he pulses
a he eal poin a ge loca ion, he phase is cons an , which
allows o a ull cons uc i e summa ion o he signal when
cohe en ly combining all he pulses wi hin he illumina ion
ime. Howe e , a any poin wi h a ce ain along- ack dis ance
nea by o he poin a ge , he phase p esen s a cons an slope.
The slope alue is p opo ional o he along- ack dis ance
om he poin a ge .
Indeed, gi en speci ic condi ions, such slope can be de e -
mined in a closed- o m exp ession. In such cases, i is possible
o ocus he su ace poin s a nea by along- ack dis ances in
a as e manne , jus applying a seconda y phase co ec ion
p opo ional o he phase slope (linea e m). Fig. 2shows he
di e ence be ween he phase slope ob ained om Fig. 1by
means o a linea i ing and he closed- o m exp ession. The
maximum e o is below 15 m ad, less han 0.02%.
HERNÁNDEZ-BURGOS e al.: ACCELERATED ALGORITHM FOR BACK-PROJECTION FF SAR ALTIMETRY 5209816
Fig. 3. E olu ion o he along- ack phase om wa e o ms ob ained a ound a
scena io wi h a single poin a ge . I we apply he back-p ojec ion co ec ions
o a su ace poin be o e he poin a ge , namely a poin A, he emaining
phase slope ob ained will be posi i e. I he ocusing poin is a e he
poin a ge (namely a poin B), he emaining phase slope ob ained will
be nega i e. As expec ed, he phase is comple ely la only when ocusing on
he su ace poin whe e he poin a ge is loca ed.
Fig. 3illus a es a quali a i e explana ion in he case
o an isola ed poin a ge . When applying he classic
back-p ojec ion algo i hm o ocus on poin A, he esul ing
azimu h phase exhibi s a linea beha io , p opo ional o he
along- ack dis ance be ween poin A and he a ge . Then,
a e ocusing on poin A, he co ec ed wa e o ms can be
eused by applying an addi ional seconda y (and linea ) phase
co ec ion o e ie e he azimu h phase co esponding o he
poin a ge . Simila ly, he seconda y phase co ec ion can
also be applied o ob ain he phase co esponding o poin
B. In his p ocess, mos o he back-p ojec ion ope a ions a e
pe o med only once o poin A, while ocusing on addi ional
poin s—such as he poin a ge and poin B— equi es only
a seconda y phase co ec ion ela i e o he dis ance om
poin A.
In gene al, by ini ially ocusing he collec ion o pulses on
a subse o su ace poin s in a coa se g id, i is possible o
ob ain ocused wa e o ms o a much ine g id. This app oach
conside ably educes he numbe o ope a ions needed o ocus
he en i e su ace, hus inc easing he compu a ional e iciency
o he back-p ojec ion algo i hm.
The me hod p oposed consis s o wo s eps.
1) A i s s ep aimed a pe o ming he classic back-
p ojec ion ocusing [9] a speci ic su ace poin s, namely
e e ence su ace poin s, which a e ypically sepa a ed in
he o de o 10 m.
TABLE I
SENTINEL-6 POSEIDON-4 ALTIMETER INSTRUMENT PARAMETERS
2) A second s ep whe e he s acks o co ec ed pulses
be o e cohe en summa ion om e e ence su ace
poin s a e used o ocus he wa e o ms a nea by su ace
poin s, namely he seconda y su ace poin s, ob aining
a inal ocusing a poin s sepa a ed ypically be ween
0.5 and 1 m o an al ime e such as Sen inel-6.
Fu he mo e, an addi ional accele a ion can be achie ed by
assuming ha he phase slope is su icien ly small, so he
azimu h phase emains ela i ely unchanged o e a ew pulses
du ing he illumina ion ime. In his case, we can a e age he
ma ix o co ec ed pulses om he e e ence su aces in o
subape u es and s o e his educed ma ix o he seconda y
poin s. This app oach u he dec eases he numbe o ope a-
ions equi ed o ocus on he seconda y poin s.
III. THEORETICAL FRAMEWORK
A. T ansmi ed Pulses
Mos high- ange esolu ion ada s, including al ime e s,
ansmi chi p pulses [21], as desc ibed by he ollowing
equa ion:
s ( )=w ( )cosh2π c −α
2 2i −Tp
2< <Tp
2.(1)
In his equa ion, w ( ) ep esen s he pulse en elope, which
is usually a ec angula window (uni o m ene gy). Mo eo e ,
cis he ca ie equency o he modula ed signal, is he ime
elapsed om he beginning o he pulse, which has a limi ed
du a ion Tp, in he o de o en hs o mic oseconds. The ime
a iable is known as as ime. The e m αis he chi p a e,
which is de ined as he a io o he pulse bandwid h B o he
pulse du a ion: α=B/Tp. The quad a ic e m in he phase
signal is a ep esen a ion o he linea equency modula ion
o he pulse. Table Ishows he Sen inel-6 nominal alues.
A e he ansmission h ough a nadi -poin ing an enna, he
pulse a els o he Ea h’s su ace, whe e pa o he ene gy
is e lec ed and e u ned back o he senso .
B. Recei ed Signal
Fig. 4illus a es a simpli ied geome y o a ada al ime e .
