This pape is a ailable online a www.p aisewo hyp ize.o g
In e na ional Re iew o Ae ospace Enginee ing (I.RE.AS.E), Vol. 12, N. 1
ISSN 1973-7459 Feb ua y 2019
Copy igh © 2019 The Au ho s. Published by P aise Wo hy P ize S. .l.
This a icle is open access published unde he CC BY-NC-ND license (h p://c ea i ecommons.o g/licenses/by-nc-nd/3.0/)
A ailable online by Feb ua y 28 h, 2019 h ps://doi.o g/10.15866/i ease. 12i1.16153
12
N oM: a Concep o Ope a ions o Pilo s
o Mul iple Remo ely Pilo ed Ai c a
Miguel-Ángel Fas-Millán, En ic Pas o
Abs ac – The concep o ope a ions p oposed he e pu sues he easibili y, om a human
ac o s pe spec i e, o ha ing a single pilo /ai c ew con olling se e al emo ely pilo ed ai c a
sys ems a once in non-seg ega ed ai space. To mee such easibili y, his mul i asking mus be
sa e and no in e e e wi h he job o he ai a ic con olle s due o delays o e o s associa ed
wi h pa allel pilo ing. To ha end, a se o measu es a se e al le els is sugges ed, which includes
wo kload p edic ion and balance, pilo ac i i y moni o ing, and a special emphasis on in e ace
usabili y and he pilo ’s si ua ional awa eness. The concep elies g ea ly on he exploi a ion o
he po en ial o Con olle -Pilo Da a Link Communica ions, an icipa ing u u e widesp ead
implemen a ion and ull use. Expe imen s compa ing he pe o mance o he same pseudo-pilo s
be o e and a e he implemen a ion o pa o he measu es showed a dec ease in he numbe o
e o s, o e sigh s and subjec i e s ess. Copy igh © 2019 The Au ho s.
Published by P aise Wo hy P ize S. .l.. This a icle is open access published unde he CC BY-NC-ND license
(h p://c ea i ecommons.o g/licenses/by-nc-nd/3.0/).
Keywo ds: RPAS, UAS, CPDLC, Mul i-RPAS, Mul i-UAS
I. In oduc ion
One o he p inciples behind he guidelines p o ided
by he In e na ional Ci il A ia ion O ganiza ion (ICAO)
is o gua an ee equali y in access o ai space. Tha was
why in 2011 he agency se ou he basis o he
de elopmen and in eg a ion o Remo ely Pilo ed
Ai c a Sys ems (RPAS). As a esponse o ha
ecommenda ion, in 2012, he Eu opean Commission
c ea ed he Eu opean RPAS S ee ing G oup o
s akeholde s, who de eloped and published he Eu opean
RPAS Roadmap [1], a wo k plan de ining he s eps,
esponsibili ies and miles ones o he g adual inse ion
o RPAS in he Single Eu opean Sky (SES) [2]. Saying
ha his in eg a ion implies a comple ely new indus y
wi h a wide se ice sec o and supply chain is no a
p edic ion anymo e, bu a eali y ha is g owing e e y
day. Howe e , e en hough he po en ial applica ions and
p omising ma ke o he unmanned ae ial sys ems (UAS)
a e well known, any acili y o hei comme cial
exploi a ion will mean an incen i e o hose in es o s
needed o boos his eme ging indus y. Tha is one o he
goals o his concep o ope a ions (ConOps): o
maximise he p oduc i i y o a lee o RPAS and hei
emo e pilo s, allowing a pilo o ai c ew o con ol
se e al ai c a wi hou losing sigh o sa e y o he
impac ha his mul i asking could cause on he wo k o
he ai a ic con olle s (ATC). This impac could come
om delays in he esponse ime, e o s o o e sigh s.
These p oblems, p one o a ise when pilo ing se e al
ai c a in pa allel, could also be ound in a one- o-one
pilo -RPAS ela ionship. So, he whole concep and i s
sa e y measu es could also be le e aged o he one- o-
one case. The supe ision o con ol o mul iple
unmanned ae ial ehicles (UAVs) is a g owing opic o
esea ch s ill in i s ea ly days, bu se e al app oaches can
al eady be ound. Some o hem a e mo e app op ia e o
speci ic kinds o missions, o he s equi e almos ull
au onomy o he ai c a , and mos o hem we e
concei ed o seg ega ed ai space. Di e en pa adigms
can be ound; he ones whe e he pilo is expec ed o pilo
he ai c a o he ones whe e he ope a o is basically in
cha ge o de ining and guiding he mission [3]. In o de
o imp o e p oduc i i y and e icacy, much esea ch,
mainly add essed o mili a y ope a ions, ocuses on
imp o ing he design o he G ound Con ol S a ion
(GCS). Some o hese imp o emen s a e based on a
ca e ul balance o cogni i e esou ces and men al
wo kload. O he s y o enhance he human scope by
eaming pilo and adap i e au oma ion [4]. The sea ch o
p oduc i i y has also some imes been based on ask
scheduling [5], [6]. N oM combines some o bo h
wo lds: i ies o a oid unmanageable le els o wo kload
by scheduling and ask suppo , bu also p o ides
measu es o back up he pilo i i appea s. The pape
begins wi h a lis o assump ions ega ding he e ms
used in he documen and he condi ions o scena ios
conside ed o a possible implemen a ion o he ConOps.
I also speci ies which kind o ope a ions would be e
le e age he N oM concep . Nex , he e is a desc ip ion
o he a chi ec u e o he sys em, i s inpu s, pa s and
hei ela ionships. La e in ha sec ion, i s di e en
ea u es a e explained g ouped by ype: usabili y,
wo kload o moni o ing. Then, he e is a desc ip ion o
M.-A. Fas-Millán, E. Pas o
Copy igh © 2019 The Au ho s. Published by P aise Wo hy P ize S. .l. In e na ional Re iew o Ae ospace Enginee ing, Vol. 12, N. 1
13
he expe imen s conduc ed o check he pe o mance o
he pseudo-pilo s wi h o wi hou some o he ea u es
implemen ed and he esul s ob ained. Finally, he
conclusions abou he po en ial bene i s o he ConOps
a e gi en.
