AMBIENT-6G D5.1: Use cases, KPIs and KVIs

Towa ds s anda dized 6G connec i i y o ambien -powe ed ene gy neu al IoT de ices
Deli e able D5.1
Use cases, KPIs and KVIs
AMBIENT-6G p ojec has ecei ed unding om he Sma Ne wo ks and Se ices Join Unde aking (SNS JU) un-
de he Eu opean Union’s Ho izon Eu ope esea ch and inno a ion p og amme unde G an Ag eemen No 101192113.
Da e o deli e y: 30 h Aug, 2025 Ve sion: 1.1
P ojec e e ence: 101192113 Call: HORIZON-JU-SNS-2024
S a da e o he p ojec : 1s Jan, 2025 Du a ion: 36 mon hs
D5.1 - Use cases, KPIs and KVIs
Documen p ope ies
Documen Numbe : D5.1
Documen Ti le: Use cases, KPIs and KVIs
Edi o (s): Rubén Al una Pé ez (TEL)
Au ho s: D agan Subo ic (IMEC), Ri esh Singh (IMEC), Same Nasse (IMEC),
Hen ique Dua e Mou a (IMEC), And ey Belogae (IMEC),
Maby Johns (IMEC), Riku Jän i (AAU),
Osmel Ma ínez Rosabal (OUL), E s a hios Ka ana as (SEQ),
Mil iadis Filippou (WIN), Aimilia Ban ouna (WIN),
Sok a is Ba mpounakis (WIN), Vasiliki Lamp ousi (WIN),
Panagio is Demes ichas (WIN), Be Cox (KUL), Daniel Pöhl (NXP),
Ja ne Van Mulde s (KUL), Guus Leende s (QKS),
Benjamin J. B. Deu schmann (TUG), Lukas D’Angelo (TUG),
Rubén Al una Pé ez (TEL), Bik amji Singh (LMF),
Hamza Khan (LMF).
Con ac ual Da e o Deli e y: 31s Aug, 2025
Dissemina ion le el: PU
S a us: Final Ve sion
Re ision: 1.1
Filename: AMBIENT-6G_5.1_ 1.1
Re ision His o y
Re ision Da e Issued by Desc ip ion
0.1 17/03/2025 TEL Baseline d a o he deli e able.
0.2 24/03/2025 TEL Me ge con ibu ions o Use Cases.
0.3 12/05/2025 TEL Inpu s in In oduc ion, Technical Desc ip ion, Use Case
So A and Conclusions.
0.4 26/05/2025 TEL In e nal e iew.
0.5 24/06/2025 TEL Ex e nal e iew.
1.0 01/08/2025 TEL Final e sion.
1.1 11/11/2025 TEL Re iewed inal e sion (au ho s lis upda ed).
Page I
D5.1 - Use cases, KPIs and KVIs
Abs ac
This deli e able desc ibes ele an use cases in di e en e ical a eas whe e ene gy-neu al
de ice ope a ion can b ing pa icula bene i s in e ms o ease o use and main enance, cos ,
and ecological oo p in .
Keywo ds
AMBIENT-6G, Ambien -IoT, Ene gy-Neu al De ices, Use Cases, Key Pe o mance Indica-
o s, Key Value Indica o s.
Disclaime
Funded by he Eu opean Union. The iews and opinions exp essed a e howe e hose o he au ho (s)
only and do no necessa ily e lec he iews o AMBIENT-6G Conso ium no hose o he Eu opean
Union o Ho izon Eu ope SNS JU. Nei he he Eu opean Union no he g an ing au ho i y can be held
esponsible o hem.
In e nal e iewe s
Hen ique Dua e Mou a (IMEC)
P iyesh Pappinisse i Puluckul (IMEC)
Rubén Al una Pé ez (TEL)
Ex e nal e iewe s
And ey Belogae (IMEC)
Lie en De S ycke (KUL)
Gilles Callebau (KUL)
Page II
D5.1 - Use cases, KPIs and KVIs
Execu i e Summa y
This documen cons i u es he issue o Deli e able D5.1: ‘Use cases, KPIs and KVIs’, wi hin
he amewo k o he p ojec i led "AMBIENT-6G – Towa ds s anda dized 6G connec i i y
o ambien ly-powe ed ene gy-neu al IoT de ices" (P ojec Ac onym: AMBIENT-6G; G an
Ag eemen No: 101192113).
This documen p esen s a comp ehensi e analysis o he po en ial o Ambien In e ne o Things
(A-IoT) in enabling inno a i e and sus ainable use cases, aiming a suppo ing u u e de elop-
men s and s anda diza ion e o s wi hin he Eu opean con ex . I is s uc u ed in o i e chap e s,
each add essing a key aspec o he opic.
•Chap e 1in oduces he mo i a ion behind he wo k, ou lines he main objec i es, and
p esen s he s uc u e o he documen .
•Chap e 2p o ides a echnical o e iew o A-IoT and he co e echnologies i builds upon,
o e ing a concep ual ounda ion o he es o he documen .
•Chap e 3 e iews he cu en s a e o he a in de ining use cases o A-IoT, ocusing on
in e na ional s anda diza ion bodies and ele an Eu opean p ojec s, bo h comple ed and
ongoing.
•Chap e 4p esen s he p oposed A-IoT use cases in de ail. Each use case includes an
analysis o unc ional equi emen s, key pe o mance indica o s (KPIs), and key alue indi-
ca o s (KVIs), co e ing en i onmen al, social, economic, and inno a ion- ela ed impac s.
•Chap e 5summa izes he main conclusions d awn om he wo k desc ibed in he docu-
men .
Page III
D5.1 - Use cases, KPIs and KVIs
Con en s
1 In oduc ion 1
1.1 Mo i a ion..................................... 1
1.2 ScopeandObjec i es............................... 1
1.3 S uc u e o he Documen . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2 Ambien In e ne o Things: Technical Desc ip ion 3
2.1 Holis ic Ene gy Neu al De ice a chi ec u e . . . . . . . . . . . . . . . . . . . 3
2.1.1 Encompassing Pe o mance s. Cos s. Ma e ial Usage T ade-o . . . 4
2.1.2 Ci cui s o Ene gy S o age, Ha es ing, and Managemen . . . . . . . 5
2.1.3 Enabling Ene gy Neu al De ice Connec i i y . . . . . . . . . . . . . . 6
2.1.4 Secu e Low-powe P o ocol Design . . . . . . . . . . . . . . . . . . . 6
2.1.5 In as uc u e Enable s o Wi eless Powe T ans e . . . . . . . . . . . 7
2.2 Cloud-Edge De ice O ches a ion, O loading and On-de ice Machine Lea ning 8
3 Use Cases: S a e o he A 11
3.1 In e na ional S anda diza ion Bodies . . . . . . . . . . . . . . . . . . . . . . . 11
3.2 Eu opean Resea ch and Inno a ion P ojec s . . . . . . . . . . . . . . . . . . . 13
4 Use Cases: AMBIENT-6G Vision 16
4.1 In e p e a ion o he Use Case Analysis . . . . . . . . . . . . . . . . . . . . . 16
4.2 Elec onicShel Label............................... 22
4.2.1 Desc ip ion ................................ 22
4.2.2 Func ional Requi emen s . . . . . . . . . . . . . . . . . . . . . . . . . 22
4.2.3 Key Pe o mance Indica o s . . . . . . . . . . . . . . . . . . . . . . . 25
4.2.4 Key Value Indica o s . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
4.3 Senso sinSma Homes ............................. 28
4.3.1 Desc ip ion ................................ 28
4.3.2 Func ional Requi emen s . . . . . . . . . . . . . . . . . . . . . . . . . 29
4.3.3 Key Pe o mance Indica o s . . . . . . . . . . . . . . . . . . . . . . . 32
4.3.4 Key Value Indica o s . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
4.4 Sma B idge Heal h Moni o ing . . . . . . . . . . . . . . . . . . . . . . . . . 35
4.4.1 Desc ip ion ................................ 35
4.4.2 Func ional Requi emen s . . . . . . . . . . . . . . . . . . . . . . . . . 36
4.4.3 Key Pe o mance Indica o s . . . . . . . . . . . . . . . . . . . . . . . 39
4.4.4 Key Value Indica o s . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
4.5 Pe sonal Belongings Finding . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
Page IV

D5.1 - Use cases, KPIs and KVIs
4.5.1 Desc ip ion ................................ 41
4.5.2 Func ional Requi emen s . . . . . . . . . . . . . . . . . . . . . . . . . 42
4.5.3 Key Pe o mance Indica o s . . . . . . . . . . . . . . . . . . . . . . . 44
4.5.4 Key Value Indica o s . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
4.6 In-body o Wea able Medical Senso s . . . . . . . . . . . . . . . . . . . . . . . 47
4.6.1 Desc ip ion ................................ 47
4.6.2 Func ional Requi emen s . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.6.3 Key Pe o mance Indica o s . . . . . . . . . . . . . . . . . . . . . . . 51
4.6.4 Key Value Indica o s . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.7 Sma Ag icul u e................................. 53
4.7.1 Desc ip ion ................................ 53
4.7.2 Func ional Requi emen s . . . . . . . . . . . . . . . . . . . . . . . . . 54
4.7.3 Key Pe o mance Indica o s . . . . . . . . . . . . . . . . . . . . . . . 56
4.7.4 Key Value Indica o s . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
4.8 Asse , P oduc , Tool, and I em T acking . . . . . . . . . . . . . . . . . . . . . 59
4.8.1 Desc ip ion ................................ 59
4.8.2 Func ional Requi emen s . . . . . . . . . . . . . . . . . . . . . . . . . 59
4.8.3 Key Pe o mance Indica o s . . . . . . . . . . . . . . . . . . . . . . . 63
4.8.4 Key Value Indica o s . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
4.9 MuseumGuide .................................. 65
4.9.1 Desc ip ion ................................ 65
4.9.2 Func ional Requi emen s . . . . . . . . . . . . . . . . . . . . . . . . . 66
4.9.3 Key Pe o mance Indica o s . . . . . . . . . . . . . . . . . . . . . . . 69
4.9.4 Key Value Indica o s . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
4.10End- o-endLogis ics ............................... 72
4.10.1 Desc ip ion ................................ 72
4.10.2 Func ional Requi emen s . . . . . . . . . . . . . . . . . . . . . . . . . 73
4.10.3 Key Pe o mance Indica o s . . . . . . . . . . . . . . . . . . . . . . . 75
4.10.4 Key Value Indica o s . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
4.11 Indus ial P edic i e Main enance . . . . . . . . . . . . . . . . . . . . . . . . . 77
4.11.1 Desc ip ion ................................ 77
4.11.2 Func ional Requi emen s . . . . . . . . . . . . . . . . . . . . . . . . . 78
4.11.3 Key Pe o mance Indica o s . . . . . . . . . . . . . . . . . . . . . . . 80
4.11.4 Key Value Indica o s . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
4.12 Coope a i e Mobile Robo s . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
4.12.1 Desc ip ion ................................ 82
4.12.2 Func ional Requi emen s . . . . . . . . . . . . . . . . . . . . . . . . . 83
4.12.3 Key Pe o mance Indica o s . . . . . . . . . . . . . . . . . . . . . . . 86
4.12.4 Key Value Indica o s . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
5 Conclusions 89
Page V
D5.1 - Use cases, KPIs and KVIs
Glossa y
3GPP 3 d Gene a ion Pa ne ship P ojec .
4G ou h-gene a ion.