The sa elli e mo es along he azimu h di ec ion (y-axis),
ansmi ing pulses wi h a nadi -poin ing an enna in a nea ly
con inuous way. A single sca e e is pe iodically e lec ing
he ansmi ed pulses du ing a limi ed illumina ion ime Ti.
The slow ime η e e s o he ime ela i e o he posi ion o he
sa elli e o each emi ed pulse. The o de o magni ude o he
5209816 IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, VOL. 63, 2025
Fig. 4. Simpli ied ada geome y scheme. The sa elli e is mo ing and
cons an ly ansmi ing pulses owa d he nadi . Pa o he ene gy o he
ansmi ed pulses is e lec ed by he Ea h’s su ace and e u ned back o he
sa elli e an enna. I a seconda y poin ysec is close enough o y e , hei ange
mig a ion di e ence R(η, y e )−R(η, ysec)can be app oxima ed by a linea
amp.
slow ime is seconds, while he sampling a e is milliseconds,
depending on he PRI.
A e ecep ion, in mode n ada al ime e s such as Sen inel-
6, he signal is i s digi ized and hen ma ched- il e ed.
The esul ing wa e o ms o a sca e e loca ed a ycan be
exp essed as ollows [16]:
S (η, ,y)
=wη(η, y)W ( )
·exp
j2π c
2
cR(η, y)−2
c(R(η, y)−R k)− d(η, y)
α
(2)
whe e wη e e s o he azimu h en elope, de ined by he
an enna pa e n. The ange en elope is now de ined in e-
quency domain W ( ), al e ed by geophysical e ec s. The
equency ange is he equency a iable in he ange
domain. The ac o d(η) =2 c /cindica es he al e a ion
in he ange o he phase due o he Dopple e ec ela ed
o he ela i e eloci y o he sa elli e . The acke ange
R k is commanded by he sa elli e onboa d. In he case o
Sen inel-6, he acke ange e e s o he cen al poin wi hin
he ange window. The e m R(η) is he ange o he sa elli e
o he a ge . The i s e m in he exponen ial is called ela i e
ange phase (RRP) and ep esen s he phase o a ion due o
he ange mig a ion. The second e m is he CW phase, whose
equency is p opo ional o he delay om he sa elli e o
he sca e e . To ocus all wa e o ms om he same sca e e ,
he back-p ojec ion algo i hm compensa es all phase e ms in
unc ion o he ange be ween he sca e e and he sa elli e,
and hen in eg a es all co ec ed echoes, achie ing maximum
cohe ence om a single sca e e on he su ace. This has o
be done o each de ined poin on he su ace, which esul s
in a high-compu a ional e o .
C. Rada Geome y Simpli ica ions
To de e mine a closed- o m exp ession o he ange mig a-
ion di e ence be ween wo poin s on he su ace, we mus
assume ha he dis ance be ween a e e ence along- ack poin
y e and a seconda y along- ack poin ysec is su icien ly small
o ensu e he alidi y o he geome ic p inciples ou lined
in Sec ion II. Unde hese igh condi ions, we can assume
some ypical geome y simpli ica ions (see [15],[16],[22]):
he Ea h is locally sphe ical wi h a adius Re, he sa elli e
mo es along he ack wi h a cons an eloci y s, and he
di e ence be ween he sa elli e al i ude zsa and he acke
ange R k is ze o.
D. Range Mig a ion Di e ence Be ween Two Sca e e Poin s
To ob ain a closed- o m exp ession o he ange mig a-
ion di e ence be ween sca e e s, we app oxima e he ange
equa ion o a quad a ic polynomial. I is impo an o ema k
ha all phase co ec ions applied o a e e ence sca e e
a e made using eal sa elli e da a, and we do no need o
assume any geome ical app oxima ion, as explained in [9].
On he o he hand, he azimu h phase co ec ion applied
o a seconda y sca e e is designed unde he quad a ic
app oxima ion. The e o e, wha we a e eally app oxima ing
is no he ange equa ion ela i e o a single sca e e , bu
he ange mig a ion di e ence be ween wo sca e e s. Fo a
poin loca ed a he along- ack posi ion y, he quad a ic ange
equa ion can be de ined as ollows:
R(η, y)=R0+ zη−y
g+ 2
eq
2R0η−y
g2
(3)
whe e R0is he ange a he poin o closes app oach (POCA),
zis he e ical mo ion (o hogonal o he azimu h eloci y),
and eq is he equi alen speed de ined as ollows (see [22]):
eq =√ s· g= s·sR0
R0+Re
(4)
whe e Reis he Ea h adius and gis he eloci y o he
pla o m p ojec ed on he g ound.
We can compu e he ange mig a ion di e ence be ween a
e e ence poin y e and any along- ack poin yas a unc ion
o he azimu h dis ance (y e −y)in he ollowing manne :
R(η, y)=R(η, y e )+ s(y e −y)
R0
η(5)
whe e cons an e ms ha e been omi ed, as hey do no con-
ibu e o he azimu h phase cohe ence. The e o e, he ange
mig a ion di e ence be ween e y close sca e e poin s can
be sa ely exp essed as a linea unc ion, because he nonlinea
e ms a e negligible compa ed o he ada pulse wa eleng h.
Fo an azimu h dis ance be ween sca e e s o 100 m, using
nominal alues o Sen inel-6 (see Table I), he linea e m o
he ange mig a ion di e ence is in he o de o 0.5 m, while
nonlinea e ms a e in he o de o mic ome e and less, a
below he o de o he pulse wa eleng h, app oxima ely 22 mm
o he Sen inel-6 Ku-band.