II. Assump ions
Th oughou he es o he pape , o simplici y, he
Ai Vehicle Ope a o (AVO) will be e e ed o as he
pilo , bu assuming ha an ai c ew could be behind he
e m. In p inciple, o he ConOps, he e is no limi in he
ole a chi ec u e o he ai c ew, as long as we could
de ine he wo kload, in e ac ions, and incompa ibili ies
among hei asks. Fo ins ance, we could ha e a payload
ope a o o se e al ai c a , while hese a e pilo ed by
di e en AVOs. In a case like ha , he sys em should
handle he a ailabili y o he payload ope a o o he
di e en ligh s besides he a ailabili y o hose ligh s’
AVOs i hese a e con olling mo e han one ai c a . The
sys em could also allow a c ew assignmen on demand
like he one sugges ed in NASA’s SPO ConOps [7] o
he hyb id g ound dispa che - i s o ice ole in manned
single-pilo ope a ions. Fo cla i y, his documen will
ocus on how he concep manages he assignmen o
ligh s o he AVOs ope a ing as a single pilo , bu a
g ea e deg ee o lexibili y could be allowed in he
quan i y and ype o oles in ol ed. Cu en unsol ed
issues ega ding he accep ance o he in eg a ion o
RPAS by he gene al public, o he legal equi emen s o
he S a es, like policies o he de ini ion o
esponsibili ies, a e le ou . In he wo s case, he
concep could s ill be use ul in seg ega ed ai space. This
wo k aims o be a b ick in he building o he RPAS
in eg a ion bu conside ing ha he ounda ions o his
in eg a ion a e jus a ma e o ime. Ano he ques ion o
ime is he widesp ead use o Con olle -Pilo Da a Link
(CPDLC) [8, Pa IV]. Being N oM a concep ha would
equi e a long- e m implemen a ion, i is assumed ha
CPDLC will al eady be he main ai -g ound means o
communica ion. Then, some o he measu es exploi i s
cu en ly unde used po en ial. I should be no ed ha he
use o he e m RPAS means ha w a human pilo in
command is conside ed. Anyway, he a ge ai c a will
be so highly au oma ed o allow he pilo o delega e a
la ge pa o he asks, which is key o be able o con ol
se e al ai c a concu en ly. An example o his appea s
in [9], whe e a mili a y pilo is expec ed o con ol
se e al unmanned ai c a while pilo ing a manned
ai c a . The GCS should allow he con ol o di e en
ai c a models wi h he e ogeneous le els o au onomy
and equipmen . [10] add esses he equi emen s and
design o his kind o con ol s a ion, a sample o which
was pa o he ambi ious DARIUS p ojec , wi h i s
Gene ic G ound S a ion able o con ol ma i ime, ae ial
o g ound unmanned sys ems used in sea ch and escue
ope a ions [11]. In ac , he N oM concep could be
applied o any ype o unmanned ehicle o o p ocess
supe ision, as he basis o he sys em managemen elies
on wo kload, ask in e ac ion cons ain s and p ocedu e
de ini ions, which a e agnos ic o he kind o ac i i y
pe o med. The kind o mission ha would bes p o i
om his ConOps would be one wi h a p ede e mined
ajec o y. Su eillance missions in which he pa h is
decided on he go would no i oo well in he concep .
While he sys em can handle he consequences o a
change o ajec o y, like a di ec o clea ance, he
ecalcula ion o he wo kload o ecas could o ce a
escheduling and nega i ely a ec o he pilo s, especially
i done o en. The echnological and compu ing
in as uc u e o ches a ing he ConOps will be e e ed
o, in sho , as he sys em, o he se e , wi hou
assuming any explici implemen a ion behind hese
e ms, since i s se ices could be dis ibu ed. The
desc ip ions o he ea u es o he concep appea in h ee
g oups, depending on he kind o measu e in ended o
achie e he easibili y o he pa adigm: wo kload,
moni o ing and usabili y. Howe e , hese aspec s a e
dependen upon one ano he ; usabili y a ec s wo kload
as much as he suppo o he moni o ing bene i s he
usabili y and he wo kload. Then, such ca ego isa ion is
no exclusi e and a ea u e can ha e some dependence;
o be e o wo se, a change in one o hem could a ec
he e icacy o ano he .
III. Sys em Desc ip ion
Fig. 1 shows he a chi ec u e o he N oM se e
p o o ype. I akes as inpu ligh de ini ions, which,
besides he usual da a in a ligh plan, can con ain
p e e ences and cons ain s, u he explained in he
sec ion abou wo kload managemen . This inpu , besides
he wo kload o ecas p o ided by he Wo kload Module
and he cu en s a us o he ligh e ie ed om he
GCS Module, is used by he Schedule o sugges o
apply he ligh -pilo assignmen s.
Fig. 1. N oM se e p o o ype
The E en Handle is in cha ge o a ending eques s
o manage unexpec ed e en s. Ac ing somewha like he
so-called decision au ho i y in adap i e au oma ion, i
will igge no i ica ions, eques s o eedback o
au oma ed ac ions. The Wo kload Moni o o he GCS
M.-A. Fas-Millán, E. Pas o
Copy igh © 2019 The Au ho s. Published by P aise Wo hy P ize S. .l. In e na ional Re iew o Ae ospace Enginee ing, Vol. 12, N. 1
14
Moni o sends hose eques s o managemen . Finally,
he GCS Manage holds he ligh -pilo ela ionships and
he s a us o he ai c a , and i sha es he module wi h
he se ices a ending o he communica ions and
commands.
III.1. Usabili y
Al hough au oma ion is essen ial o allow a pilo o
con ol se e al ai c a , he usabili y o he GCS is key o
allowing a quick esponse and app op ia e awa eness. A
clea example is he wo k ha led o he Noodle Mode
[12] o d aw he in en o each ai c a . While o ien ed o
a di e en scena io, su eillance missions on seg ega ed
ai space, he lesson lea ned is ha he ype o
in o ma ion, he way o displaying i , and he kind o
con ols p o ided a e basic pilla s o sa e y and
p oduc i i y. N oM in ends o be GCS-agnos ic; we
could ha e any GCS plugged in o he sys em ia he
N oM pilo in e ace (NPI), p o ided ha i o e ed an
applica ion p og amming in e ace (API) o he NPI o
access he ligh ins umen s eadings, send commands o
moni o hose execu ed by he pilo . Then, he design
decisions desc ibed he e a e limi ed o he NPI, which is
displayed on a dedica ed moni o , e en when i also
ollows guidelines add essed o he usabili y o he
emo e pilo s a ions [13]. While a inal implemen a ion
o he NPI would equi e a p ocess o e inemen and
su ey pilo s o de e mine which in o ma ion and
no i ica ions hey p e e , we can expec he usabili y o
he GCS and he ai c a au oma ion o be he main
de e minan s o he le el o wo kload associa ed o each
ask, and he e o e he numbe o ligh s pe pilo . The
NPI (Fig. 2) shows a lis o he assigned ligh s, each one
gi en as a s ip wi h he main eadings and hei
associa ed imes amp. The ela i e posi ion o he ai c a
in he cu en segmen can be seen a a glance, unde he
emaining dis ance and ime o he nex waypoin . The
nex segmen shows he minimum link quali y on i i he
cu en link is kep . A he igh end o he s ip, a bu on
allows he ligh de ini ion, p e ious pilo s’ sequence,
and no es o be ead. Reminde s and wa ning icons will
appea nex o he bu on. The in e ace also con ains he
CPDLC display, and he e is oom le o no i ica ions,
con i ma ion dialogues, and a cha oom o communica e
wi h he supe iso  a ole ha will be desc ibed la e 
o o he pilo s ope a ing in he N oM en i onmen ( he
use o ins an messages o coo dina ion is al eady
common in mili a y ope a ions [14]). Di e en au al
indica o s, depending on he u gency and ype,
some imes accompany he p in ing o eminde s o
wa nings.