5G i h-gene a ion.
6G six h-gene a ion.
ADC analog- o-digi al con e e .
AI a i icial in elligence.
A-IoT Ambien In e ne o Things.
AOA angle-o -a i al.
AP access poin .
BLE Blue oo h low ene gy.
BS base s a ion.
CDMA code di ision-mul iple access.
CJT cohe en join ansmission.
CN co e ne wo k.
CO con inuous ope a ion.
COTS comme cial o - he-shel .
CRC cyclic edundancy check.
CSI channel s a e in o ma ion.
D-MIMO dis ibu ed MIMO.
DB du y-cycled ba ebone.
DLI di ec link in e e ence.
DP du y-cycled pe sis en .
DPPM dynamic powe pa h managemen .
DTLS Da ag am T anspo Laye Secu i y.
Page VI
D5.1 - Use cases, KPIs and KVIs
EMF Elec omagne ic Field.
eNB E ol ed Node B.
END ene gy-neu al de ice.
EoL end o li e.
ESL elec onic shel label.
ET Ene gy T ansmi e .
ETSI Eu opean Telecommunica ions S anda ds Ins i u e.
FCC Fede al Communica ions Commission.
FEC o wa d e o co ec ion.
FFT as Fou ie ans o m.
GDPR Gene al Da a P o ec ion Regula ion.
GNSS global na iga ion sa elli e sys em.
GPS Global Posi ioning Sys em.
HF high equency.
IC in eg a ed ci cui .
IEEE Ins i u e o Elec ical and Elec onics Enginee s.
IMU ine ial measu emen uni .
IoT In e ne o Things.
ISAC In eg a ed Sensing and Communica ion.
ISM indus ial, scien i ic and medical.
KPI key pe o mance indica o .
KVI key alue indica o .
LCA li e cycle assessmen .
LCD liquid c ys al display.
LED Ligh Emi ing Diode.
LoRa long ange.
LoRaWAN long- ange wide-a ea ne wo k.
LoS line-o -sigh .
LPWAN low-powe wide-a ea ne wo k.
LTE Long Te m E olu ion.
Page VII
D5.1 - Use cases, KPIs and KVIs
M&O managemen and o ches a ion.
MAC Medium Access Con ol.
MCU mic ocon olle uni .
MEMS mic o-elec omechanical sys em.
MIMO mul iple-inpu mul iple-ou pu .
ML machine lea ning.
NB-IoT na owband IoT.
NFC nea - ield communica ion.
NOMA non-o hogonal mul iple access.
OBU on-boa d uni .
OFDM o hogonal equency-di ision mul iplexing.
OLoS obs uc ed-line-o -sigh .
OT Ope a ional Technology.
OTA o e - he-ai .
OTA-TinyML O e - he-ai TinyML.
PHY physical.
PMIC powe managemen in eg a ed ci cui .
POS poin -o -sale.
QoE Quali y-o -Expe ience.
QoS quali y-o -se ice.
QR quick esponse.
RAM andom-access memo y.
RBAC Role-Based Access Con ol.
RF adio equency.
RFID adio equency iden i ica ion.
RIS e lec i e in elligen su ace.
ROI e u n o in es men .
RSS ecei ed signal s eng h.
RSSI ecei ed signal s eng h indica o .
RTI eusable anspo i em.
Page VIII
D5.1 - Use cases, KPIs and KVIs
echnologies con inue o e ol e, hey will play a pi o al ole in he widesp ead deploymen o
sus ainable and au onomous A-IoT sys ems.
2.1.3 Enabling Ene gy Neu al De ice Connec i i y
Ambien -powe ed ENDs su e om in e mi en ly u ning o due o insu icien powe o com-
munica ion and ope a ion o he adio and mic ocon olle uni (MCU). The inconsis ency o
he ene gy ha es ing is he majo challenge, as ENDs canno ely on a s able ha es ing a e o
p edic i hey ha e enough powe o pe o m a communica ion cycle. The mos challenging is
he suppo o ac i e ENDs ha can gene a e he signal on hei own ins ead o communica ing
using back-sca e ing. Feasibili y o di e en connec i i y p o ocols ha e been s udied in li e -
a u e: long- ange wide-a ea ne wo k (LoRaWAN) [9]–[11], Blue oo h low ene gy (BLE) [12]–
[14]na owband IoT (NB-IoT): [15]–[17].
In 6G, ac i e ENDs a e supposed o unc ion simila o legacy IoT de ices wi h a signi ican ly
lowe complexi y [18]. Legacy IoT de ices in cellula ne wo ks ope a e acco ding o NB-IoT
s anda d. The majo pa o he ene gy consump ion o NB-IoT de ices is spen on channel
access. The de aul channel access mechanism o hese de ices a e e u ning om he sleep
s a e is andom channel access, when a andom p eamble om a p ede ined se is sen o eques
a g an . Howe e , o enable low-la ency c i ical IoT applica ions, wo o he channel access
mechanisms has been in oduced: g an - ee and as uplink g an , also known as con igu ed
g an . When g an - ee access is used, de ices a e allowed o ansmi da a in andomly selec ed
esou ces wi hou nego ia ion wi h he base s a ion. The e o e, he la ency is signi ican ly
educed, bu unde an expense o a highe chance o unsuccess ul ansmissions due o collisions.
Fas uplink g an allows he base s a ion o p o ide g an s o he associa ed de ices in ad ance,
be o e da a packe s appea in hei queues. The e iciency o his mechanism signi ican ly
depends on how accu a ely he base s a ion p edic s he ac i i y o de ices. Fo ENDs, he
selec ion o he channel access mechanism is especially impo an , as di e en mechanisms
equi e di e en ene gy consump ion a di e en ne wo k condi ions. S udies [16], [17] compa e
he pe o mance o hese mechanisms in e ms o ou age p obabili y and ene gy consump ion o
ambien powe ed ENDs. Bo h s udies e eal a high po en ial o he as uplink g an scheme,
especially o p edic able a ic pa e ns.
2.1.4 Secu e Low-powe P o ocol Design
ENDs ha e an ex emely low ene gy budge and hus a e ypically esou ce cons ained, making
hem less sui ed o implemen ing adi ional secu i y mechanisms [19]. Due o limi ed p ocess-
ing powe , memo y, and ene gy a ailabili y, adding complex c yp og aphic p o ocols has been
challenging. Since communica ion ypically accoun s o he la ges sha e o powe consump ion
in IoT de ices [20], secu i y solu ions ha inc ease communica ion o e head o equi e ex ensi e
compu a ion a e ypically iewed as imp ac ical o oo cos ly o such nodes.
This pe spec i e especially applies o secu i y amewo ks ha in ol e equen o con inuous
connec ions wi h o he de ices o cloud in as uc u es. P o ocols like T anspo Laye Secu-
i y (TLS) and Da ag am T anspo Laye Secu i y (DTLS), while p o iding s ong secu i y
gua an ees in adi ional IoT o In e ne sys ems, end o impose signi ican compu a ional and
communica ion o e head ha is unusable o ul a-low-powe sys ems [21]. The handshakes,
Page 6 o 94

D5.1 - Use cases, KPIs and KVIs
key exchanges and mul iple message exchanges in ol ed can quickly d ain he limi ed ene gy
budge s o ENDs.
To o e come hese limi a ions, esea ch has been shi ed owa d ligh weigh c yp og aphic p o-
ocols and e icien key managemen schemes designed speci ically o cons ained de ices. In
he con ex o low-powe wide-a ea ne wo ks (LPWANs) — such as LoRaWAN,NB-IoT, and
Sig ox — secu i y is ypically in eg a ed a he ne wo k and applica ion laye s wi h a s ong
emphasis on minimizing ene gy consump ion [22]. Fo example, LoRaWAN employs symme ic
AES-128 enc yp ion o bo h ne wo k au hen ica ion and payload con iden iali y, elying on p e-
dis ibu ed keys and a ligh weigh join p ocedu e. NB-IoT, as pa o he Long Te m E olu ion
(LTE)/ i h-gene a ion (5G) amily, inhe i s he obus au hen ica ion and enc yp ion ea u es
o he cellula s ack while ailo ing i s signaling and powe beha io o IoT scena ios. Sig ox,
despi e i s ul a-na owband and highly cons ained message o ma , suppo s op ional payload
enc yp ion and sequence-based eplay p o ec ion mechanisms. These p o ocols exempli y how
secu i y can be designed o coexis wi h s ic ene gy budge s — limi ing handshake complexi y,
minimizing message exchanges, and elying on p e-sha ed o e icien ly nego ia ed keys. Al hough
such mechanisms may no ma ch he lexibili y o end- o-end gua an ees o adi ional In e ne -
g ade p o ocols like TLS, hey ep esen a balance be ween secu i y and ene gy e iciency o
ENDs.
In addi ion o p o ocol-le el app oaches, he e is g owing in e es in physical (PHY) laye secu i y
echniques, especially o backsca e and ene gy-ha es ing de ices. PHY secu i y le e ages
he inhe en cha ac e is ics o he wi eless medium - such as channel andomness o ha dwa e-
speci ic impe ec ions (e.g., adio equency inge p in s) - o enable ligh weigh au hen ica ion
and key gene a ion wi hou elying on hea y c yp og aphic compu a ions. These echniques
a e pa icula ly a ac i e o ul a-low-powe and in e mi en ly powe ed de ices, as hey can
educe o e en elimina e he need o ene gy-expensi e message exchanges [23]. Al hough
cu en implemen a ions ace challenges ela ed o en i onmen al a iabili y and de ice s abili y,
ecen esea ch has shown encou aging p og ess in imp o ing he eliabili y and obus ness o
PHY-laye secu i y mechanisms [24], [25].
2.1.5 In as uc u e Enable s o Wi eless Powe T ans e
RF-WPT is a p omising solu ion o powe ing la ge-scale ENDs ecosys ems, o e coming he
limi a ions o adi ional ba e y-powe ed ENDs in e ms o main enance, scalabili y, and en i-
onmen al impac . I s ands among WPT echnologies o i s abili y o suppo simul aneous
cha ging o mul iple ENDs, enable mobili y, and ope a e in non-line-o -sigh condi ions—albei
wi h some e iciency ade-o s. Unlike ambien ene gy ha es ing sys ems, he pe o mance o
WPT-enabled ne wo ks is dic a ed by he end- o-end con e sion e iciency—which encompasses
he Ene gy T ansmi e (ET), he wi eless channel, and he RF ene gy ha es ing ci cui s—
sa e y egula ions, and co e age. This is because he ETs p o ide a con ollable con ac less
powe deli e y o sus ain he long- e m ope a ion o ENDs.
Con enien ly deploying mul iple spa ially dis ibu ed ETs is c i ical o elimina e blind spo s in he
ne wo k and dis ibu e he ene gy acco ding o he applica ion equi emen s. Robo ic WPT im-
plemen a ions wi h mo ing o lying ETs [26] b ing signi ican ad an ages o e adi ional s a ic
ETs due o hei abili y o dynamically sho en he cha ging dis ance. Mo eo e , obo ic WPT
po en ially educes deploymen cos s as ewe ETs may be equi ed o accomplish he same
Page 7 o 94
D5.1 - Use cases, KPIs and KVIs
ask. Mo e impo an ly, obo ic WPT can cope mo e e icien ly wi h empo al se ice equi e-
men s. Dis ibu ed ET in as uc u es inhe en ly ha e physically la ge ape u es, which bene i
WPT h ough (i) be e egula o y compliance, (ii) highe WPT e iciency, (iii) inhe en in e e -
ence mi iga ion and channel ha dening [27]. Cohe en join ansmission (CJT) is an eme ging
pa adigm o dis ibu ed mul iple-inpu mul iple-ou pu (MIMO) a chi ec u es, p omising join ly
phase-cohe en downlink beam o ming wi h spa ially dis ibu ed ETs, bu necessi a es es ablish-
ing igh empo al synch oniza ion and phase calib a ion o ETs wi h dis ibu ed clocks [28],
[29].