HERNÁNDEZ-BURGOS e al.: ACCELERATED ALGORITHM FOR BACK-PROJECTION FF SAR ALTIMETRY 5209816
E. Range Misalignmen
The mos c i ical aspec o he simpli ica ions made ea lie
is he esidual misalignmen o he ac oss- ack wa e o ms in
ange. Recalling he phase o he L1A wa e o ms om (2)
θ(η, ,y)
=2π c
2
cR(η, y)−2
c(R(η, y)−R k)− d(η, y)
α .
(6)
The accele a ed back-p ojec ion me hod co ec s he RRP,
which co esponds o he i s e m ( he one mul iplied by
he ca ie equency c), by applying a seconda y phase
co ec ion. Howe e , he ange phase delay co ec ion, com-
monly known as ange cell mig a ion co ec ion (RCMC),
is no applied o seconda y su aces. The a ionale behind
his omission is o a oid an addi ional complex co ec ion
and ange comp ession ( ia FFT) o each seconda y su ace
poin . Since back-p ojec ion in eg a es all pulses wi hin he
illumina ion ime o achie e p ope ocusing, i is c ucial ha
all pulses align accu a ely a he same sca e e delay sample.
None heless, his esidual e o is una oidable, and he only
iable mi iga ion s a egy is o ensu e ha he dis ance be ween
he e e ence and he seconda y poin emains su icien ly
small o keep he esidual e o wi hin negligible limi s.
Wi hin his con ex , i a ange comp ession ( ange FFT) is
applied o a seconda y su ace poin unde he assump ion
o uni o m ecei ed ene gy, he esul ing signal would be
a delayed sinc unc ion p opo ional o he ange-mig a ion
di e ence be ween he e e ence and he seconda y poin ,
as well as he di e ence be ween hei espec i e Dopple -
delay shi s
S (η, ,y)=sinc[B( −1τ ·η)]wη(η, y)
·expj2π c
2
cR(η, y)(7)
whe e 1τ ep esen s he ange-mig a ion di e ence, as de ined
in (5), plus he di e ence in he Dopple delay shi d.
We de e mine he c i e ia such ha misaligned wa e o ms
a e s ill cohe en ly in eg a ed i he wa e o ms a he ex emes
o he syn he ic ape u e a e wi hin he ac oss- ack esolu ion
δ o he wa e o m ela i e o he middle o he syn he ic
ape u e. Wi h hese c i e ia, we could de e mine he maximum
dis ance a ailable om a e e ence poin as ollows:
y e −y≤δ ·R0
sTi
(8)
whe e o he sake o simplici y we ha e omi ed he Dopple
delay shi d, as o e y close a ge s (less han 100 m), he
impac o he la e on he maximum dis ance equi emen
is in he o de o millime e s, hus conside ed negligible
in his case. As e iden , he maximum a ailable azimu h
dis ance be ween bo h poin s dec eases as he in eg a ion ime
inc eases, gi en he inc eased ange mig a ion di e ence a
he syn he ic ape u e ex emes. In he case o Sen inel-6, he
ac oss- ack esolu ion is app oxima ely 0.42 m [16]. Taking
he nominal al i ude as he ange o closes app oach, 1336 km,
he nominal speed o he Sen inel-6 sa elli e, 7200 m/s and
using 3.4 s o illumina ion ime ( he maximum a ailable o
Fig. 5. Sen inel-6 L1A no malized powe (dB) om a poin a ge a e
ange comp ession. (a) Powe a he poin a ge loca ion. (b) Powe a 40 m
along- ack dis ance om he poin a ge loca ion. The ange misalignmen
is due o he ange mig a ion di e ence be ween bo h along- ack poin s.
he Dopple bandwid h wi hou aliasing [16]), he maximum
dis ance be ween a e e ence poin and a seconda y poin o
keep he wa e o ms aligned is a ound 23 m.
Fig. 5p esen s a 2-D image o he wa e o ms om a
Sen inel-6 pass o e a ansponde loca ed in C e e [18]
a e aligning all wa e o ms. The le image (a) shows he
aligned wa e o ms a he exac ansponde loca ion, while
he igh image (b) illus a es he miss-aligned wa e o ms a a
40-m along- ack dis ance om he ansponde . As obse ed,
wa e o ms a e sligh ly misaligned due o he ange mig a ion
di e ence be ween he wo along- ack posi ions. The una oid-
able miss-alignmen leads o an ine i able loss o esolu ion
in bo h dimensions, i.e., ac oss- ack and along- ack.
Assuming a symme ical ecei ed powe along he illumi-
na ion ime a ound he POCA, his misco ec ion should no
a ec ange measu emen accu acy, as he posi i e and nega-
i e unco ec ed delays cancel ou du ing azimu h in eg a ion,
e ec i ely cen e ing he ocused pulse a he co ec poin .