When he message needs o be mo e a en ion-
g abbing o desc ip i e, speech syn hesis is used. The lis
o ligh s allows he pilo o choose which one o load in
he GCS. As his loading could be ela i ely slow, he
numbe o imes i is equi ed ies o be minimised by
he summa y o eadings, he au oma ion o asks, and
making i possible o send commands and
communica ions o ligh s in he backg ound (any ligh
no cu en ly loaded in he GCS o selec ed in he NPI).
The GCS, which should be connec ed o he NPI, was
simula ed wi h an embedded se o commands, which
was enough conside ing he kind o expe imen s and he
ac ha he pa icipan s we e no p o essional pilo s. A
his poin , an impo an ques ion ha migh a ise is who
is keeping he connec ion o he es o he RPAs ali e
while jus one o hem is loaded in he GCS. To N oM
his would be jus a ma e o implemen a ion, and
logis ic o pe o mance a ailabili y. Fig. 3 shows wo
possible con igu a ions. One (Fig. 3, op), shows a GCS
able o suppo mul iple Command, Con ol and
Communica ions (C3) links, as ei he an embedded
capabili y in he GCS, o p o ided by an ex e nal swi ch.
An example o his kind o GCS is Gene al A omics’
He esy mul i-mission con ol, a so wa e ha a single
pilo can use o con ol up o six la ge UASs like he
MQ-9 om a common lap op including he possibili y o
ans e he con ol o o he pilo s in he ne wo k [15].
Depending on he op ion, he access o he NPI o he
C3 would be ia he GCS API, o di ec ly connec ing o
he swi ch. The second op ion (Fig. 3, bo om) shows a
GCS ac ing as a me e e minal, wi hou any kind o
connec i i y wi h he ai c a , which is ully managed by
he se e . This seems an unlikely con igu a ion o a
eal implemen a ion, as i would add he ne wo k delay o
he one o he C3 link, bu i is he one ep esen ing he
cu en N oM p o o ype. In he p o o ype, he se e
eads he ai c a has simula ed ADS-B eadings om he
socke o a single ins ance o RAISE. RAISE is a
simula ion en i onmen mixing eDEP simula o [16]
a ic and RPAS a ic, de eloped by ICARUS esea ch
g oup o e alua e he impac o he RPAS in eg a ion in
he Ai T a ic Managemen (ATM) [17]. The se e
o wa ds he eadings o he pilo ’s clien applica ion, and
ice e sa o apply he commands o he pilo o he
ai c a . The bene i o his con igu a ion is in he use o
he RTI Connex connec i i y amewo k [18], complian
wi h he Da a Dis ibu ion Se ice (DDS) s anda d, o
communica e se e and clien s.
This allows a simula ion o di e en scena ios o
Quali y o Se ice (QoS), p o iding a amewo k o es
con ingencies ela ed o link quali y o “de e mine wha
compensa o y beha iou s, i any, ai a ic con olle s
and UAV ope a o s adop in esponse o communica ions
delays” [19] o p o ide aining in he associa ed
p ocedu es.
Following he usabili y, and con ibu ing o educing
he o al wo kload, he e a e ask assis ance ea u es like
he possibili y o execu ion and eply o ecei ed
ins uc ions wi h a single click, enhanced CPDLC
message composi ion (including oice composi ion) o
au oma ed epo s wi h o wi hou condi ional e en s.
The CPDLC display was designed aking in o
conside a ion he possibili y o pilo ing se e al ai c a ,
so i ies o educe he head down o he minimum [8,
Pa I, Appendix o Ch. 2, 6- ]. Fu he de ails can be
ound in [20] and sc eencas s in [21].
M.-A. Fas-Millán, E. Pas o
Copy igh © 2019 The Au ho s. Published by P aise Wo hy P ize S. .l. In e na ional Re iew o Ae ospace Enginee ing, Vol. 12, N. 1
15
Fig. 2. The N oM Pilo In e ace. 1) Assigned ac i e ligh s. Each s ip con ains he main eadings, eminde s and access o he ligh in o ma ion.
2) S ack o no i ica ions and dialogues used o eques eedback o p o ide mo e in o ma i e wa nings. 3) Embedded ai c a con ol commands. 4)
CPDLC display. 5) N oM ne wo k Cha oom
A delica e momen du ing he con ol o he RPAS is
also add essed: he hando e p ocess. I has been a
sou ce o e o s wi h d ama ic consequences when he
p ocedu e equi ed manual con igu a ion o he GCS
would no clea ly show he s a us o he p ocess [22].
Ne e heless, in N oM, o ches a ed by he se e , i
a oids human e o s and gua an ees an o de ly
p ocedu e. A pilo ha is selec ed o be assigned wi h a
ligh will see a con i ma ion dialogue o accep o ejec
i . I accep ed, a new s ip o he ecei ed ligh appea s,
amed in a di e en colou o highligh ha i s ill
canno be con olled. Howe e , he eadings a e upda ed,
and he pilo can access he ligh plan and he his o y o
CPDLC o o m a iew o he si ua ion. While no
implemen ed in he p o o ype, he in e ace could allow
pilo s o in oduce anno a ions a ailable o he ollowing
pilo s. A he momen , i shows he lis o p e ious pilo s;
ha way he ecei ing pilo can check who is ans e ing
he ligh and use he cha oom o coo dina e he ans e .
The ecei e pilo , once p epa ed o con ol he ligh ,
in o ms he ans e ing pilo , who applies he ans e .
The ecei e can hen con ol he new ligh and he
p e ious pilo sees he ligh s ip disappea om he
NPI. The sc eencas in [23] illus a es he p ocess. I
conside ed use ul by con olle s, a ee ex message, sen
au oma ically, could ale hem abou he p ocess o
ans e , as i could delay he communica ions o he
execu ion o ins uc ions. We could e en conside
ale ing nea by a ic using an ADS-B b oadcas .
T ans e s could be scheduled, he esul o a
con ingency measu e, o eques ed by he pilo o he
supe iso .