Ene gy beam o ming plays a c ucial ole in enhancing end- o-end con e sion e iciency wi hou
equi ing an inc ease in he ETs’ ansmi powe . E o s o de elop cos -e ec i e beam o ming
solu ions p ima ily ocus on wo key domains: sys em a chi ec u e and signal p ocessing op-
imiza ion. The o me esea ch di ec ion ocuses on di ec modi ica ions o he a chi ec u e
o he ETs, whe eas he la e ocuses on minimizing he o e head o he signal p ocessing
algo i hm.
Popula echniques in he a chi ec u al domain include he use o low- esolu ion analog- o-
digi al con e e s (ADCs) and simpli ica ions o he analog on -end ia an enna selec ion
a chi ec u es, pa asi ic a ays, and e en implemen a ions wi hou RF chains. Mo i a ed by he
idea o hyb id beam o ming, ecen e o s ha e ocused on de eloping no el a chi ec u es ha
can e icien ly handle nume ous an ennas wi h a limi ed numbe o RF chains. This includes
lens an enna a ays [30], dynamic me asu ace an ennas [31], and e lec i e in elligen su ace
(RIS)-equipped ETs [32]. O he e o s, such as he de elopmen o mo able an ennas, ha e
been di ec ed o gi e he ETs su icien lexibili y o explo e he channel in he ques o he
bes con igu a ion ha inc eases he ha es ed powe and wi h a limi ed numbe o an ennas.
Ano he p omising esea ch di ec ion ha allows he use o ene gy-e icien low- esolu ion ADCs
is di ec link in e e ence (DLI) mi iga ion in bis a ic o mul is a ic ET in as uc u es [33], [34].
F om he signal p ocessing pe spec i e, he main challenge o ene gy beam o ming lies in bal-
ancing compu a ional e iciency wi h pe o mance gains, necessi a ing inno a i e app oaches o
mi iga e complexi y. No ably, he bene i s o accu a e channel s a e in o ma ion (CSI)-based
s a egies quickly anish, and may e en e e se, as he numbe o ENDs inc eases due o he
ene gy-demanding aining p ocess. Tha is why al e na i e beam o ming s a egies ha e been
p oposed o ely on s a is ical CSI [35], ecei ed ene gy eedback [36], and he posi ions o he
ENDs [37], which a e easie o acqui e and a y slowly.
2.2 Cloud-Edge De ice O ches a ion, O loading and On-
de ice Machine Lea ning
Cloud-edge de ice o ches a ion and o loading a e c i ical o low-powe A-IoT de ices due
o hei inhe en esou ce cons ain s, including limi ed compu a ional powe , memo y, and
ene gy supply. The adi ional cloud-cen ic AI pa adigm, which o loads all da a p ocessing o
cen alized se e s, in oduces signi ican la ency, bandwid h limi a ions, and p i acy conce ns,
making i unsui able o eal- ime and sensi i e A-IoT applica ions. Howe e , by dis ibu ing
in elligence close o he de ice, a he edge, hese limi a ions can be mi iga ed.
The p ima y impo ance o cloud-edge o ches a ion and o loading o A-IoT de ices lies in en-
abling on-de ice and edge in elligence, which signi ican ly imp o es esponse imes by elimina ing
Page 8 o 94
D5.1 - Use cases, KPIs and KVIs
he need o cons an da a ansmission o he cloud. These ansmissions a e cos ly o ENDs
Local p ocessing allows de ices o analyse senso y inpu s and execu e ac ions ins an aneously,
os e ing g ea e au onomy and e iciency. Fu he mo e, p ocessing da a a he de ice enhances
p i acy by educing he exposu e o sensi i e in o ma ion ac oss ne wo ks. I also enables pe -
sonaliza ion, allowing de ices o adap con inuously o use beha iou s and p e e ences wi hou
equi ing cons an connec i i y. C ucially, by educing equen da a ansmissions, de ice in elli-
gence con ibu es o subs an ial ene gy e iciency, a i al ac o o ba e y-ope a ed IoT de ices
and essen ial o ENDs. Howe e , he local p ocessing comes a a cos o highe complexi y
and ene gy/ esou ce consump ion. The e o e, i is necessa y o analyse whe he hese ad e se
e ec s a e compensa ed by he bene i s o local p ocessing.
On-de ice ML o ENDs is a apidly ad ancing ield ha enables in elligen da a p ocessing
di ec ly on ul a-low-powe IoT nodes. These de ices ope a e unde se e e cons ain s in e ms
o ene gy, memo y, and compu e capaci y. To add ess hese limi a ions, esea che s ha e de el-
oped highly op imized ML models using echniques, such as quan iza ion, p uning, and knowl-
edge dis illa ion, allowing in e ence o un on mic ocon olle s wi h as li le as 32–256 KB o
andom-access memo y (RAM). Two ounda ional ools in his space a e Tenso Flow Li e o
Mic ocon olle s, which enables he deploymen o compac ML models on ba e-me al sys ems
wi hou an ope a ing sys em [38] and CMSIS-NN, a lib a y o highly op imised neu al ne wo k
ke nels o ARM Co ex-M p ocesso s ha signi ican ly boos s in e ence e iciency on embedded
ha dwa e [39]. These amewo ks a e o en used oge he o deli e eal- ime ML capabili ies
on de ices wi h minimal ene gy budge s.
These capabili ies a e being applied ac oss a wide ange o use cases. In coope a i e mobile
obo ics, ENDs equipped wi h on-de ice ML can pe o m local obs acle de ec ion o ges u e
ecogni ion, enabling eal- ime in e ac ion wi hou elying on con inuous connec i i y. In sma
ag icul u e, de ices can classi y soil mois u e le els o de ec c op diseases using ligh weigh clas-
si ie s. Wea able medical senso s use on-de ice ML o de ec anomalies in hea a e o mo ion
pa e ns, while in sma logis ics, ENDs moni o package condi ions and p edic equipmen ail-
u es. E en in cul u al he i age se ings, such as museums, ambien powe ed ags wi h embedded
ML could pe sonalize isi o expe iences by ecognizing use beha iou o p e e ences. These
applica ions bene i om he abili y o p ocess da a locally, which educes la ency, p ese es
p i acy, and minimizes ene gy consump ion [40].
Recen esea ch highligh s he g owing in e es in enabling no jus in e ence bu also on-de ice
lea ning. Techniques o con inual lea ning and ede a ed lea ning a e being adap ed o ope a e
unde he cons ain s o ENDs, allowing de ices o pe sonalize models wi hou ansmi ing
aw da a. Fo example, he au ho s in [41] p o ide a comp ehensi e su ey o on-de ice ML
om an algo i hmic and lea ning heo y pe spec i e, emphasizing he impo ance o esou ce-
cons ained lea ning. Meanwhile, newe wo k explo es he easibili y o on-de ice aining and
adap a ion in eal-wo ld sys ems [42]. As 6G and A-IoT ecosys ems e ol e, on-de ice ML is
expec ed o become a ounda ional capabili y, enabling scalable, au onomous, and sus ainable
in elligence ac oss sec o s anging om indus ial au oma ion and sma homes o heal hca e
and en i onmen al moni o ing.
On he o he hand, he e a e some ad an ages o p ocessing da a a he edge. A p ima y ben-
e i is ha compu a ionally in ensi e asks can be shi ed om local de ices o mo e powe ul
edge se e s o cloud in as uc u e. O loading om he de ice e ec i ely o e comes inhe en
ha dwa e limi a ions o he IoT de ices, allowing hem o pa icipa e in complex analy ical asks
Page 9 o 94
D5.1 - Use cases, KPIs and KVIs
and machine lea ning in e ences ha would o he wise be beyond hei na i e p ocessing capa-
bili ies, ul ima ely expanding hei unc ional scope and enabling ad anced unc ionali ies like
pe sonalized se ices and in elligen au oma ion.
Thus, he e is a balance ha mus be achie ed be ween local p ocessing and o loading he ask
o he edge/cloud. Se e al echniques a e employed o acili a e cloud-edge o ches a ion and
o loading o low-powe de ices. De ice Edge O loading is a key s a egy whe e compu a ional
asks a e o loaded om esou ce-cons ained de ices o mo e powe ul edge se e s. This al-
lows complex machine lea ning in e ences o be emo ely execu ed in p oximi y o he de ices,
op imizing la ency and ne wo k bandwid h, pa icula ly o ENDs. Techniques such as model
comp ession, ede a ed lea ning, and esou ce-e icien ene gy-awa e algo i hms a e c ucial o
o e coming ha dwa e limi a ions and enabling Tiny Machine Lea ning (TinyML) on cons ained
de ices. Ene gy-awa e ask scheduling is also i al, op imizing ene gy consump ion by p io i izing
essen ial asks, dynamically alloca ing p ocesses, and le e aging echniques like du y cycling o
ensu e eliable ope a ion despi e luc ua ing powe a ailabili y om ambien sou ces. Spli lea n-
ing and ede a ed lea ning can be in eg a ed o enhance pe o mance and lea ning, add essing
bo h pe sonaliza ion and gene aliza ion in edge AI en i onmen s. O e - he-ai TinyML (OTA-
TinyML) [43] is ano he me hod ha enables emo e upda es, con igu a ion, and execu ion o
TinyML models, enhancing scalabili y and lexibili y.
Despi e he nume ous ad an ages, se e al challenges hinde he ull ealiza ion o cloud-edge de-
ice o ches a ion and o loading o low-powe de ices. Fo emos among hese a e he inhe en
limi a ions o ha dwa e esou ces, pa icula ly o ENDs, and he ene gy a ailabili y. T aining
complex models on such de ices emains imp ac ical, necessi a ing ad anced echniques like
model comp ession and ede a ed lea ning. Upda ing models dynamically on IoT de ices is also
a signi ican challenge, o en equi ing physical access o complex communica ion p o ocols.
The in e mi en powe a ailabili y in ba e y-less en i onmen s can lead o dis up ed in e -
ence and dec eased model accu acy, despi e he use o echniques like du y cycling and model
quan iza ion. Fu he mo e, deep lea ning models o en demand subs an ial memo y and compu-
a ional esou ces, c ea ing ade-o s be ween in e ence complexi y and de ice longe i y. While
edge p ocessing educes cloud dependencies, main aining model upda es and adap abili y can be
challenging i connec i i y is no consis en ly s able.
Page 10 o 94
D5.1 - Use cases, KPIs and KVIs
Chap e 3
Use Cases: S a e o he A
The pu pose o his chap e is o p o ide an o e iew o he s a e o he a on he use cases o
A-IoT. As A-IoT con inues o gain impo ance, iden i ying and analysing i s po en ial applica ions
has become a p io i y o bo h esea ch and s anda disa ion communi ies. This chap e p esen s
he main e o s o de ine A-IoT use cases om wo p ima y sou ces: in e na ional SDOs and
publicly unded esea ch and inno a ion p ojec s. Fi s , his chap e explo es how o ganiza ions
like 3 d Gene a ion Pa ne ship P ojec (3GPP) and he Ins i u e o Elec ical and Elec onics
Enginee s (IEEE) ha e con ibu ed o shaping he A-IoT landscape by iden i ying key applica ion
and echnical equi emen s. Second, i e iews ele an use cases p oposed in Eu opean esea ch
p ojec s, highligh ing eal-wo ld deploymen s.