Howe e , he ecei ed powe om a sca e e poin du ing he
illumina ion ime may no be symme ic, consequen ly, one o
he sides o he syn he ic ape u e ecei es mo e powe han
he o he ex eme. As a esul , a ange bias inc easing wi h
in eg a ion ime can be in oduced when all wa e o ms a e
in eg a ed. This e ec is app eciable o e y b igh a ge s,
such as ansponde s, and ex eme back-p ojec ion con igu-
a ions. Fo example, o 20 m o dis ance o he e e ence
poin and 3.4 s o in eg a ion ime, he ange bias is in he
o de o 3 mm. In he case o na u al scena ios, his e ec
will depend on he geome y con ex bu is expec ed o be
much lowe (below he millime e ). Thus, he e is a ade-o
be ween he spacing o e e ence poin s and he accu acy o
he measu emen s. A de ailed s udy o his e ec is p o ided
in he alida ion sec ion.
IV. METHODOLOGY
As men ioned ea lie , he e e ence su ace poin s a e
ocused using he classic back-p ojec ion me hod desc ibed
in [9]. The seconda y su ace poin s a e ocused by eusing
he se o co ec ed wa e o ms om a nea by e e ence
poin be o e in eg a ion, wi h a seconda y phase co ec ion
applied, p opo ional o he dis ance be ween he e e ence and
seconda y poin s. In his sec ion, we p o ide a summa y o he
classic back-p ojec ion me hod in he case o Sen inel-6, along
5209816 IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, VOL. 63, 2025
wi h a de ailed explana ion o he ma hema ical de elopmen
behind he seconda y ocusing.
A. Re e ence Su ace Poin s: Classic Back-P ojec ion
The back-p ojec ion algo i hm is explained in [9] in he
case o C yoSa -2, which employs a de- amp il e ins ead o
ma ched il e ing as in he Sen inel-6 case. The only di e ence
be ween hem is he absence o he esidual ideo phase (RVP)
e m in he ma ched il e case. The es o he app oach is
applicable. To gi e some con ex , a summa y o he classic
back-p ojec ion is explained. The FF-SAR back-p ojec ion
algo i hm o a ma ched- il e case sys em consis s he e o e
o h ee main well-known s eps.
1) Range Cell Mig a ion Co ec ion: The RCMC in ol es
co ec ing he sca e e delay di e ence be ween a poin
loca ed a he along- ack posi ion yp and he posi ion o
he sa elli e o each pulse. The ange-mig a ion is de ined
as δR(η, yp )=R(η, yp )−R0, whe e R0is he ange a he
POCA. We can exp ess he RCMC as ollows:
SRCMCη, τ, yp =S η, ,yp
·exp" 2
cδRη, yp − dη, yp
α! #.
(9)
A e his mul iplica ion, a ange-comp ession is applied
using a Fou ie T ans o m. The 2-D signal in he empo al
domain is
S cη, τ, yp ≈TpsincB(τ−γ0)
·expj2π c
2
cRη, yp (10)
whe e γ0=(R0−R k)·c/2. The emaining phase e m is he
RRP, which is he dominan phase e m.
2) Along-T ack Focusing: The along- ack ocusing
in ol es co ec ing he RRP and in eg a ing all he cohe en
wa e o ms. This is done o all he poin s de ined on
he su ace. Fo a poin a ge loca ed a yp , he azimu h
co ec ion is de ined as ollows:
SATη, τ, yp |y=S cη, τ, yp ·exp−j2π c
2
cR(η, y)
=TpsincB(τ−γ0)
·expj2π c
2
cRη, yp −R(η, y)(11)
whe e y e e s o a speci ic along- ack poin on he su ace.
I he e alua ed poin ycoincides wi h he posi ion o he
poin a ge yp , he RRP is ully co ec ed, and he ene gy
a e in eg a ing all he pulses will be maximum. The 2-D
PTR can be compu ed as he in eg a ion o all wa e o ms in
he azimu h di ec ion a e he RRP co ec ion, in unc ion o
he ange delay τand he along- ack poin y
χp (τ, y)=X
η∈Ti
SATη, τ, yp |y
≈TpTisincB(τ−γi)sinc2Ti c s
cR0y−yp .
(12)
B. Seconda y Focusing
S a ing om (11), we can exp ess he azimu h-co ec ed
signal e alua ed a a e e ence su ace poin y e as ollows:
SATτ, η, yp |y e =TpsincB(τ−γ0)
·expj2π c
2
cRη, yp −R(η, y e ).
(13)
Subs i u ing he ange mig a ion di e ence de ined in (5)
in o he la e equa ion, we can exp ess he azimu h-co ec ed
signal as a unc ion o he wa e o ms e alua ed a a e e ence
su ace poin mul iplied by an exponen ial linea phase e m
SATτ, η, yp |y e =TpsincB(τ−γ0)
·exp"j2π c
2 sy e −yp
cR0
η#.(14)
Thus, o any seconda y along- ack poin y, we can exp ess
he azimu h ocusing as he wa e o ms co ec ed a he e e -
ence su ace poin (la e equa ion) mul iplied by a seconda y
phase co ec ion, p opo ional o he dis ance be ween yand
he e e ence poin y e
SATτ, η, yp |y e →y=SATτ, η, yp |y e
·expj2π c
2 s(y−y e )
cR0
η
=TpsincB(τ−γ0)
·exp"j2π c
2 sy−yp
cR0
η#.(15)
In summa y, we apply he RCMC and RRP co ec ions
ela i e o a e e ence poin using eal sa elli e da a o ob ain
SAT(τ, η, yp |y e ). Then, we apply he seconda y phase co ec-
ion o a seconda y along- ack poin y, close o he e e ence
poin y e , ob aining an exp ession equi alen o (11)
SATτ, η, yp |y e →y≈SATη, τ, yp |y.(16)
A e ha , we can ob ain he 2-D PTR by in eg a ing all
wa e o ms in he azimu h di ec ion, yielding he same esul
as in (12).