Fig. 3. Connec i i y con igu a ions
[24] desc ibes an in e ace concep wi h common
poin s wi h he NPI. The concep ollows Endsley’s [25]
guidelines o any sma in e ace. A inal NPI
implemen a ion should ollow hem, and he p o o ype
al eady does his o some ex en . Fi s , i should p o ide
a big pic u e o he si ua ion, wi h high-le el in o ma ion,
M.-A. Fas-Millán, E. Pas o
Copy igh © 2019 The Au ho s. Published by P aise Wo hy P ize S. .l. In e na ional Re iew o Ae ospace Enginee ing, Vol. 12, N. 1
16
which could be he de ini ion o he ligh s ips. Second,
he pilo should be able o decide which in o ma ion is
displayed. In his espec , in N oM he ligh loaded in
he GCS would no necessa ily be he one ac i e in he
NPI; you could ha e one ligh loaded and eply o he
ATC o a di e en ligh , o apply he se o commands
associa ed o he ins uc ion ecei ed in a backg ound
ligh . Fligh s ips can be ea anged, bu in he o iginal
mockup, i was planned o place non-exclusi e o de
p e e ences, like close o nex waypoin i s , pending
ac ions i s o u gen ac ions i s . This would be
op ional, as some pilo s could p e e o ind always he
same ligh in he same place. Thi d, Endsley s esses he
impo ance o p o iding in o ma ion abou pas e en s
ha could help pilo s o o m an idea o he cu en
s a us. In N oM, a pilo ecei ing a ligh would ha e
access o he CPDLC his o y, pending e en s, lis o
p e ious pilo s, hei anno a ions, o au oma ed e en
logging (in he NPI all his in o ma ion would appea ,
hiding he CPDLC display by using he bu on wi h he
clip olde icon in he s ip). Las ly, he e is he
impo ance o p o iding a pe sonalised expe ience, o
conside each use ’s p e e ences and limi a ions. The
p e e ences could be sa is ied by p o iding some
lexibili y abou how and which eminde s and wa nings
a e igge ed bu wi h some limi a ions, like c i ical
wa nings ha could no be u ned o . Rega ding i ing
pilo limi a ions, his con enien ly leads us o he nex
sec ion, which de ails how i would be pu sued.
III.2. Wo kload
The pla o m needs o allow he pilo o com o ably
ope a e mo e han one RPAS, which is an ob ious sa e y
measu e and could de e mine i s implemen a ion pe mi ;
e.g. we ind egula ions ha add ess he maximum
wo kload on indi idual c ew membe s o de e mine he
ai wo hiness o he ai plane in [26, FAR Ti le 14, Pa
25.1523 Minimum Fligh C ew]. This objec i e is
possible because many cu en RPAs a e highly
au oma ed and no all ligh phases a e equally
demanding, some imes educed o long and mono onous
supe ision [27], which leads o disengagemen and loss
o awa eness. In his espec , he mul i-pilo ing induces a
sel -mo i a ing a i ude o awa eness (Ye kes-Dodson
law), as could be checked du ing he i s se o
expe imen s when pa icipan s’ pe o mance and
esponse ime o ATC we e be e when hey had ou
ligh s compa ed o ha ing jus one [28]. Bu he sys em
mus gua an ee a iable load o asks pe pilo a all
imes. To ha end, i needs o p edic , e alua e, limi and
dis ibu e he wo kload. [28] desc ibed he equi emen s
o he wo kload desc ip ion o he asks, exempli ied by
he p ocess o measu emen ca ied ou du ing he i s
se o expe imen s. This wo kload, conside ed oughly as
he le el o a en ion and ime equi ed o pe o m he
ask, depends no only on he cu en ask pe o med, bu
also on he concu en ones, and is ep esen ed in a kind
o pilo ’s wo kload p o ile ha would e ol e o e ime
as he pilo gains expe ience, a ac shown in [29] ( he
a o emen ioned pilo limi a ions). The wo kload
measu ed while execu ing a ask, o an o e lap o asks,
is an inpu o he sys em. Such a measu e would be GCS-
speci ic, o p ocedu al-speci ic, al hough, as discussed in
[30], which dissec s he ac o s con ibu ing o he
wo kload o he emo e pilo s, a us wo hy
ep esen a ion o wo kload would ake in o accoun all
he en i onmen al d i e s, no only hose ela ed o he
GCS o he speci ic p ocedu es. N oM accep s any kind
o ask wo kload measu emen . I could be ob ained
using biome ics, pe o mance e alua ion, subjec i e
assessmen s, o a mix o hese echniques. In [31], he
ask wo kload and he o al wo kload o ask
o e lappings is es ima ed wi h analy ically gene a ed
wo kload p o iles (i.e. models p epa ed by an expe
based on he expe ience o he pilo s). The ou pu o ha
wo k is a clea example o one o he main inpu s o
N oM o be able o manage and p edic he mul i-UAS
wo kload. Du ing he i s N oM expe imen s, he
measu emen was done based on pe o mance e alua ion
and subjec i e assessmen , bu i would be desi able o
use mo e eliable desc ip o s, like biome ics able o
e lec a en ion o s ess. Tha will be he nex s ep o
de elopmen . Ne e heless, wha e e he indica o (s)
used, he ep esen a ion should be consis en h oughou
he sys em; he Schedule , he Wo kload Lea ne , he
Wo kload Moni o and he ini ial p o iles o he pilo s
should all e alua e he wo kload using he same
pa ame e s o me hod, o a leas p o ide a no malisa ion
o he ep esen a ions. Once hose alues a e ob ained,
hey should be simpli ied in he pai s o ime and a e age
le el o a en ion equi ed ha would use he Schedule
o he o al wo kload agg ega ion a e e y momen . I
should be no iced ha he e is no in en ion o ep esen
how o when he asks a e pe o med exac ly, especially
when hey o e lap; ha is le o he pilo ’s will. The
main in e es is o de e mine he ex en o a ailabili y o
he pilo o handle mo e asks. This espec o he pilo ’s
pe sonal s a egy o he wo kload managemen no only
ollows a guideline abou RPAS in e ace usabili y o
pilo accep ance, i also means an abs ac ion o he
ep esen a ion ha decouples he sys em om any
speci ic GCS o p ocedu e, making i possible o easily
adap he concep o he e ogeneous ehicles, mission-
speci ic asks, and le els o au oma ion. A ques ion ha
may a ise is wha exac ly cons i u es a ask. I s de ini ion
is i ele an o he Schedule , bu an app op ia e le el o
g anula i y should be chosen: no oo wide o be coun e -
p oduc i e, no being able o le e age he wo kload
op imisa ion; and no oo ine o be GCS-speci ic, o
a oid an o e i ing ha could hypo hesise oo much
abou he pilo ’s indi idual wo kload s a egy. [32]
shows an example o ask analysis, aking he whole
mission as he oo o a ee ha i e a i ely b anches (e.g.
mission, phase, segmen , unc ion, ask) up o he senso y
esou ces needs. [33] also exempli ies a ho ough
cogni i e ask analysis o a conside ably mo e
demanding mili a y scena io and mission, o guide he

M.-A. Fas-Millán, E. Pas o
Copy igh © 2019 The Au ho s. Published by P aise Wo hy P ize S. .l. In e na ional Re iew o Ae ospace Enginee ing, Vol. 12, N. 1
17
design o he GCS eam, inding he unc ional
equi emen s and in o ma ion equi ed o app op ia e
awa eness. In N oM, a ask de ini ion does no need o i
a consis en concep ual agmen a ion like hese,
al hough a simila app oach could help o de ine hem.