3.1 In e na ional S anda diza ion Bodies
In ecen yea s, s anda diza ion o A-IoT has gained momen um wi h se e al o ganiza ions
ecognizing i s po en ial. 3GPP has aken a leading ole in ying o in eg a e A-IoT in o he
5G ecosys em. In addi ion o 3GPP, en i ies such as IEEE ha e also s a ed wo king on A-IoT
s anda disa ion.
3GPP has ecognized A-IoT as an IoT echnology wi h he po en ial o enable la ge-scale
connec i i y wi h ne ze o ene gy consump ion. The concep is based on he in eg a ion o
ENDs in o cellula and non-cellula ne wo ks, enabling ubiqui ous connec i i y in a numbe o
applica ions.
In Release 18, 3GPP ini ia ed p elimina y s udies on A-IoT as pa o i s e o s o enhance
5G. Release 18 laid he g oundwo k by explo ing he po en ial applica ions o ene gy-ha es ing
IoT de ices. The s udy i em Technical Repo (TR) 38.848 “S udy on Ambien IoT (In e ne
o Things) in RAN” [44] add esses he easibili y o mee ing design a ge s o ele an use
cases o A-IoT. The s udy conside ed a ious aspec s (including de ice ypes, communica ion
me hods, and ne wo k equi emen s) o suppo he ope a ion o A-IoT de ices in 5G. I also
iden i ied se e al use cases o A-IoT, emphasizing hei applicabili y in scena ios whe e ba e y
eplacemen is challenging. Mo e speci ically, wo main use case g oups we e p oposed. The
i s g oup, o G oup A, is based on he deploymen en i onmen , and consis s o h ee sub-
ca ego ies: Indoo , Ou doo and Indoo /Ou doo . The second g oup, o g oup B, is based on
he unc ionali y/applica ion o he use case, and i is o med by ou subca ego ies: In en o y,
Page 11 o 94

D5.1 - Use cases, KPIs and KVIs
Sensing, Posi ioning and Command. By combining bo h g oups, eigh di e en ca ego ies a e
gene a ed, which we e called ep esen a i e Use Cases ( UC) in TR 38.848 [44]. Table 3.1
de ines hese eigh UCs.
Table 3.1: Use cases de ined in TR 38.848 [44].
UC Desc ip ion
Indoo In en o y Disco e wha goods a e p esen in a speci ic a ea. A-IoT de ices a ached o
hese goods can epo an iden i ie and addi ional in o upon eques om he
ne wo k.
Ou doo In en o y
Indoo Sensing A-IoT de ices a e associa ed wi h a senso . A-IoT de ices send da a ob ained by
he senso igge ed by di e en condi ions (pe iodically, igge ed by he ne wo k,
when he A-IoT de ice is comple ely powe ed...).
Ou doo Sensing
Indoo Posi ioning De e mine he loca ion o ce ain goods. A-IoT de ices a ached o hese goods
epo in o ma ion like iden i ica ion o loca ion.
Ou doo Posi ioning
Indoo Command A-IoT de ice is associa ed wi h an ac ua o . Commands o con ol he ac ua o
a e usually sen by he ne wo k.
Ou doo Command
In Release 19, 3GPP con inued o de elop he ision in oduced in Release 18. While he
e m “Ambien IoT” is no always explici ly used, he use cases being discussed align closely
wi h i s p inciples: massi e-scale deploymen o low-cos , low-powe de ices wi h in e mi en
o passi e connec i i y o mobile ne wo ks. These use cases a e being explo ed wi hin se e al
3GPP wo king g oups, mos no ably SA1 (Se ices) and SA2 (Sys em A chi ec u e).
In TR 22.840 "S udy on Ambien powe -enabled In e ne o Things" [45], 3GPP p esen ed
a easibili y s udy on he po en ial suppo o A-IoT wi hin 3GPP sys ems. This documen
ep esen ed he i s o mal e o wi hin 3GPP o analyse he applicabili y o A-IoT o exis ing
and u u e mobile ne wo k. The objec i e o TR 22.840 was o e alua e use cases ha e lec
ealis ic applica ions o A-IoT ac oss mul iple e ical domains, and o iden i y he echnical
equi emen s and limi a ions in cu en 3GPP speci ica ions. The epo ocused on explo a o y
analysis o suppo u u e no ma i e wo k.
In Technical Speci ica ion (TS) 22.369 "Se ice equi emen s o ambien powe -enabled IoT"
[46], which used he ou come om TR 22.840, he unc ional equi emen s co esponding o
he UCs in TR 38.848 we e de ined. These equi emen s co e aspec s such as communica-
ion, posi ioning and loca ion se ices, de ice and se ice managemen , in o ma ion collec ion,
ne wo k capabili y exposu e, cha ging models, and secu i y. Mo eo e , TS 22.369 also p o ides
some gene al KPI de ini ions o some o he UC.
In addi ion o he s anda diza ion e o s wi hin 3GPP, he IEEE is also ocusing on A-IoT. The
IEEE 802.11bp Ambien Powe (AMP) Task G oup s a ed a echnical in es iga ion o suppo
A-IoT de ices o e wi eless local a ea ne wo ks (WLANs). In Ma ch 2023, he g oup published
a echnical epo i led “Suppo o Ambien IoT De ices in WLAN” (IEEE 802.11-23/0436 0)
[47], which explo es he equi emen s, challenges, and possible enhancemen s o he IEEE 802.11
s anda d o include A-IoT. The epo iden i ies use cases ha show how A-IoT de ices can be
in eg a ed in o WLAN en i onmen s. Each use case imposes di e en equi emen s in e ms o
ene gy e iciency, da a collec ion equency, ne wo k access p ocedu es, and de ice iden i ica ion
mechanisms. Table 3.2 shows a summa y o hese use cases.
Page 12 o 94
D5.1 - Use cases, KPIs and KVIs
Table 3.2: Use cases de ined in IEEE 802.11-23/0436 0 “Suppo o Ambien IoT De ices in WLAN”[47].
Use Case Desc ip ion
Sma Manu ac u ing Real- ime acking o asse s, wo ke s, and p oduc s using low-cos and ba e y-
ee de ices o dense deploymen .
Da a Cen e En i onmen al and asse moni o ing h ough wi eless senso s. Sha es simi-
la equi emen s wi h sma manu ac u ing: ba e y- ee, long-li e, and low-
main enance.
Logis ics/Wa ehouse Accu a e in en o y checks and so ing using ba e y- ee ags. Needs high iden-
i ica ion accu acy, small size, and suppo o mo ing i ems.
Sma Home Compac low-powe senso s o sa e y, com o , and loca ing pe sonal i ems wi hin
home en i onmen s.
Sma Ag icul u e Moni o ing en i onmen al condi ions and asse s in la ge ou doo a eas. Requi es
wide co e age, ba e y- ee ope a ion, and he abili y o handle housands o
de ices.
Indoo Posi ioning Na iga ion and acking in la ge indoo spaces. Needs high-densi y, low-cos ags
wi h 1–3 m ho izon al and 1–2 m e ical accu acy.
Sma Powe G id Moni o ing equipmen and en i onmen al condi ions in subs a ions and ansmis-
sion lines wi h ba e y- ee, low-main enance senso s wi h long- ange co e age.
F esh Food Supply Chain T acking o ood anspo condi ions using ul a-low-cos , ba e y- ee de ices
wi h long in e als be ween da a ansmissions.
3.2 Eu opean Resea ch and Inno a ion P ojec s
Beyond s anda diza ion ac i i ies in 3GPP and IEEE, se e al Eu opean esea ch and inno a ion
p ojec s ha e con ibu ed o he de ini ion o use cases aligned wi h A-IoT. Al hough no all o
hese p ojec s explici ly use he e m "Ambien IoT", many o hem add ess i s key enable s.
Table 3.3 p o ides an o e iew o some o he Eu opean p ojec s ha p esen use cases wi h
s ong compa ibili y wi h A-IoT.
The ADROIT6G p ojec ocuses on enabling dis ibu ed in elligence o 6G AI and cloud-na i e
ne wo ks. Some o i s use cases de ined in [48], [49] can also be applied o A-IoT:
•Assis ing Fi s Responde s: ul a-low-powe wea ables can be used o ansmi medical
and en i onmen al da a wi hou ba e y eplacemen ,
•Rail Au oma ion suppo ed by Non-Te es ial Ne wo ks: A-IoT nodes can ha es ambien
ene gy and send epo s ia sa elli e links, minimizing main enance cos s,
•Collabo a i e Cons uc ion Robo s: sel -powe ed senso s can con inuously epo s a us
da a o nea by obo s in he cons uc ion en i onmen , gi ing obo s he necessa y con ex
o coo dina e mo emen s, a oid obs acles, and adjus asks as si e condi ions change.
The Hexa-X p ojec , which is a lagship p ojec o 6G ision, de ines i e main use case ami-
lies: Sus ainable De elopmen , Imme si e Telep esence o Enhanced In e ac ions, Local T us
Zones o Human and Machine, Massi e Twinning, and Robo s and Cobo s [50]. A-IoT can
complemen ou o hem:
Page 13 o 94
D5.1 - Use cases, KPIs and KVIs
•Sus ainable De elopmen : sel -powe ed senso s and ac ua o s can be deployed in he 6G
ecosys em, enhancing ene gy e iciency and educing main enance cos s,
•Local T us Zones o Human and Machine: ENDs can p o ide sho - ange da a collec ion
and dis ibu ed in elligence a he edge, hus enabling low-la ency decision making,
•Massi e Twinning: ene gy-ha es ing senso a ays can enable digi al- win pla o ms wi h
con inuous s a us upda es on in as uc u e o indus ial asse s, ensu ing ha i ual epli-
cas emain up- o-da e wi hou powe -hung y o complex logis ics,
•Robo s and Cobo s: ENDs can be embedded in wo kspaces o guide and coo dina e and
command collabo a i e obo s.
In he Hexa-X-II p ojec , which con inues he wo k pe o med in Hexa-X, many o he use case
amilies mi o hose o Hexa-X, and ou —Collabo a i e Robo s, Physical Awa eness, Digi al
Twins, and T us ed En i onmen s— emain compa ible wi h A-IoT [51].
The In ellIoT p ojec de elops in eg a ed, dis ibu ed, human-cen ed and us wo hy IoT
amewo ks applicable o a numbe o scena ios [52], [53]. Th ee use cases in which A-IoT
can be po en ially applied a e p oposed in his p ojec :
•Ag icul u e: A-IoT de ices can ha es ene gy om sunligh , he mal g adien s, o RF
sou ces o p o ide en i onmen al da a, educing main enance cos s and suppo ing sus-
ainable a ming,
•Heal hca e: A-IoT wea ables can be used o moni o medical da a wi hou equen ba e y
eplacemen s, enabling long- e m emo e ca e and minimizing elec onic was e.
•Manu ac u ing: A-IoT de ices embedded on equipmen can elimina e ba e y main enance
o senso s ha moni o machine pe o mance, enabling p edic i e main enance and p ocess
op imiza ion while in eg a ing in o exis ing Indus ial IoT ne wo ks.