The p incipal ad an age is ha , by u ilizing a single
e e ence-co ec ed ma ix (14),SAT(τ, η, yp |y e ), we can
ocus mul iple seconda y poin s wi h only an addi ional
seconda y co ec ion (15),SAT(τ, η, yp |y e →y). The com-
pu a ional cos o pe o ming he classic back-p ojec ion o
e e y along- ack poin y(11) is much la ge han pe o ming
a single e e ence co ec ion (14) and hen apply as many
seconda y co ec ions (15) as needed.
I is impo an o no e ha he linea app oxima ion is only
used o he seconda y phase co ec ion (15), which is alid o
azimu h dis ances below ∼23 m, as p e iously demons a ed.
C. Subape u e A e aging
An addi ional s ep o enhance he compu a ional e iciency
o he FF-SAR back-p ojec ion algo i hm in ol es a e aging
s acks o Nsub wa e o ms, p e iously aligned wi h espec o
a e e ence sca e e and s o ing his educed ma ix o he
seconda y phase co ec ion and azimu h in eg a ion. In his
HERNÁNDEZ-BURGOS e al.: ACCELERATED ALGORITHM FOR BACK-PROJECTION FF SAR ALTIMETRY 5209816
Fig. 6. Simula ed poin a ge powe loss (dB) using Sen inel-6 nominal
pa ame e s o di e en dis ances o e e ence poin s, in unc ion o he s ack
size.
app oxima ion, we assume ha he RRP is cons an du ing a
ce ain numbe o consecu i e pulses, allowing he in eg a ion
o all he pulses o emain cohe en wi hou needing o
apply he RRP co ec ion. The use o subape u es o educe
compu a ional e o has al eady been s udied o SAR
imaging, speci ically in he as ac o ized back-p ojec ion
algo i hm [17]. The ocusing p ocess is he e o e di ided in o
wo in eg a ions: one be o e he ex a azimu h co ec ion o
a small ime o subape u es Tsub and one a e he seconda y
phase co ec ion o he en i e in eg a ion ime bu wi h ewe
poin s.
The leng h o each s ack should be small enough so he
RRP wi hin he subape u e emains su icien ly cons an o
ha e a cohe en summa ion. S a ing om (14), his p ocess
consis s o in eg a ing he pulses wi hin a shi ing window. The
in eg a ion is ollowed by a educ ion s ep, whe e he esul ing
ec o is scaled p opo ionally o he window size
SAT,Tsub τ, η, yp |y e
=Tp
Tsub
sincB(τ−γ0)
·
Nsub−1
X
k=0(exp"j2π c
2 sy e −yp
cR0·η+k·Tp i#)
(17)
whe e Tp i is he pulse epe i ion in e al in seconds. The esul
o his in eg a ion is a 2-D ma ix wi h a educ ion ac o
p opo ional o he s ack leng h. The slow- ime a iable ηnow
has a pulse epe i ion in e al p opo ional o he subape u e
ime Tsub. A e he s ack a e aging and educ ion, he sec-
onda y poin s a e ocused as usual, applying a seconda y phase
co ec ion and in eg a ing all he pulses along ack
χp (τ, y)=X
η∈TillSAT,Tsub τ, η, yp |y e
·expj2π c
2 s(y−y e )
cR0
η.(18)
I he phase e m a e he s ack a e aging we e ze o, he
in eg a ion would be ully cohe en , and we would ob ain he
same PTR as in (12). To assess he e ec o he subape u e
app oxima ion, we mus analyze he cohe ence loss o he
signal due o he emaining RRP phase ha was no co ec ed
when applying he s ack a e aging. Thus, e alua ing he sum-
ma y o (17) as a con inuous in eg al
SAT,Tsub (τ, η, y|y e )
≈Tp
Tsub
sincB(τ−γ0)
·ZTsub/2
−Tsub/2
exp"j2π c
2 sy e −yp
cR0η+η′#dη′.(19)
We can exp ess he e e ence-co ec ed ma ix a e sec-
onda y ocusing as ollows:
SAT,Tsub τ, η, yp |y e →y=sincB(τ−γ0)
·sinc" c
2 sy e −yp
cR0
Tsub#
·expj2π c
2 s(y−y e )
cR0
η.
(20)
The second sinc unc ion is esponsible o he subape -
u e app oxima ion. As obse ed, he powe decays apidly
wi h espec o he subape u e cohe ence ime Tsub and he
along- ack dis ance be ween he poin a ge and he e e ence
poin . Mo eo e , i is essen ial o no e ha his e m depends
on he dis ance be ween he e e ence poin and he poin
a ge loca ion, bu no on he seconda y su ace poin s (y).
This implies ha he ocused PTR will emain unchanged,
as he azimu h exponen ial phase is s ill main ained o pe o m
he ocused esponse. Thus, he along- ack esolu ion emains
dependen on he o al illumina ion ime Ti, i espec i e o he
s ack leng h. In ac , i Tsub is se o 64 pulses and we pe o m
a powe (incohe en ) in eg a ion along he ape u e, we would
ob ain he well-known un ocused SAR al ime y esponse [6],
[9],[15]. Fo he FF PTR, s ack a e aging di ec ly impac s
he signal- o-noise a io (SNR), leading o a noisie PTR.