The eason is ha while he p e ious examples seek o
op imise he pe o mance in highly demanding mili a y
scena ios, N oM ocuses on con enience o sa e
scheduling. Then, i would oughly de ine a ask like a
se o pa e n o ac ions equi ing a a he p edic able
amoun o ime, wi hou oo long wai ing imes in
be ween ha could jus i y b eaking i in o smalle asks
o le e age hose idle imes. These ask de ini ions would
be dependen on he mission, he GCS o he ATM’s
es ablished p ocedu es. A simple example o a ask in he
expe imen s was he change o link, which includes
checking i he scheduled equency keeps he link
quali y expec ed and hen applying he change a an
app op ia e momen . Ano he ask was he change o he
los link p ocedu e (LLP) i he con olle asked o i ;
o he wise, i was assumed au oma ically upda ed when
necessa y, ollowing ei he he ligh plan o assuming a
dynamic con ingency e alua ion [34]. The p ocedu e
consis ed o : eplying wi h a WILCO, consul ing an
al e na i e plan, applying he change and epo ing he
new plan o he ATC wi h a ee ex message. A hi d
example was he execu ion o a di ec o clea ance,
assumed p e iously eques ed. The pilo had o check i
he e was link co e age a ailable o he segmen and,
only i i exis ed, he pilo eplied wi h a WILCO and
hen could send he commands o he ai c a . A e ha ,
i he LLP had o be changed o he new leg, he s eps
we e simila o he change o he p ocedu e by eques .
Each o hese asks had i s own wo s -case es ima ed
ime equi ed and a wo kload le el ha would de e mine
he concu en asks allowed. The assignmen o he
ligh s can be decided be o e he ligh ; he e he sys em
would play he ole o a wo k o ce schedule , helping o
o ganise he s a ime able. Du ing he ligh , when
eques ed, he Schedule ac s as an online planne , ying
o ind he mos app op ia e pilo o ecei e a speci ic
ligh . This ans e could ha e been scheduled due o
end o wo kday, p e e ence, cons ain o co e age
issues. I could be he esul o a eques om he pilo ,
who is acing an unexpec ed con ingency o subjec i ely
conside s ha he cu en wo kload is oo high. The
p edic ion o he cogni i e wo kload is no an easy ask,
and is a he me cy o he emo ional o psychological
si ua ion; hen he pilo has he op ion o selec ligh s
and ask o hem o be ans e ed. An icipa ing ha , a
ans e could be a decision o he Wo kload Moni o ,
which conside s ha he pilo should be eleased o he
load. The supe iso can o ce ans e s o wha e e
eason. Finally, a con ol mig a ion could be a measu e
o he E en Manage i no ac i i y and eedback om
he pilo is ecei ed o he pilo is b eaching ins uc ions.
The Schedule calcula es he candida e assignmen s
ha bes i he cons ain s and p e e ences o he ligh
and also comply wi h he limi on he cu en and u u e
wo kload o he pilo . These sugges ions could be
au oma ically applied o subjec o con i ma ion by a
supe iso . A simila dynamic c ew alloca ion and
supe iso ole can be ound in [4] when he dispa che
needs o hand o any ligh o o e dedica ed suppo as
a i s o ice and he supe iso is in cha ge o ha e-
assignmen ; in N oM, he bes candida es would be
au oma ically calcula ed and he supe iso , i p esen ,
jus needs o selec and apply a sugges ion. A di e en
and in e es ing app oach, which is wo h men ioning,
appea s in [35]. On i , like in N oM, each ai c a is no
pe manen ly assigned o a eam o pilo . Ins ead, he lee
is conside ed able o mainly p oceed au onomously.
When an ai c a equi es some in e ac ion om he
ope a o , i s pe i ion is queued un il i can be se ed, and
he ai c a wai s o i s u n, pe o ming some holding
pa e n. While his kind o concep a oids any ask
o e lap among ligh s  he ope a o (s) a ends only one
a a ime, a.k.a. ehicle-based con ol  and he e o e he
wo kload is always limi ed, he ac ha he ai c a
could suddenly en e in a holding pa e n o wai o he
pilo ’s a en ion, some hing which may be unexpec ed by
con olle s, seems mo e accep able in seg ega ed
ai space. The wo kload o ecas is p edic ed based on he
asks scheduled in he ligh plan and hose added/dele ed
du ing he ligh because o ATC’s ins uc ions (a
sc eencas o his beha iou appea s in [36]). Bu i
would also ake in o conside a ion eal- ime in o ma ion
on ex e nal sou ces like wea he epo s and an in e nal
map o he wo kload buil by he Wo kload Lea ne . The
idea behind his map is o ind pa e ns o wo kload and
place hem in da e- ime-al i ude-posi ion coo dina es.
This wo kload could be in e ed om he agg ega ion
o di e en sou ces like ATC in e ac ion, wea he , a ic
densi y, link quali y, subjec i e assessmen s o
biome ics, in sho , no ask- ela ed wo kload, bu
ci cums an ial wo kload, o he pilo ’s pe cep ion o i .
This map would be ini ialised om his o ic da a which
could p o ide some clues abou he wo kload, such as
ha o e ed by Eu ocon ol’s Demand Da a Reposi o y
(DDR) da abase. Fo ins ance, [37] uses machine
lea ning o p edic he le el o wo kload om he
con olle -pilo oice communica ions. I would be
g adually upda ed om he ac i i y o he ligh s o ge a
wo kload es ima ion pic u e o he speci ic RPAS case
and he pilo ’s pe spec i e. The inal con ibu ion o his
map o he wo kload p edic ion would depend on he
p obabili y o occu ence o he sou ces o wo kload
egis e ed a he poin s in ol ed. Fo he cu en
p o o ype and expe imen s, only he scheduled asks’
wo kload is aken in o accoun o he assignmen s.