In he REINDEER p ojec , ou applica ion scena ios a e in oduced: Adap a i e Robo ized Lo-
gis ics, Human-Machine In e ac ion in Ca e En i onmen s, Imme si e En e ainmen o C owds,
and Sma Homes [54]. A-IoT can be applied o h ee o i s ou use cases.
•Adap i e Robo ized Logis ics: END can be used in wa ehouse in as uc u e and au-
onomous ehicles, enabling con inuous acking o goods and ou e op imiza ion wi hou
ba e y eplacemen ,
•Human-Machine In e ac ion in Ca e En i onmen s: senso s powe ed by body hea o am-
bien ligh can moni o medical and en i onmen al condi ions, suppo ing assis ed li ing
wi hou equen main enance,
•Sma Homes: A-IoT senso s can be used o moni o o modi y he en i onmen al con-
di ions o he house, as well as o moni o i s s uc u al condi ions o enable p edic i e
main enance.
In he SUPERIOT p ojec , h ee main applica ion scena ios a e analysed: Sma Tags and
Labels, La ge-scale Sensing and Ac ua ion, and Enhanced IoT Communica ions in Demanding
En i onmen s [55]. Each o hese h ee applica ion scena ios a e di ided in o mo e concise use
cases. A-IoT echnology can be applied o he h ee main applica ion scena ios o SUPERIOT
as ollows:
Page 14 o 94
D5.1 - Use cases, KPIs and KVIs
•Sma Tags and Labels: A-IoT ags can be a ached o ma ke p oduc s, heal hca e
equipmen in hospi als and medical cen es, indus ial equipmen in ac o ies and p oduc-
ion lines, o e en pe s o acking and moni o ing. This enables sus ainable and low-cos
acking and loca ion in mul iple scena ios.
•La ge-scale Sensing and Ac ua ion: ENDs can be used o moni o and ac ua e o e la ge
like sma buildings and acili ies, cons uc ion en i onmen s, sma ci ies and u al loca-
ions. This is possible hanks o he low main enance equi emen s and long au onomy o
ENDs.
•Enhanced IoT Communica ions in Demanding En i onmen s: A-IoT is di icul o imple-
men in his applica ion scena io due o he pa icula ly high le el o secu i y and p i acy
equi ed. Howe e , ENDs can be used as senso s in on-body/in-body applica ions o in
emo e zones o deploy e icien , low-main enance and long-las ing ne wo ks.
Table 3.3: Eu opean esea ch and inno a ion p ojec s.
P ojec Name Focus Rele ance o A-IoT.Re e ence
ADROIT6G AI and cloud-na i e design o 6G. In ol es in elligen , con ex -awa e con-
nec i i y o massi e low-cos de ices.
[48], [49]
Hexa-X Flagship o 6G ision and a chi-
ec u e.
In oduces concep s like ambien connec-
i i y, massi e IoT, and 6G loca ion se -
ices.
[50]
Hexa-X-II Con inua ion o Hexa-X wi h
sys em-le el pilo s.
Conside s ex eme ene gy e iciency and
ul a-dense IoT de ices deploymen in 6G.
[51]
In ellIoT In eg a ed, dis ibu ed, human-
cen ed and us wo hy IoT
amewo k in heal hca e, a ming
and indus y.
Ene gy-awa e IoT ne wo k managemen ,
applica ion unc ion alloca ion, edge com-
pu ing, AI/ML...
[52], [53]
REINDEER Cell- ee, dis ibu ed beam o m-
ing and dis ibu ed in elligen p o-
cessing.
Enables ene gy-e icien , p og ammable
wi eless en i onmen s o ubiqui ous sens-
ing and acking, as well as in e ac ion
wi h ENDs.
[54]
SUPERIOT Sus ainable, elixble and adap -
able IoT sys em based on op ical
and adio communica ions.
De elops IoT nodes based on ene gy ha -
es ing o enhance ene gy e iciency and
en i onmen al sus ainabili y.
[55]
Page 15 o 94
D5.1 - Use cases, KPIs and KVIs
4.2 Elec onic Shel Label
Table 4.3: Use Case Families o Elec onic Shel Label use case.
Use Case Families
Ha d- o- each de ices Massi e-scale senso ne wo ks Ex eme-li e ime applica ions
Table 4.4: END classes o Elec onic Shel Label use case.
END Classes
UP ENDs DB ENDs DP ENDs CO ENDs
4.2.1 Desc ip ion
The use o elec onic shel label (ESL) ags is becoming mo e widesp ead on e ail shel es
ha house la ge numbe s o p oduc s. The ESL ag p o ides in o ma ion abou he p ice o a
p oduc , along wi h o he po en ial pa ame e s, such as he expi a ion da e, sales o e s, ba
codes, quick esponse (QR) codes, e c. ESL ags a e usually in on o a p oduc o a g oup o
p oduc s on he shel , p o iding a con enien way o manage o upda e p oduc in o ma ion o
se dynamic p icing o e - he-ai (OTA), using a mobile handheld eade (connec ed o a adio
uni e.g., Mac o bases a ion), o s a ic on-si e (Mic o bases a ion), o emo e eade (connec ed
ia an in e media e node) accompanied by a use - iendly applica ion in e ace. This app oach
sa es conside able ime and labou cos s compa ed o commonly used s a ic labels (e.g., pape
labels), while also enhancing he shopping expe ience [45].
A ypical ESL ag is composed o se e al componen s ha a e equi ed o communica ion and
in o ma ion display, including he RF module, MCU, ba e y, and display uni . To manage he
in o ma ion on he ESL ag, communica ion occu s ia a wi eless link be ween he ag and he
eade connec ed o a se e applica ion. Some examples o wi eless echnologies include RF,
in a ed, blue oo h [56], and isible ligh [57], whe e each echnology o e s a di e en le el o
eliabili y and ange.
Among he di e en a ailable echnologies, adio equency iden i ica ion (RFID)-based com-
munica ion o e s a signi ican ad an age o low in as uc u e cos s. Howe e , his comes a
he expense o poo e quali y-o -se ice (QoS) compa ed o o he communica ion echnologies,
which ha e ex ensi e p o ocol s acks, dedica ed spec a, and cen alized con ol o suppo OTA
communica ion.
4.2.2 Func ional Requi emen s
Communica ion
ESL ags mus suppo unique iden i ica ion wi hin a local se ice a ea. The ag iden i ie can
include hi d-pa y componen s in addi ion o he de ice unique iden i ie (UID) assigned by
he co e ne wo k (CN). The hi d pa y can ensu e he uniqueness o ag UIDs wi hin hei
Page 22 o 94

D5.1 - Use cases, KPIs and KVIs
I em I1 I em I2
CN/ESL ne wo k
ESL ags
In e media e node
Mobile handheld
eade
Remo e eade
Radio uni On-si e eade
Figu e 4.1: Illus a ion o Elec onic Shel Label use case.
local se ice a ea, such as wa ehouses, shopping malls, g oce y s o es, e c, and hus elax he
cons ain s on ensu ing uniqueness on a global scale.
The abundance o ags wi hin a small se ice a ea manda es ha communica ion p o ocols
should include ea u es o obus esou ce selec ion, such as e icien andom access and e -
ec i e collision esolu ion echniques o suppo bulk ope a ions, e.g., bulk w i e ope a ions.
Addi ionally, he wi eless channels expe ience empo a y shadow ades due o obs acles, i.e.,
mo ing ca s, humans, obo s nea shel es, aisles, e c. The e o e, i is essen ial o suppo
mul i- eade ope a ion o ci cum en empo a y co e age holes and ensu e high a ailabili y.
Posi ioning/Loca ion
Gi en ha in his use case all ESL-based ENDs ha e s a ic loca ions, upda es o hei spa ial
posi ioning du ing un ime a e no equi ed. A hi d pa y ensu es he mapping o ESL ags and
hei spa ial con ex in a da abase, which is main ained as pa o he applica ion’s unc ionali y.
Managemen
Rega ding managemen , he sys em mus be equipped wi h ea u es and mechanisms o ensu e:
•pe manen disabling, decommissioning, and/o o -boa ding o ENDs,
• empo a y enabling and disabling,
•ac i a ion-on-demand o newly deployed ENDs,
•so wa e/ i mwa e upda es ia cen alized, au oma ed p o isioning and con igu a ion,
•g oup ope a ion,
•secu i y managemen (a he CN and applica ion le els, a a minimum),
• aul de ec ion,
Page 23 o 94
D5.1 - Use cases, KPIs and KVIs
•mul i- eade ope a ion, and
•ENDs owne ship managemen , pa icula ly i he de ice UID includes hi d-pa y compo-
nen s.
Ne wo k Secu i y
The ne wo k mus be equipped wi h obus secu i y p o ocols designed o low-powe IoT sys ems
o ensu e he in eg i y and con iden iali y o he in o ma ion handled by ENDs. The sys em
should suppo s ong au hen ica ion p ocedu es o ESL by ne wo k nodes, as well as addi ional
au hen ica ion p ocedu es o ne wo k nodes a he ag/de ice side, i he ENDs’ capabili ies
allow.
Ene gy
The ESL-based ENDs would ypically be placed in indoo en i onmen s, o a he e y leas , in
a eas whe e na u al ambien ene gy sou ces a e limi ed. The e o e, o ope a e e ec i ely, ENDs
mus suppo ene gy ha es ing om RF, isible ligh , o bo h. To ensu e eliable ope a ion and
longe ansmission du a ion, ENDs mus inco po a e a small ene gy s o age uni (e.g., capaci o ,
mic o-ba e y, o supe capaci o ) o main ain unc ionali y du ing empo a y una ailabili y o
ambien ene gy sou ces.
Fo ESL ags in ended o ou doo use, he s o age uni mus be obus enough o wi hs and
co osi e en i onmen al condi ions (such as mois u e, unmanned ehicle (UV) adia ion, dus ,
e c.) o e ex ended pe iods. Especially in applica ions in ol ing dynamic p icing, he ENDs
should suppo ea u es like apid cha ging unde low inpu powe condi ions ( anging om µW
o mW) o acili a e ela i ely s enuous ansmission ope a ions.
Use -Sa e y and Robus ness
The use case is p edominan ly an indoo scena io, and he e o e, a ia ions in humidi y and
empe a u e should no be d as ic. Consequen ly, he ENDs can be designed o ope a e unde
mo e elaxed condi ions, wi h a ela i ely na owe ange o empe a u e and humidi y. On
he o he hand, he use case equi es equen handling o p oduc s by cus ome s (e.g., picking
up, ouching, e c.), which may expose he ENDs o equen shocks and ib a ions. F om his
pe spec i e, he ENDs should be designed wi h shock and ib a ion-abso bing ma e ials o be
obus enough o wi hs and such condi ions. Addi ionally, hey should be compac enough o
main ain use com o and objec unc ionali y, bu no so compac ha hey cause di icul y
o cus ome s when eading in o ma ion om he ESL-equipped liquid c ys al display (LCD).
The mos impo an conside a ion is ha END ags should be manu ac u ed using non- oxic
ma e ials.
AI/ML
Fo ex emely passi e ENDs, employing AI/ML echniques on he de ices would be challenging
due o unce ain y in ene gy a ailabili y, limi ed ene gy s o age, and low p ocessing capabili ies.