Fig. 6illus a es he powe decay a he poin a ge as
we inc ease he cohe ence in eg a ion ime and he dis ance
o he e e ence poin . Fo a 20 m dis ance o he e e ence
poin , we obse e a powe loss o 1 dB o a cohe ence
ime o 27.5 ms, which co esponds o a s ack leng h o
Nsub =247 pulses. Howe e , he imp o emen in compu a-
ional e iciency sa u a es close o 30–40 pulses, esul ing in
a powe loss o less han 20 mdB.
This powe decay is in luenced by sa elli e and ada al ime-
e pa ame e s such as pla o m eloci y and PRF, meaning
hese alues will a y o o he ada al ime e missions, such
as Sen inel-3 and C yoSa -2.
D. Algo i hm Implemen a ion
This sec ion ou lines he s eps in ol ed in implemen ing
he accele a ed back-p ojec ion algo i hm. Fig. 7shows he
lowcha o he back-p ojec ion algo i hm, and Fig. 8illus-
a es he algo i hm unc ions scheme o bo h classic and
5209816 IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, VOL. 63, 2025
Fig. 7. Algo i hm lowcha o he accele a ed back-p ojec ion. Fo each g id poin , we i s p ocess a e e ence su ace (g een pa h) and compu e i s
wa e o m. The classic back-p ojec ion is applied and he wa e o ms be o e ull in eg a ion a e s o ed in a bu e . The size o he wa e o ms depends on
he numbe o pulses o he s ack a e aging size Nsub and he ze o-padded ange samples N zp . All seconda y poin s associa ed wi h a single e e ence a e
ocused a e wa d (pu ple pa hs). This is epea ed o all poin s de ined in he g id. The ou pu is a ange wa e o m o each poin on he su ace.
Fig. 8. (Top) FF-SAR accele a ed back-p ojec ion algo i hm scheme.
(Bo om) Each seconda y poin (in pu ple) is associa ed wi h he closes
e e ence poin (in ed) in he su ace poin s.
accele a ed back-p ojec ion ocusing p ocedu es. Fi s , all su -
ace poin s a e de ined by p ojec ing he sa elli e coo dina es
in o he g ound. Then, each su ace poin is classi ied ei he
as a e e ence poin o a seconda y poin . Then, we i e -
a e o e each de ined poin , whe e he e e ence poin s a e
ocused using he classic back-p ojec ion me hod, as desc ibed
in [9] and he seconda y poin s a e ocused as desc ibed
in his a icle. The e o e, gi en a e e ence su ace poin ,
all back-p ojec ion co ec ions a e applied, and he se o
phase-co ec ed wa e o ms p io o azimu h in eg a ion is
sa ed. The e e ence poin is ocused on applying he azimu h
in eg a ion as usual. The sa ed ma ix is hen a e aged and
decima ed by a ac o o Nsub (s ack a e aging). The educed
ma ix is used o ocus all seconda y poin s nea by o he
e e ence su ace poin , applying a linea azimu h phase co -
ec ion. Subsequen ly, he collec ion o co ec ed wa e o ms
is cohe en ly in eg a ed as usual, yielding a seconda y- ocused
look. The dis ance be ween e e ence along- ack poin s is
con igu able, depending on use p e e ences, along wi h o he
pa ame e s such as along- ack spacing, azimu h in eg a ion
ime, and subape u e in eg a ion leng h.
V. ALGORITHM CHARACTERIZATION
To cha ac e ize he limi s and pe o mance o he accele -
a ed back-p ojec ion, we ha e conduc ed a es wi h Sen inel-6
da a o e a ansponde loca ed in Ga dos [20] wi h di -
e en p ocessing con igu a ions. Ac i e elemen s such as
ansponde s o e a high signal- o-clu e a io (SCR) and
allow us o cha ac e ize he eal PTR and compa e di e en
algo i hms. Addi ionally, conduc ing epea ed es s on he
same ansponde o e ime enables us o de e mine long-
e m s a is ics, such as he ange bias, in compa ison be ween
di e en algo i hms and measu emen me hods. In his sec ion,
esul s om he accele a ed back-p ojec ion a e compa ed wi h
he classic back-p ojec ion algo i hm. The pe o mance o
he classic back-p ojec ion algo i hm o e ansponde s and
co ne e lec o s has been al eady shown in [23] ( ansponde )
and [19] (co ne - e lec o ).
HERNÁNDEZ-BURGOS e al.: ACCELERATED ALGORITHM FOR BACK-PROJECTION FF SAR ALTIMETRY 5209816
Fig. 9. 2-D PTR no malized powe in dB om he ansponde loca ed in Ga dos. Accele a ed back-p ojec ion has been p ocessed o e he di e en dis ances
be ween e e ence poin s, no ed a he op o each image. Fo all cases, 3.4 s o syn he ic ape u e has been in eg a ed. The poin a ge is loca ed in he wo s
case scena io: in he middle be ween wo e e ence poin s. The 2-D PTR deg ades as we inc ease he dis ance be ween e e ence poin s, as expec ed.