Du ing scheduling, he maximum expec ed wo kload
a any ime is no he sole de e minan o choosing a
candida e. Cons ain s and p e e ences we e men ioned
as pa o he ligh de ini ions. A cons ain could be
ela ed o he kind o mission o he ai c a model, like a
equi ed pilo ’s licence o ce i ica e. O he s could be
associa ed o au ho i ies’ egula ions, like wo king and
es hou dis ibu ion. P e e ences co espond o he
M.-A. Fas-Millán, E. Pas o
Copy igh © 2019 The Au ho s. Published by P aise Wo hy P ize S. .l. In e na ional Re iew o Ae ospace Enginee ing, Vol. 12, N. 1
18
possibili y o de ine he so-called s ages (no o be
con used wi h ligh s ages). These could be delimi ed by
empo a y o geog aphic c i e ia, he ligh s age o a he
pilo ’s will (when he limi is subjec i e o canno be
clea ly calcula ed, like wea he condi ions) and speci y
which pilo s o which condi ions should hold hose pilo s
o be assigned o he s age. Fo ins ance, a c i e ion could
be o selec pilo s wi h g ea e expe ise o some asks
( unc ional mig a ion [22]) o jus he opposi e, selec
hose who a e less expe ienced o allow hem o keep o
imp o e hei skills. Anyway, s ages only ep esen
p e e ences (i.e. so cons ain s) adding poin s lead o e
o he pilo s ha also comply wi h he i s il e o
wo kload es ic ions and any speci ied cons ain s (i.e.
ha d cons ain s); bu hese can help he ai line o mo e
a ou ably o ganise he s a ime able. Also, hey could
be de ined dynamically du ing he ligh . Mo e de ails
abou s ages can be ound in [28]. Once all he p e ious
a iables ha con ibu e o he pilo ’s assignmen ha e
been conside ed, ano he signi ican ac o should
de e mine he inal decision: o a oid hando e s as much
as possible, as his is a delica e p ocedu e implying high
cogni i e esou ces. When ecei ing a new ligh , he
pilo needs o s udy he ligh plan and c ea e a pic u e o
he ligh s a us, pending and nex asks, and i his new
equi emen o a en ion in o he es o he cu en
ligh s. Conside now a candida e pilo s ill eligible a e
wo kload limi s and cons ain s il e s ha e been passed.
Howe e , he wo kday o his pilo inishes in a pe iod
o ime ha is less han he emaining ime o he ligh ,
which would o ce ano he ans e . The e o e, i
possible, he schedule should choose a di e en pilo . A
balance o wo kload among pilo s would also con ibu e
o he mig a ion a oidance. I we do no push he pilo s
o hei limi o wo kload, any unexpec ed inc ease in i
will no easily imply a ans e . Fo ha eason, he
policy o p e e ably choosing hose pilo s wi h a lowe
o al wo kload a any ime should be ollowed. This
whole idea o he e alua ion o he cu en and u u e
s a us o he ai c a and wo kload o he pilo o
de e mine he esponse o he sys em can be ound a a
ask le el in [24], a sugges ion o a pilo / ehicle
adap i e cockpi in e ace ocused on he a he mo e
demanding needs o he comba pilo , bu wi h a inal
aim in line wi h he equi emen s o N oM concu en
pilo ing: “imp o e he pilo ’s si ua ion awa eness and
decision-making while alle ia ing wo kload”. On i , he
wo kload e alua ion will be used o selec he kind o
in o ma ion p o ided o he pilo o pe o m he ask and
balance he s imulus modali y ( isual, audi o y, hap ic).
[24] akes in o accoun he impac o he ask swi ch in
si ua ion awa eness, simila ly as N oM needs o conside
his impac when he swi ch is made be ween ligh s oo.
Bu i also ad oca es he need o e lec he indi idual
p e e ences and limi a ions o cope wi h he wo kload,
which could e ol e as expe ience is gained, some hing
ha would also be e lec ed in he N oM pilo wo kload
p o ile, which de ines each ask’s needs o a pa icula
pilo . [5] p o ides a ma hema ical model o hese
ope a o -speci ic equi emen s o ime and wo kload o
pe o m a ask, besides a ho ough e iew o he
li e a u e on di e en aspec s o he opic. The model is
based on ask wo kload le els ha a e subjec i ely a ed
and is sugges ed o he dynamic assignmen o asks o
ai c a and pilo s, op imising he dis ibu ion o asks on
he go. This concep schedules asks, while N oM
schedules ligh s. Bu he way he o al wo kload is
calcula ed a a gi en momen also akes in o accoun he
swi ching cos o cos o concu ence be ween asks,
some hing ha in N oM will be sepa a ed in o wo
di e en alues: he cos o he in- ligh ask swap and
he cos o he ligh swap; his di e en ia ion comes
om he especially high cos o he ligh swap. Ano he
di e ence is ha he p e ious pa adigm elies on he
scheduling o he asks pa ly in he e ec i eness bo h o
he ope a o o he UAV, while N oM, mainly ocused on
ci il ope a ions, p io i ises o he aspec s: balance o he
wo kload no only a a gi en momen o sa e y easons
o o a oid delays in communica ions and ins uc ion
execu ion, bu also h oughou he wo kday o a oid
bu nou ; allowing ask o a ion o main ain skills,
conside ing ha by de aul a pilo is assigned a whole
ligh , no a ask (scheduling based on pe o mance will
end o specialisa ion); and a oiding hando e s as much
as possible. One issue o be de e mined is he way o
upda e a pilo ’s wo kload p o ile. Recen expe imen s
[38] ound a dec ease in he b ain ac i i y o UAV
pseudo-pilo s as hey gained expe ience wi h he
simula o . I seems clea ha he ull po en ial o a pilo
would be unde used i he imp o emen o skills, o he
capaci y o assume a highe wo kload, is no e lec ed in
he wo kload p o ile. In he wo k men ioned, his
imp o emen was de ec ed by unc ional nea -in a ed
spec oscopy. Howe e , he Wo kload Lea ne should y
o de ec au oma ically such an imp o emen bo h in he
ime equi ed and in he cogni i e load. While his seems
qui e s aigh o wa d o he a e age ime o pe o m he
ask, clues abou he dec ease in he men al wo kload
equi ed a e no ob ious. In any case, we could in e an
inc ease in ha load om a dec ease in pe o mance. I
he e is no o he way o do i , pilo s should pe iodically
go h ough measu emen s o upda e hei wo kload
p o ile.