Howe e , o ela i ely capable ENDs, such echniques can be deployed o unde s and cus ome
engagemen , beha iou al choices, and p oduc demand, especially i he ESL ag is supplemen ed
Page 24 o 94
D5.1 - Use cases, KPIs and KVIs
wi h addi ional ha dwa e, senso s and in as uc u e, such as ene gy sys ems o suppo la ge
p ocessing cos , o senso like in a ed senso o de ec cus ome p oximi y o in e es .
Al e na i ely, he AI/ML echniques can be deployed on he eade side (base s a ion, o in-
e media e use equipmen (UE)), o in he cloud. These echniques can enhance cus ome
engagemen analysis, helping o boos business ou pu . Addi ionally, hey can assis wi h chan-
nel and adio condi ions es ima ion and e o co ec ions, hus imp o ing he communica ion
link.
4.2.3 Key Pe o mance Indica o s
Table 4.5: KPIs o Elec onic Shel Label use case.
De ice KPIs Ne wo k KPIs
KPI Value KPI Value
Sensing accu acy N.A. End- o-end la ency <1 s
De ice au onomy >20 yea s Age o in o ma ion N.A.
A g. powe consump ion <10 µW Packe loss a e <1 %
Peak powe consump ion <100 µW Message size <100 By es
Communica ion ange <50 m
Se ice a ea dimension <20000 m2
Max simul aneous de ices <30000 de ices
T ans e in e al 0.3 - 2 hou s
De ice speed S a ic
Use -expe ienced da a a e 0.8 kbps
Applica ion KPIs AI/ML KPIs
KPI Value KPI Value
Communica ion se ice a ailabili y > 99.9 % T aining complexi y N.A.
Posi ioning/Loca ion se ice a ailabili y No equi ed In e encing accu acy N.A.
Posi ioning/Loca ion se ice accu acy N.A. In e encing la ency N.A.
Command se ice a ailabili y No equi ed
AI/ML capabili ies No equi ed
4.2.4 Key Value Indica o s
En i onmen al KVIs
Impac Sco e: 3 T adi ional labelling me hods ypically ely on RFID-based echnology o
manual labelling, bo h o which ha e en i onmen al impac s. RFID in eg a ion equi es p ecise
implemen a ion, and o en includes ba e ies which need o be eplaced pe iodically. This in-
c eases cos s and equi es addi ional human e o . Mo eo e , he communica ion links a e no
always op imal, leading o highe ene gy consump ion (due o inc eased collisions o e ans-
missions), as well as he need o dense deploymen o eade s, which inc eases in as uc u e
cos s, ma e ial cos s, and pu s addi ional s ain on he esou ces.
Page 25 o 94
D5.1 - Use cases, KPIs and KVIs
A he same ime, ESLs con ibu e o sus ainabili y by educing pape was e ( ypically used in
pape ags), as well as minimizing ood and p oduc was e h ough be e managemen , acking,
and upda es.
On he o he hand, manual labelling equi es a signi ican human p esence, leading o delays and
being highly p one o e o s. This esul s in was e o human e o , which could be minimized
h ough au oma ion wi h be e -equipped echnology, such as ESL.
Social KVIs
Impac Sco e: 3 ESLs ha e a modes social impac by enhancing bo h he cus ome and shop-
ping expe ience and s o e e iciency. Thei use in businesses inc eases anspa ency by p o iding
egula , as , and accu a e p icing upda es. They also sa e ime by educing human e o s and
in equen cus ome que ies associa ed wi h pape ags and can ensu e as e checkou s i ESLs
in eg a ed wi h he poin -o -sale (POS) sys em. Fu he mo e, wi h dynamic p icing s a egies,
ESLs can help accele a e he sale o p oduc s wi h nea -expi y da es h ough a ge ed discoun s
and sales, hus educing was age.
Economic KVIs
Impac Sco e: 4 ESLs p o ides a signi ican economic impac o e aile s, s o e owne s, o
wa ehouse manage s by educing cos s and imp o ing e enue. Al hough he e may be subs an-
ial ini ial cos s o in as uc u e deploymen , hese can be ou weighed by he gains, especially
wi h la ge-scale u iliza ion in supe ma ke s o hype ma ke s, mega ma s, malls, and simila
enues. ESLs o e a highly au oma ed solu ion ha educes p icing e o s and inco ec p icing,
signi ican ly minimizing inancial losses. They also lowe cos s associa ed wi h labou o manual
label changes o o de ul ilmen when in eg a ed wi h he POS sys em, as well as ma e ial cos s
o pape , ink, and p in ing used in adi ional pape ags. Fu he mo e, ESLs ensu es e icien
in en o y managemen by enabling eal- ime acking and educing was e. Addi ionally, hey
acili a e dynamic p icing, which can po en ially inc ease e enue h ough a ge ed p omo ions
and p ice op imiza ion.
Inno a ion KVIs
Impac Sco e: 4 ESL ags powe ed by ambien sou ces unning on cellula echnologies can
o e a wide ange o di e se and ich ea u es. Some o hese ea u es can be emo ely managed
i no di ec ly inco po a ed in o he ESL ags due o hei small o m ac o . This allows
businesses o op imize ope a ions, educe cos s, and minimize was e. The da a collec ed om
ags can be combined wi h o he senso s, such as empe a u e o in a ed senso s, depending
on he use case, o enable imp o ed da a analysis and p edic ion. Fo example, a empe a u e
senso could allow ESL ags o dynamically adjus expi a ion imes/da es o modi y he p ices
o empe a u e-sensi i e ood i ems.
In summa y, he inno a ion will be d i en by he ollowing indica o s, which will p opel he
adop ion o ambien -powe ed ESL ags in he e ail wo ld:
•ba e y-less ( educes cos , easy adop ion and o e s mo e lexibili y),
•u iliza ion cellula echnologies’ ea u es, cen alized con ol, eliable communica ion link,
ich da a collec ion, analysis, and p edic ion ea u es.
Page 26 o 94
D5.1 - Use cases, KPIs and KVIs
En i onmen al
Social
Economic
Inno a ion
12345
Figu e 4.2: Impac Sco es o Elec onic Shel Label.
Page 27 o 94

D5.1 - Use cases, KPIs and KVIs
4.3 Senso s in Sma Homes
Table 4.6: Use Case Families o Senso s in Sma Homes use case.
Use Case Families
Ha d- o- each de ices Massi e-scale senso ne wo ks Ex eme-li e ime applica ions
Table 4.7: END classes o Senso s in Sma Homes use case.
END Classes
UP ENDs DB ENDs DP ENDs CO ENDs
4.3.1 Desc ip ion
In he las decade, sma homes ha e seen an inc ease in popula i y due o hei imp o ed
com o and quali y o li e, and he way hey can ans o m ou daily li es. In such sma homes,
e e yday de ices and appliances (such as he mos a s, ligh s, e ige a o s, o ens, e c.) ge
connec ed o a cen al moni o ing and con ol uni , o en connec ed o he In e ne o emo e
access. This eshapes he way we engage wi h ou home en i onmen , po en ially enhancing
ene gy e iciency, sus ainabili y, home secu i y and heal h ou comes. Howe e , ge ing e e yday
objec s sma by connec ing hem o he in e ne makes hem ulne able o hacking (secu i y
and p i acy), has a ce ain le el o complexi y, and cos s money. This use case add esses
ano he hidden challenge inside sma homes. Sma objec s equi e ex a ha dwa e o (i)
communica ing wi h he cen al con olle in case o wi eless connec ion, and o (ii) p ocessing
he incoming o ou going in o ma ion o o om he de ice, inc easing he powe consump ion
a each uni . Resea ch shows ha in sma ligh ing [58], [59] he s andby powe consump ion o
he elec onics o sma Ligh Emi ing Diode (LED) bulbs’ con ol and adio communica ion
lies a ound 0.4 W, esul ing in an addi ional 3.5 kWh pe yea . The ac i e powe o a LED bulb
lies a ound 10 W o a 800 Lm lamp. This means ha e e y 20 hou s, a sma lamp uses as
much as a gene ic LED lamp in ac i e mode du ing 1 hou . RF backsca e communica ion
in indoo en i onmen s in combina ion wi h ul a-low powe designs can coun e ac his unseen
ene gy consump ion. In his use case, we explo e how we can se up an indoo backsca e ing
ne wo k o moni o ing o ambien pa ame e s (ligh , empe a u e, ai quali y). These de ices
ope a e wi hou an in eg a ed powe supply by le e aging backsca e communica ion, whe ein
hey modula e and e lec inciden RF signals emi ed by an ex e nal RF sou ce o ansmi
da a. The in o ma ion ha will be added o he channel consis s o small da a packe s wi h
a low upda e a e. An example o he sma home wi h backsca e ing in eg a ion is depic ed
in Figu e 4.3. I consis s o se e al ENDs, cap u ing, p ocessing and ansmi ing he ecei ed
senso da a and a cen al con olle ha ecei es, demodula es and p ocesses he backsca e ed
senso da a.
Page 28 o 94
D5.1 - Use cases, KPIs and KVIs
Figu e 4.3: Illus a ion o Senso s in Sma Homes use case. Typical examples a e empe a u e, humidi y,
ai quali y, wa e and powe consump ion and p esence de ec ion senso s.
4.3.2 Func ional Requi emen s
Communica ion
The sys em shall enable ul a-low-powe , low- h oughpu communica ion be ween semi-passi e
backsca e ENDs and UEs by le e aging a cus om backsca e modula ion scheme. The p o o-
col mus suppo su icien da a a es o acili a e he pe iodic ansmission o senso eadings
o a cen al con olle , while main aining minimal ac i e powe d aw on he ag side. Commu-
nica ion should be op imized o sho - ange, asymme ic links, wi h emphasis on ag simplici y
and eade -side p ocessing complexi y.
The cus om backsca e communica ion p o ocol shall inco po a e ligh weigh bu e ec i e syn-
ch oniza ion p imi i es and e o de ec ion mechanisms, such as p eamble-based iming align-
men and cyclic edundancy check (CRC), o ensu e da a in eg i y. To accoun o he challenges
inhe en o e lec i e and in e e ence-p one RF en i onmen s, he p o ocol should suppo adap-
i e symbol iming, collision a oidance h ough ime o equency domain sepa a ion, equency
hopping, code di ision-mul iple access (CDMA), o enable scalable, collision- esis an ope a ion
and obus en elope de ec ion s a egies. Addi ional edundancy o o wa d e o co ec ion
(FEC) may be in eg a ed as needed based on he applica ion’s eliabili y equi emen s.
The communica ion s ack should be designed o seamless in eg a ion wi h comme cial o - he-
shel (COTS) mobile and IoT de ices by employing s anda d-complian physical and MAC laye
abs ac ions whe e applicable. While le e aging cus om backsca e modula ion echniques, he
p o ocol should main ain compa ibili y wi h widely adop ed wi eless amewo ks (e.g., BLE,Wi-
Fi, o sub-GHz indus ial, scien i ic and medical (ISM) bands) h ough p oxy ga eways o p o ocol
ansla ion laye s. This ensu es scalable deploymen and in e ope abili y wi hin he e ogeneous
IoT ecosys ems, while acili a ing s aigh o wa d use access ia sma phones o o he exis ing
in as uc u e.
The sys em mus suppo concu en in e ac ions om mul iple backsca e ags o use de ices
wi hin a sha ed RF domain wi hou incu ing packe collisions o iden i y ambigui ies. Tag iden-
i ica ion mus be obus agains signal in e e ence and closely spaced ansmissions, ensu ing
Page 29 o 94
D5.1 - Use cases, KPIs and KVIs
accu a e di e en ia ion and acking o mul iple ac i e en i ies in dynamic en i onmen s.