Fig. 10. Sen inel-6 PTR no malized powe (dB) om he ansponde loca ed
in Ga dos, p ocessed wi h classic and accele a ed back-p ojec ion algo i hm
o di e en dis ances be ween e e ence su aces and 3.4 s o in eg a ion
ime. (Le ) Ac oss- ack cu . (Righ ) Along- ack cu .
Fig. 11. PTR esolu ion deg ada ion om he ansponde (Ga dos),
p ocessed wi h classic and accele a ed back-p ojec ion algo i hm o di e en
dis ances be ween e e ence poin s and 3.4 s o in eg a ion ime. The eso-
lu ion deg ada ion is ela i e o he e e ence alue ob ained wi h he classic
back-p ojec ion.
To pe o m he cha ac e iza ion es s, he ansponde coo -
dina es a e loca ed in he middle be ween wo e e ence poin s,
ha is he a hes seconda y poin wi h espec o a e e ence
poin and, he e o e, he wo s case o he con igu a ion in
e ms o esidual e ec s. All es s ha e been pe o med wi h
3.4 s o in eg a ion ime. Mo eo e , he dis ance be ween
along- ack su ace poin s is 0.05 m and he ange wa e o ms
ha e been o e sampled wi h a ac o o 8, which esul s in
a ange bin spacing a ound 0.05 m. Any an enna pa e n
co ec ion has been applied.
Fig. 12. PTR powe decay in dB wi h espec o he classic back-p ojec ion
om he ansponde loca ed in Ga dos, o di e en dis ances be ween
e e ence poin s and 3.4 s o in eg a ion ime.
Fig. 13. Peak sidelobe (PSL) om he ansponde loca ed in Ga dos,
p ocessed wi h classic and accele a ed back-p ojec ion algo i hm o di e en
dis ances be ween e e ence poin s and 3.4 s o in eg a ion ime. The PSL is
calcula ed ela i e o he maximum powe .
A. PTR: Re e ence Poin Dis ance Analysis
Fig. 9illus a es he deg ada ion o he ocused 2-D PTR
as we inc ease he dis ance be ween e e ence poin s, no ed a
he op o each image. A 40 m, he expec ed deg ada ion
o he 2-D PTR becomes isible. Fig. 10 showcases he
c oss and along- ack cu s a he poin o maximum powe .
As we inc ease he dis ance be ween e e ence poin s, he
sinc esponse ge s wide , abso bing ene gy om he sidelobes.
Fig. 11 shows he ac oss- ack and along- ack esolu ion e o
wi h espec o he classic back-p ojec ion in he unc ion
o he dis ance be ween wo e e ence poin s. As we can
see, he e o inc eases apidly as we inc ease he dis ance
be ween e e ence poin s abo e 40 m. Mo eo e , Fig. 12
depic s he powe loss a he peak o he PTR, while Fig. 13
5209816 IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, VOL. 63, 2025
Fe an Gibe (Membe , IEEE) ecei ed he M.S.
deg ee in ae ospace enginee ing and he Ph.D.
deg ee in ae ospace science and echnology om
he Uni e si a Poli ècnica de Ca alunya, Ba celona,
Ca alonia, in 2011 and 2016, espec i ely, ocused
on he design o he mal diagnos ics expe imen s o
he LISA Pa h inde mission.
F om 2016 o 2018, he was a Pos -Doc o al
Resea che wi h he Uni e si y o T en o, T en o,
I aly, om whe e he suppo ed he LISA Pa h inde
ope a ions and da a analysis phase. Since June 2018,
he has been a Resea ch and De elopmen Enginee wi h isa dSAT, S.L.,
Ba celona, being cu en ly in ol ed in he de elopmen o algo i hms o ada
al ime e p ocesso s.
An oni B oque as (Li e Membe , IEEE) was bo n
in Ba celona, Ca alonia, in 1959. He ecei ed he
Enginee ing and Doc o Enginee ing deg ees in
elecommunica ion enginee ing om he Uni e si a
Poli ècnica de Ca alunya (UPC), Ba celona, Ca -
alonia, in 1985 and 1989, espec i ely, whe e his
disse a ion was on mic owa e omog aphy.
In 1986, he was a Resea ch Assis an wi h
Po smou h Poly echnic, Po smou h, U.K., whe e
he was in ol ed in p opaga ion s udies. In 1987,
he joined he Depa men o Signal Theo y
and Communica ions, School o Telecommunica ion Enginee ing, UPC.
F om 1998 o 2002, he was he Subdi ec o o Resea ch a he Ins i u e
o Geoma ics, Ba celona. Since 1999, he has been a Full P o esso wi h
UPC, whe e he is in ol ed in esea ch on ada imaging and emo e sensing.
F om 2003 o 2006, he was he Di ec o o he Depa men o Signal Theo y
and Communica ions, UPC. F om 2005 o 2009, he p oposed and comanaged
he Knowledge To Ma ke (K2M) P og am a UPC o suppo uni e si y spin-
o s. He has published mo e han 200 a icles on mic owa e omog aphy,
ada , ISAR, and syn he ic ape u e ada (SAR) sys ems, SAR p ocessing,
and in e e ome y. His p esen esea ch in e es s include SAR calib a ion
echniques, GMTI SAR pe o mance analysis, GEOSAR, and mmW g ound-
based SAR, in suppo o u u e SAR missions and ada emo e sensing
inno a ion.