III.3. Moni o ing
Ha ing he sys em access he ligh plan, s a us o he
ai c a , ATC ins uc ions and pilo ’s ac ions, he GCS
Moni o will igge wa nings o eme gency p ocedu es
in case an unexpec ed si ua ion is de ec ed. I can send
eminde s oo, a sa e y measu e when con olling se e al
ai c a , which inc eases he possibili y o o ge ing a
ask, in pa icula condi ional ins uc ions ha should be
applied a a ime o posi ion ahead. Rega ding
compliance wi h he ATC ins uc ions, he use o
CPDLC ins ead o oice allows he sys em o easily
unde s and he seman ics o he message and moni o i s
execu ion, as ecommended in [8, Pa I, ch. 2, 5.9]:
M.-A. Fas-Millán, E. Pas o
Copy igh © 2019 The Au ho s. Published by P aise Wo hy P ize S. .l. In e na ional Re iew o Ae ospace Enginee ing, Vol. 12, N. 1
19
“Moni o ing au oma ion capable o ques ioning ce ain
classes o ope a o ac ions ha can po en ially
comp omise sa e y mus be designed in o he sys em”.
The e o e, i he sys em de ec s ha hese a e igno ed
o con adic ed, a e wa ning he pilo and supe iso ,
he con ol o he ai c a could be au oma ically
ans e ed, wi h no need o app o al om he cu en
pilo . This kind o moni o ing is impo an when pilo s
con ol se e al ligh s, as hey could mix in en ions, as
was shown du ing he expe imen s. I also se es as a
c osscheck in wha will p obably be single pilo
ope a ions. While CPDLC p o ides unambiguous and
pe sis en communica ion, he c oss- e i ica ion by bo h
pilo s be o e he execu ion o he clea ances has been
ecommended in manned ai c a same as is done wi h
oice communica ions [39]; in his sense, he N oM
moni o ing would play he ole o he i s o ice . An
example o his need using CPDLC is epo ed in [40],
when pilo s do no ealise he condi ion o he
condi ional clea ances and execu e he ins uc ion
immedia ely. The e was also a p oblem wi h he e ms
“AT” and “BY”, which esul ed in di e en
in e p e a ions o he condi ions [41]. Al hough hese
kinds o p oblems ha e led o a ew i ing o some o he
message elemen s, his illus a es ha , wi h a condi ion
ha can be acked, like posi ion o ime, he GCS
Moni o can check i he pilo is execu ing when equi ed
and send a wa ning o con i ma ion eques i no .
Anyway, when possible, he display ies o educe
possible inpu e o s [8, Pa I, 3.6] by au oma ing eplies
o p o iding bu ons ha apply he whole sequence o
commands associa ed o he ins uc ion. Sc eencas s o
moni o ing and wa nings can be ound in [42]. Al hough
he cu en implemen a ion o he p o o ype elies on
CPDLC, oice ins uc ions could also be acked as long
as we could use eliable oice ecogni ion o he sys em
o pa se he meaning behind he speech. [43] is a ecen
wo k ha shows he easibili y o his op ion, and [44] is
a p oposal o de ec ing non-compliance ac ions o bo h
oice and da a link messages ins uc ions. The sys em
could keep an eye on he pilo ’s heal h. The simples way
would be o ack he ac i i y in he NPI o he GCS and
ask o eedback, as a kind o dead man’s pedal, a e a
ce ain pe iod o inac i i y. I could also use biome ic
de ices, as sugges ed in SPO ConOps [4], which also
conside ed he possibili y o a ideo eed o he cockpi
o check he pilo ’s beha iou al s a e and clea ly
de e mine any possible pilo incapaci a ion and he need
o ans e con ol o g ound ope a o s o he onboa d
au oma ion. Wa nings sen by he ai c a associa ed wi h
ailu es o unexpec ed eadings should also be
highligh ed, especially o ligh s in he backg ound. F ee
ex messages o ATC wa ning abou hese issues could
be ollowed by an au oma ic dependen su eillance
(ADS) epo con aining da a like g ound ec o , ai
ec o o sho - e m in en . An LLP ac i a ion could be
ollowed by a no i ica ion o he cu en da a au ho i y i
possible (i he link is los wi h he GCS bu no wi h he
ATC) and a b oadcas o he in en o he nea by a ic.
One o he oles o he supe iso , as could be in e ed
by now, is o p o ide an op ional human con i ma ion o
selec ion o he E en Handle sugges ions. Supe iso s
could o ce assignmen s and ans e s, o e w i ing he
Schedule decisions, o c ea e s ages and cons ain s. Fo
a la ge lee , he supe iso could be assumed as he
ope a ions manage o he ai line ope a ions cen e. In
smalle lee s, his ole could be mixed wi h he one o
he dispa che , assuming o he asks like eques o
eedback ( he supe iso has a cha oom a ailable wi h
e e y pilo ), cla i ying ac ions and assis ing pilo s.
Besides he usual in o ma ion ha an ope a ions
manage has a ailable, he N oM supe iso in e ace
would p o ide: he dynamic ep esen a ion o wo kload
o ecas (sc eencas in [45]), he his o y o ac ions and
communica ions, he sequence o pilo s o a ligh , and
he log o any ele an e en like wa nings, los links o
eme gency p ocedu es loaded. Modes o le els o
au oma ion would de e mine he asks o he supe iso ;
ha way, a supe iso who is busy a ending some
eme gency could delega e o he E en Handle he
decisions o apply assignmen s, a oiding con i ma ion
eques s. Rega ding he p ocess o con i ma ion o an
assignmen sugges ion, he Schedule would p o ide an
o de ed lis o candida e pilo s ( hose le a e
conside ing he ha d cons ain s) and he alues o he
a iables ha de e mined he anking. In his way, in he
e en ha he bes op ion calcula ed could no be applied
o any eason, he supe iso could ake an in o med
decision, so ing by a iables like cu en wo kload,
wo k ime le , cu en numbe o ai c a o p e e ences
me by he pilo o he s age. The moni o ing can also be
applied o he link quali y. When de ec ing a dange ous
link deg ada ion ha could a ec he se ice, he sys em
could wa n, and sugges o assign, i possible, an
al e na i e pilo wi h be e co e age. Conside ha o
N oM, pilo s could be sca e ed a ound he wo ld.
Finally, moni o ing and wo kload aspec s a e
combined in he Wo kload Moni o . This pa o he
sys em cons an ly e alua es he cu en wo kload o he
pilo , which could be done using se e al sou ces:
biome ics (physiological pa ame e s like hea a e o
pupil dila a ion [46]; in any case a non-in asi e me hod
ha could hampe he pilo ing), he ma ch o he cu en
ask(s) pe o mance wi h he wo kload p o ile, o he
esponse ime o eply and execu e ATC ins uc ions. I i
becomes highe han ha allowed by he h esholds
obse ed, di e en measu es could be applied. These
should be ag eed wi h pilo s. A ans e o con ol when a
high peak o wo kload is de ec ed could be mo e
annoying o ecei ing and ans e ing pilo s han jus
dealing wi h his high wo kload i i is only expec ed o
las a sho ime. The da a ob ained by his module would
be used o upda e he pilo -independen map o expec ed
wo kload.