Posi ioning/Loca ion
The sys em a chi ec u e shall no ac i ely de e mine o upda e he spa ial posi ioning o ene gy-
neu al de ices du ing un ime. Ins ead, de ice commissioning, including he assignmen o
spa ial con ex , shall be conduc ed manually du ing ini ial sys em se up. To minimize complexi y
and conse e ene gy, spa ial g anula i y should be limi ed o he oom o zone le el a he han
equi ing p ecise localiza ion. This app oach aligns wi h he cons ain s o ene gy-ha es ing
de ices and suppo s scalable deploymen in in as uc u e-limi ed en i onmen s.
Managemen
Ene gy neu al de ices shall embed a UID wi hin hei ansmi ed payloads, ensu ing aceable
associa ion be ween senso da a o de ice s a e and he o igina ing ha dwa e ins ance. This UID
is c i ical o he a o emen ioned sys em commissioning, enabling de e minis ic mapping o da a
s eams o speci ic de ices du ing deploymen and ope a ion. While spa ial me ada a, such as
oom-le el loca ion, should be de ined du ing ini ial se up, he sys em shall suppo con olled
upda es o his me ada a o accommoda e econ igu a ion o eloca ion.
The sys em should suppo con igu able mechanisms o empo a ily o pe manen ly disable non-
c i ical ambien IoT senso s, he eby op imizing ene gy usage and minimizing unnecessa y RF
ac i i y. This unc ionali y should be con ollable ia cen alized policies o local igge s, allowing
o dynamic adap a ion o en i onmen al condi ions, use p e e ences, o ope a ional p io i ies.
To ensu e ope a ional in eg i y, de ices mus unde go pe iodic heal h checks o e i y ac i e s a-
us, unc ional co ec ness, and communica ion eliabili y. Faul condi ions (e.g., ansmission
ailu es, senso anomalies, o ene gy deple ion) shall be logged and lagged o main enance. Fu -
he mo e, pe o mance me ics, including signal s eng h, da a la ency, and ac i i y equency,
should be con inuously moni o ed and analysed o in o m diagnos ics, suppo p edic i e main-
enance, and ensu e long- e m sys em obus ness.
Collec ed In o ma ion and Ne wo k Exposu e
De ices shall ansmi use -associa ed eleme y, such as en i onmen al eadings (e.g., em-
pe a u e) o de ice s a e in o ma ion, o he cen al con olle o pu poses including sys em
analy ics, beha iou al in e ence, o pe sonaliza ion. While he ansmi ed da a may appea
non-sensi i e in isola ion, app op ia e sa egua ds mus be implemen ed o p ese e use p i acy.
All da a mus be anonymized o pseudonymized in acco dance wi h es ablished p i acy s anda ds,
ensu ing ha indi iduals canno be di ec ly iden i ied h ough he da ase wi hou au ho ized
co ela ion.
Back-end sys ems mus p o ide mechanisms o secu e da a access, wi h ull suppo o use -
ini ia ed da a e asu e and audi ing o s o ed eco ds. Fu he mo e, he sys em a chi ec u e shall
inco po a e obus cybe secu i y measu es, including au hen ica ion, enc yp ion, access con ol,
and in usion de ec ion, o sa egua d agains unau ho ized access, da a b eaches, and malicious
ampe ing.
Page 30 o 94
D5.1 - Use cases, KPIs and KVIs
Ene gy
This use case depends on de ice classes equi ing limi ed da a bu e ing o sho -du a ion com-
pu a ional asks (e.g., Class 3 de ices). The in eg a ion o small-scale ene gy s o age elemen s,
such as supe capaci o s o high-e iciency capaci o s may be employed. These s o age com-
ponen s enable ansien high-powe ope a ions, such as ime-c i ical esponse coo dina ion o
da a aming, while p ese ing he sys em’s o e all ul a-low powe oo p in . In sho - e m
deploymen s, small ba e y-assis ed ags may be p o isionally u ilized. Howe e , his s a egy is
gene ally discou aged due o en i onmen al conce ns associa ed wi h ba e y disposal and he
egula o y implica ions o haza dous ma e ial classi ica ion. In some scena ios, ambien ligh
ha es ing can enable long- e m, ba e yless ope a ion.
Use -Sa e y and Robus ness
RF emissions om he sys em mus adhe e o all applicable egional egula o y s anda ds (e.g.,
Fede al Communica ions Commission (FCC),Eu opean Telecommunica ions S anda ds Ins i u e
(ETSI)), ensu ing compliance wi h spec al powe densi y limi s and in e e ence h esholds. De-
ices shall be designed o ope a e in a non-in usi e manne , exhibi ing minimal elec omagne ic
o physical impac on hei immedia e en i onmen . In he e en o sys em ailu e, communi-
ca ion loss, o deg aded pe o mance, c i ical allback unc ionali ies, such as manual con ol
in e aces (e.g., ligh swi ch ac i a ion), mus emain accessible o he use . Fu he mo e, he
sys em shall suppo au onomous aul eco e y mechanisms, enabling sel - econ igu a ion o
e-synch onisa ion wi hou use in e en ion o es o e nominal ope a ion. This is possible wi h
low-powe wa chdog ime s, moni o ing i mwa e execu ion and igge sys em ese s o sa e
s a e allback i abno mal beha iou is de ec ed. Pe iodic diagnos ics o heal h beacons (whe e
ene gy budge pe mi s) can p o ide minimal s a us upda es o indica e connec i i y, senso s a e,
o cha ge s a us o local ene gy s o age.
AI/ML
Ad anced signal p ocessing a he cen al p ocessing uni (access poin (AP)) side, such as adap-
i e il e ing and in e e ence supp ession, can enhance ag de ec ion in challenging RF backsca -
e en i onmen s, pa icula ly in mul ipa h- ich o spec ally conges ed condi ions. While such
echniques a e aluable, many sma home applica ions cu en ly ely on simple p ocessing
app oaches, which will be benchma ked agains AI/ML-based me hods.
A-IoT models can be le e aged a he sys em le el o analyse use in e ac ion pa e ns, en i on-
men al senso da a, and ac ua ion beha iou . These models enable ene gy op imiza ion, clima e
con ol adap a ion, and suppo au oma ed de ice commissioning, educing manual se up e -
o s.
Addi ionally, ML models play a key ole in p edic i e main enance by de ec ing anomalies such
as i egula RSS o delayed esponses. A no able applica ion is he ea ly de ec ion o aul s in
household ba e y sys ems o powe elec onics, helping p e en haza ds such as o e hea ing o
i e, while main aining a ligh weigh oo p in on indi idual de ices.
Howe e , in many p ac ical sma home use cases, such complexi y may no be necessa y, basic
signal p ocessing echniques a e o en su icien o achie e eliable pe o mance, pa icula ly in
s a ic, low-in e e ence en i onmen s. As pa o he sys em e alua ion, AI/ML models should
Page 31 o 94
D5.1 - Use cases, KPIs and KVIs
he b idge o embedded in o he s uc u e, ene gy ha es ing om ambien sou ces migh no
be iable as a s and-alone solu ion. This is why, he ga eways can ac he imes o ene gy ans-
mi e s cha ging he ENDs using RF wa es. No ably, he ENDs deploymen and hei ene gy
equi emen s will a ec he op imal deploymen o such ga eways.
Use -Sa e y and Robus ness
ENDs mus ope a e unde ha sh condi ions, including la ge empe a u e a ia ions, high humidi y,
co osion, and mechanical ib a ions. Wa e /dus -p oo enclosu es, e.g., IP-g aded, will be
equi ed o main ain he ENDs ope a ional. In addi ion, physical-secu i y measu es including
ampe - esis an housings, lockable moun s, and low-complexi y in usion de ec ion mechanisms
mus be in place o p e en unau ho ized access o andalisms in ENDs wi hin he each o
ci izens. Finally, o ensu e apid eco e y om mal unc ions, ENDs should suppo manual o
au onomous eco e y mechanisms depending on he ins alla ion condi ions. Manual op ions
include ho -swappable modules o ield echnicians, while au onomous eco e y can encompass
buil -in wa chdogs, sel - eboo ing i mwa e, and sel -healing so wa e ou ines.
AI/ML
The amoun o measu emen da a gene a ed by he ENDs on he b idge can g ow signi ican ly
o e ime. No ably, AI/ML-aided echniques can help educe he olume o gene a ed da a
by dynamically adjus ing he h esholds ha igge sensing ope a ions. Addi ionally, AI/ML
me hods can assis enginee s esponsible o b idge main enance by de ec ing anomalies in he
collec ed da a, he eby helping o p edic po en ial isks, schedule imely main enance ope a ions,
and implemen a ic es ic ions o load managemen measu es when necessa y.
AI/ML echniques can also be used o communica ion asks such as il e ing, in e e ence sup-
p ession, join channel es ima ion and signal de ec ion, and well as o boos ing ene gy e iciency,
e.g., ia AI/ML-based du y-cycling. No ably, he applica ion o AI/ML in his con ex mus be
op imized o ope a e wi hin he limi ed p ocessing powe , memo y, and ene gy cons ain s o
ENDs, wi h compu a ionally in ensi e wo kloads delega ed o he ga eways.
Page 38 o 94

D5.1 - Use cases, KPIs and KVIs
4.4.3 Key Pe o mance Indica o s
Table 4.11: KPIs o B idge Heal h Moni o ing use case.
De ice KPIs Ne wo k KPIs
KPI Value KPI Value
Sensing accu acy Depends on magni ude End- o-end la ency < 1 s
De ice au onomy > 20 yea s Age o in o ma ion < 10 s
A g. powe consump ion < 1 mW Packe loss a e <1 %
Peak powe consump ion < 10 mW Message size <1 kbi
Communica ion ange < 100 m
Se ice a ea dimension B idge size
Max simul aneous de ices < 60 de ices
T ans e in e al < 10 s
De ice speed S a ic
Use -expe ienced da a a e 100 kbps
Applica ion KPIs AI/ML KPIs
KPI Value KPI Value
Communica ion se ice a ailabili y > 99.9 % T aining complexi y 200 MFLOPS
Posi ioning/Loca ion se ice a ailabili y No equi ed In e encing accu acy > 95 %
Posi ioning/Loca ion se ice accu acy N.A. In e encing la ency < 0.1 s
Command se ice a ailabili y No equi ed
AI/ML capabili ies Yes
4.4.4 Key Value Indica o s
En i onmen al KVIs
Impac Sco e: 3 While indi ec , he en i onmen al bene i s o sma b idge heal h moni o ing
a e s ill no able. By educing he numbe and equency o inspec ion ips and hea y equipmen
use, he sys em con ibu es o lowe emissions om logis ics ope a ions. I also suppo s esou ce
e iciency, as ewe eac i e main enance ac i i ies mean educed consump ion o ma e ials and
ene gy. Impo an ly, he sys em p omo es long- e m sus ainabili y by p ese ing he s uc u al
in eg i y o he b idge, educing he need o majo eno a ions o eplacemen s and hus lowe ing
he en i onmen al oo p in o e he b idge’s li e cycle.
Ene gy-sa ing and ene gy-ha es ing echnologies minimize (and e en elimina e) he dependence
on ba e ies. This no only p olongs he li espan o he moni o ing sys em bu also educes he
en i onmen al pollu ion associa ed wi h imp ope was e handling.