Pol Villal illa ecei ed he M.S. deg ee in
elecommunica ions enginee ing om he Uni e si-
a Poli ècnica de Ca alunya, Ba celona, Ca alonia,
in 2020.
F om 2021 o 2022, he comple ed a one-yea
esea ch s ay a he Ea h Obse a ion and Remo e
Sensing Depa men , Swiss Fede al Ins i u e o
Technology (ETH Zü ich), Zü ich, Swi ze land, pe -
o ming signal p ocessing echniques o e measu e-
men s in alpine a eas o assis Pe sis en Sca e e
In e e ome y. Since 2022, he has joined isa dSAT,
S.L., Ba celona, o wo k on he de elopmen and implemen a ion o p ocess-
ing algo i hms o ada al ime y o u u e ESA missions.
Alejand o Egido ecei ed he mas e ’s deg ee in
elecommunica ions enginee ing om he Uni e -
sidad de Za agoza, Za agoza, Spain, in 2007, and
he Ph.D. deg ee in elecommunica ions enginee -
ing om he Uni e si a Poli ècnica de Ca alunya,
Ba celona, Spain, in 2013.
In 2007, he joined S a lab Ba celona, Ba celona,
as a Scien i ic Resea che . In 2015, he mo ed o he
Labo a o y o Sa elli e Al ime y, Na ional Oceanic
and A mosphe ic Adminis a ion (NOAA), College
Pa k, MD, USA, whe e he wo ked on de eloping
imp o ed SAR al ime y p oduc s and algo i hms o physical oceanog aphy
and sea ice applica ions. In 2016, he was a NOAA’s Measu emen Sys em
Enginee o he Jason al ime e missions. Since 2022, he has been an
Ocean and Hyd ology Ea h Obse a ion Scien is in he Ea h Su ace and
In e io Sec ion o Eu opean Space Agency’s Mission and Science Di ision,
based a he Eu opean Space Resea ch and Technology Cen e, Noo dwijk,
The Ne he lands. His main esea ch in e es s include using ada al ime y
o cha ac e ize open ocean, coas al, and pola zones, and le e aging global
na iga ion sa elli e sys em signals as sou ces o oppo uni y o emo e sensing
applica ions.
D . Egido ecei ed he NOAA’s Da id S. Johnson Awa d o his wo k on
ully ocused SAR al ime y in 2020.
Go ka Moyano ecei ed he M.Sc. deg ee in
elecommunica ions enginee ing and he M.Sc.
deg ee in esea ch on in o ma ion and communica-
ion echnologies om he Uni e si a Poli ècnica de
Ca alunya, Ba celona, Ca alonia, in 2009.
F om 2010 o 2013, he wo ked a Ind a Espacio,
Ba celona, as a Radiona iga ion So wa e Enginee ,
whe e he de eloped GNSS moni o ing sys ems o
ai a ic con ol and au oma ed es ing ools o
he Galileo Senso S a ion Co e Compu e . In 2013,
he joined isa dSAT, S.L., Ba celona. He is cu en ly
a So wa e A chi ec Enginee liaising scien i ic algo i hms and so wa e
de elopmen . His skills a e key in p ojec s, such as Sen inel-6 P4 L1 G ound
P ocesso P o o ype (GPP) de elopmen , he Al ime e L1 P oduc Gene a ion
Func ion (PGF) p ocesso s wi hin he Sen inel-6 PDAP, he CRISTAL IRIS
L1&GR GPP, and he EPS-SG MWS/MWI/ICI L1b Local P ocesso s e-
enginee ing.
Ad ián Flo es de la C uz ecei ed he M.S. and
Ph.D. deg ees in elecommunica ions enginee ing
om he Uni e sidad Poli écnica de Ca agena, Mu -
cia, Spain, in 2014 and 2019, espec i ely.
F om 2017 o 2019, he was wi h Hisdesa SA,
Mad id, Spain, as a Cal/Val Enginee , whe e he
wo ked on he PAZ Mission. Since Ma ch 2019,
he has been a Resea ch and De elopmen Enginee
wi h isa dSAT, S.L., Ba celona, Ca alonia, wo k-
ing on he de elopmen and implemen a ion o
da a p ocessing echniques wi h a special ocus on
ins umen al calib a ion.
Ma co Fo na i ecei ed he M.Sc. deg ee in
elecommunica ions enginee ing om he Uni e si y
o Rome “La Sapienza,” Rome, I aly, in 2006.
He hen joined he Ea h Obse a ion Di ec-
o a e, ESA-ESTEC, Noo dwijk, The Ne he lands,
i s as a Young G adua e T ainee and hen as a
RHEA/S a ion con ac o un il p esen . Since 2007,
he has been in ol ed in se e al ESA al ime y mis-
sions. He was ini ially pa o he C yoSa -2 P ojec
as Ins umen Calib a ion Enginee , and s ill suppo s
C yoSa -2 Ope a ions ia he ESA-ESRIN Team.
In 2011, he joined he Sen inel-6 P ojec as a G ound Segmen Enginee ,
leading he de elopmen o he al ime e g ound p o o ype p ocesso s and
s ill wo king on hei u u e e olu ions, such as he ully ocused p ocessing.
Mo e ecen ly, he has been suppo ing CRISTAL and he S3NG-Topo P ojec s
as an al ime e ins umen and da a p ocessing expe .