Besides, pai ing he eco d o ac i i y wi h
pe o mance-based e alua ions and/o physiological
measu emen s would allow he pilo ’s wo kload p o ile
o be upda ed.
M.-A. Fas-Millán, E. Pas o
Copy igh © 2019 The Au ho s. Published by P aise Wo hy P ize S. .l. In e na ional Re iew o Ae ospace Enginee ing, Vol. 12, N. 1
20
IV. Me hod
Ea ly expe imen s, desc ibed in [28], we e add essed
o ou lining a basic desc ip ion o he a e aged cogni i e
wo kload o each ask and ime equi ed o ca y hem
ou when using he minimalis ic GCS embedded in he
NPI. Tha way, we could es ima e he limi o wo kload
o be allowed by he Schedule , and disco e , om he
usual p oblems ound by he pa icipan s, any o he aid
ha could be implemen ed apa om hose al eady
p ojec ed. These expe imen s ied o show he p ocedu e
o e alua e and desc ibe he wo kload in a way
manageable by he Schedule , al hough i lacked an
objec i e measu emen o he cogni i e wo kload, like
biome ics, which was subs i u ed by subjec i e alues
based on he obse a ion o he pa icipan s’
pe o mance. Tha p ocedu e should be ex apola ed o
each speci ic GCS and p ocedu es in ol ed. Du ing
hose i s expe imen s, guidelines like eminde s and
moni o ing we e no a ailable. The eason is ha he
scena ios we e designed o iden i y p oblema ic le els o
wo kload and ask o e lapping. In his way, he wo kload
h eshold could be de e mined and incompa ibili ies in
ask o e lapping de ec ed. E alua ing he wo kload
wi hou he p e ious aids ollows he policy ha pilo s
should no ely on eminde s and wa nings. The sys em,
ac ing jus as a helpe o a oid any dange ous
consequences, does no y o a oid p oac i e awa eness.
In he s age o de elopmen desc ibed he e, he same
ini ial expe imen s we e ca ied ou , wi h he same
pa icipan s, bu wi h some o he aids ac i a ed. These
we e expec ed o ob iously con ibu e o educing he
numbe o e o s, bu i was an oppo uni y o check o
wha ex en , and e alua e how app op ia e was hei how
and when. Se e al mon hs had passed om he i s es s
and pa icipan s eques ed he chance o e esh he
concep s, p ocedu es and usage o he in e ace, and
could no emembe he e en s o he scena ios. I did no
seem likely ha a possible imp o emen in pe o mance
could be jus due o an inc ease in accumula ed
expe ience. Du ing he i s ound o expe imen s, he
wo kload h eshold was in ui i ely se o 10. Ha ing a
ligh wi h no asks a a gi en momen con ibu ed wi h
one uni o wo kload o ep esen he pe iodic checking
o eadings. Ne e heless, he h eshold would be easily
and g ea ly exceeded when ATC ins uc ions a i ed
while doing scheduled asks; a eques o an LLP upda e
added h ee uni s and a di ec o clea ance, ou . Some
e o s began o appea when he wo kload was o 12, so
he h eshold was educed o 6 in he second se o
expe imen s. Anyway, he scena ios we e designed o
p o oke o e whelming si ua ions, like demanding
con olle eques s a i ing while a couple o scheduled
asks in di e en ligh s we e concu en on ime, which
gene a ed se e al peaks in a ange be ween 14 and 20.
These si ua ions could be a oided in p ac ice hanks o
he wo kload map lea ned by he Wo kload Module. I
hese could no be a oided, he e is ano he h eshold o
be conside ed: he one handled by he Wo kload
Moni o , which would eques he E en Handle o ac
when an excessi e wo kload is de ec ed o a pilo .
Reducing he Schedule h eshold also educes he
possibili y o equi e ha in e en ion. Anyway, in his
case, he dec ease in he h eshold did no educe he
numbe o ligh s, which was ou in bo h scena ios, and
he Wo kload Moni o was no enabled. In summa y, he
esul s should be obse ed as ela ed o a wo s case
scena io, wi h peaks o wo kload ha would p obably no
be allowed by he sys em in p ac ice, bu able o e eal
he po en ial o he weaknesses o he ea u es and design
decision. The aids included in his se o expe imen s
we e he ollowing:
A) Link upda e eminde s. The link changes we e
scheduled a speci ic waypoin s ep esen ing he close
end o co e age o he cu en link, o he s a ing o an
a ea wi h be e quali y using a di e en equency. The
mos delica e case o co e age ansi ion would be ha
ep esen ed in Fig. 4. I he pilo upda es he link in 1,
whe e he e is no co e age o he nex equency ye ,
he e will be a link loss, and a e he pe iod o ime
speci ied by he los link imeou alue, he onboa d
compu e will execu e he LLP loaded, which could
nega i ely a ec he a ic and he ask o he
con olle s. The link would also be los i he pilo
exhaus s he bu e a ea wi hou changing he link. To
a oid hese mishaps a couple o wa nings we e
implemen ed. One o hem w i es he wo ds “LINK
CHANGE” in blue in he s ip when a scheduled link
change is 30 s ahead. Pa icipan s we e old ha his was
a eminde , no necessa ily an app op ia e momen o
change he link, as i depends on he los link imeou
(Fig. 4). They could use he ime emaining o he nex
waypoin o decide he momen . The second wa ning is
igge ed some seconds a e passing he waypoin i he
change has no been egis e ed ye . The wo ds “IN
BUFFER AREA” a e p in ed in ed, and a syn hesised
oice eads hem, p eceded by he callsign. A sc eencas
can be ound in [47].
B) Value checking. When ecei ing a clea ance, i he
pilo in oduces a command wi h a alue ha does no
ma ch ha speci ied in he clea ance, such as inco ec
waypoin , speed o al i ude, a eques o con i ma ion
showing he expec ed alue is ecei ed be o e applying
he command. I he pilo con i ms anyway, he
supe iso ecei es a no i ica ion and he con olle could
also ecei e a ee ex message no i ying abou his. A
sc eencas is shown in [39]. (This kind o p oblem
disappea s when he ins uc ion can be applied om he
same message in he CPDLC display.)
C) Unexpec ed ac ions. When an ac ion ha should be
p eceded by a clea ance is de ec ed, as be o e, a
con i ma ion is eques ed, and bo h he con olle and
supe iso a e no i ied i con i med. A sc eencas is
shown in [39].
D) Moni o ing o ins uc ions. The sys em moni o s
he execu ion o he ATC eques s only i hese ha e
been eplied wi h a WILCO. I hese mus be
immedia ely execu ed and hey a e no , he pilo ecei es
a no i ica ion a e a while. I i is s ill igno ed, he