Social KVIs
Impac Sco e: 4 A sma b idge heal h moni o ing sys em has a high social impac , p ima ily
by enhancing public sa e y. Real- ime moni o ing enables ea ly de ec ion o s uc u al issues,
signi ican ly educing he isk o ca as ophic ailu e and p o ec ing human li es. Addi ionally, he
p esence o such a sys em os e s communi y us in in as uc u e, as use s will eel sa e using
Page 39 o 94
D5.1 - Use cases, KPIs and KVIs
b idges ha a e con inuously moni o ed. The sys em also con ibu es o g ea e accessibili y
and educed dis up ion by minimizing he equency and du a ion o closu es needed o manual
inspec ions, hus ensu ing smoo he a ic low and access o essen ial se ices.
Economic KVIs
Impac Sco e: 5 Economically, while up on cos s a e highe han o unmoni o ed b idges
due o he ENDs, ga eways, and he in as uc u e o suppo he A-IoTs ne wo k in gene al, he
applica ion o e s subs an ial alue by minimizing he b idge’s ope a ional cos s. Th ough p e-
dic i e main enance, i enables imely epai s ha a e less cos ly han eme gency in e en ions,
while also ex ending he li espan o he s uc u e. The sys em enhances ope a ional e iciency by
educing he need o labou -in ensi e inspec ions and allowing main enance c ews o ocus on
a ge ed a eas. Fu he mo e, by de ec ing and add essing ulne abili ies ea ly, he sys em helps
mi iga e inancial isk ela ed o liabili y om po en ial s uc u al ailu es o se ice dis up ions.
Minimizing main enance ope a ions on he ENDs has a signi ican economic impac in sma
b idge heal h moni o ing sys em. Conside ing he a la ge numbe o ENDs migh be deployed
on he b idge, equen main enance becomes cos ly, ime-consuming, o may equi e a ic
dis up ions. Beyond di ec main enance sa ings, he alue o sma b idge heal h moni o ing
lies in ensu ing con inuous se ice, public sa e y, and logis ical eliabili y, all c i ical o egional
economic p oduc i i y.
Inno a ion KVIs
Impac Sco e: 2 Inco po a ing con ex -awa eness o ene gy ha es ing and wake-up ech-
nologies b ings signi ican bene i s o scale b idge heal h moni o ing ope a ions. No ably, he
inno a i e use o A-IoT sys ems enables he implemen a ion o low-complexi y senso s by exploi -
ing ambien ene gy sou ces no only o cha ge he ENDs, bu also o in e physical magni udes
di ec ly om he cha ac e is ics o he ha es ed ene gy. Simila ly, ene gy ha es ing ci cui s can
se e as e en -d i en wake-up igge s, allowing he ENDs o emain inac i e un il he physical
magni ude o be measu ed changes beyond p ede ined h esholds.
En i onmen al
Social
Economic
Inno a ion
12345
Figu e 4.6: Impac Sco es o Sma B idge Heal h Moni o ing.
Page 40 o 94
D5.1 - Use cases, KPIs and KVIs
4.5 Pe sonal Belongings Finding
Table 4.12: Use Case Families o Pe sonal Belongings Finding use case.
Use Case Families
Ha d- o- each de ices Massi e-scale senso ne wo ks Ex eme-li e ime applica ions
Table 4.13: END classes o Pe sonal Belongings Finding use case.
END Classes
UP ENDs DB ENDs DP ENDs CO ENDs
4.5.1 Desc ip ion
In p esen days, people ha e wi h hem and use all kinds o pe sonal belongings in hei daily
li es. These pe sonal i ems ange om elec onic de ices like headphones o sma wa ches, o
less sophis ica ed i ems like keys, ID ca ds o walle s. In mos cases, hese pe sonal i ems ha e
a high economical o pe sonal alue, which means ha losing hem could ep esen a p oblem.
Howe e , he loss o pe sonal belongings in he domes ic en i onmen a ec s people o all ages
in all con ex s on a egula basis.
Sea ching o hese los i ems can esul in ime loss, in e up ions in daily ac i i ies, and high
le els o s ess and anxie y. This a ec s p oduc i i y and emo ional well-being, and can also
impac social and p o essional li e. The A-IoT echnology can p o ide a sus ainable solu ion o
his p oblem.
This use case uses A-IoT echnology o enhance he e e yday expe ience o loca ing misplaced
pe sonal i ems wi hin a domes ic en i onmen . The sys em in eg a es END ags, which a e
small, low-cos , and ene gy-e icien de ices ha can be designed o be a ached o pe sonal
i ems. These END ags enable in e ac ion wi h he su ounding in as uc u e wi hou equi ing
cons an use in e en ion.
When needed, he END ag ge s ene gy om a RF signal p o ided by he UE o he pe sonal
i em’s owne (e.g. a sma phone). Addi ionally, o he ambien ene gy sou ces can be used in
o de o ensu e obus ope a ion in mul iple scena ios. When a pe sonal i em is los , he owne
ini ia es he inding p ocedu e wi h hei UE. The UE sends a RF signal o ene gize he END
ag a ached o he los i em and s a s he posi ioning p ocedu e. When ene gized, he END
ag modula es he ecei ed RF signal elying on backsca e ing modula ion, hus being able o
espond o he UE’s posi ioning eques . Though END ags do no ha e ac i e localiza ion
capabili ies like Global Posi ioning Sys em (GPS) due o hei simplici y and powe cons ain s,
hei loca ion can be in e ed h ough passi e echniques such as signal iangula ion, p oximi y
es ima ion, o RF inge p in ing. By e lec ing RF signals o mul iple ecei e s like sma phones
Page 41 o 94
D5.1 - Use cases, KPIs and KVIs
Figu e 4.7: Illus a ion o Pe sonal Belongings Finding use case.
o o he IoT nodes, he sys em can es ima e he END ag’s posi ion based on signal s eng h,
ToF o AOA da a. This enables use s o de ec he p esence o a los i em and o ecei e
app oxima e loca ion guidance, enhancing he e ec i eness o he sea ch p ocess. Wi h espec
o o he low powe echnologies (Ul a-Wideband (UWB),BLE,RFID...), A-IoT uses simple
and less expensi e de ices. Fu he mo e, he ac ha A-IoT does no equi e ba e ies o
wo k also ep esen s an ad an age in e ms o bo h complexi y and en i onmen al conce ns.
Figu e 4.7 depic s a possible scena io o he Pe sonal Belongings Finding use case.
4.5.2 Func ional Requi emen s
Communica ion
The sys em mus suppo ul a-low powe communica ions while ensu ing su icien da a a es
and packe sizes o enable he ansmission o iden i ie s om END ags o UEs and IoT nodes.
To gua an ee eliable ope a ion e en in challenging en i onmen s, END ags loca ed in ha d-
o- each a eas o egions wi h limi ed co e age mus be capable o es ablishing communica ion
wi h bo h UEs and IoT nodes. Robus ness mus be achie ed wi h simple and e icien e o
de ec ion and synch oniza ion mechanisms, a oiding any inc ease in he complexi y o powe
consump ion o he END ags. END ags mus ely on backsca e ing modula ion o encode
in o ma ion on o ex e nal RF ca ie s, enabling he ansmission o da a o UEs and IoT nodes.
Fu he mo e, communica ion should be ini ia ed by UEs, he eby minimizing unnecessa y da a
exchanges and enabling e icien and low-powe ope a ion.
Posi ioning/Loca ion
The sys em mus ope a e in all domes ic en i onmen s, including bed ooms, li ing ooms,
ga ages, and ga dens, whe e signal condi ions and physical layou s may a y signi ican ly. Mo e-
o e , i should also ex end i s unc ionali y ou side he home, enabling use s o loca e i ems in
semi-public o ex e nal places i he use has access o nea by ne wo k nodes. To ensu e usabil-
i y in di e se con ex s, he sys em mus o e su icien posi ioning g anula i y o de e mine he
loca ion o objec s in small ooms, whe e spa ial esolu ion is essen ial o i em inding.
Page 42 o 94
D5.1 - Use cases, KPIs and KVIs
Managemen
Each END ag mus ha e a UID ha allows i o be dis inguished om o he s. The sys em
mus suppo he egis a ion and associa ion o each END ag wi h a speci ic physical objec .
To main ain lexibili y, he sys em mus also allow o dynamic me ada a upda es, enabling
use s o modi y he objec associa ed wi h a gi en END ag. Addi ionally, he sys em mus
inco po a e mechanisms o pe iodically moni o he ope a ional s a us o he ags by e alua ing
heal h indica o s like backsca e signal quali y and esponse ime. Any de ec ed aul s should be
logged, and use s mus ha e access o he aul his o y o acili a e in e en ion o eplacemen
o mal unc ioning ENDs ags.
Collec ed In o ma ion and Ne wo k Exposu e
Since END ags in his use case ypically emain unpowe ed, he amoun o in o ma ion hey can
collec is limi ed. Ne e heless, ce ain da a—such as he numbe o ac i a ions o he mos
ecen loca ions whe e a ag was de ec ed—can be eco ded and ansmi ed on demand when
equi ed by he use . To ensu e he in eg i y and con iden iali y o his in o ma ion as well as
p i acy o use ’s da a and loca ion, he sys em mus implemen low-powe ye obus secu i y
p o ocols, including da a access con ol and enc yp ion mechanisms, o p o ec communica ions
agains po en ial in e cep ions o da a leaks.
Ene gy
The p ima y powe ing me hod o END ags in his use case mus be WPT, as los i ems may
end up in loca ions whe e no ambien ene gy sou ces a e a ailable. None heless, complemen a y
ene gy ha es ing echniques like ib a ion, he mal g adien s, pho o ol aic cells, o RF ene gy
ha es ing can be conside ed o suppo END ags ope a ion. In scena ios whe e he physical size
o he END ag is no a limi ing ac o and does no in e e e wi h he no mal use o he objec
o which i is a ached, he use o supe capaci o s o o he low-ene gy s o age elemen s can be
conside ed o enhance END ag unc ionali y. Howe e , due o en i onmen al conside a ions,
such s o age componen s should only be employed when necessa y.
Use -Sa e y and Robus ness
END ags mus be designed o wo k unde di e en condi ions, including shocks, humidi y,
ib a ions, and wide empe a u e anges. This ensu es obus ope a ion in di e en scena ios.
A he same ime, hey mus be compac enough no o in e e e wi h he no mal use o he
pe sonal objec s o which hey a e a ached, p ese ing use com o and objec unc ionali y.
Addi ionally, use sa e y mus be a p io i y; END ags should be manu ac u ed using non- oxic
ma e ials and mus inco po a e bo h ha dwa e and so wa e-le el sa e y mechanisms o p e en
mal unc ions ha could po en ially lead o use ha m.
AI/ML
Due o hei low complexi y and limi ed capabili ies, i is di icul o use AI/ML echniques
(including hose aimed a low-powe applica ions) in END ags. Howe e , AI/ML can be deployed
in he UE o in he cloud. These AI/ML unc ionali ies can be used o ack ypical use
ajec o ies o equen ly isi ed loca ions, which can help in sea ching o los objec s. AI/ML
Page 43 o 94

D5.1 - Use cases, KPIs and KVIs
models can also be deployed in he UE o implemen ad anced signal p ocessing echniques.
AI/ML can help in p ocesses like channel es ima ion, equaliza ion, decoding o demodula ion o
educe e o a e and imp o e communica ion om END ags o UEs.
4.5.3 Key Pe o mance Indica o s
Table 4.14: KPIs o Pe sonal Belongings Finding use case.
De ice KPIs Ne wo k KPIs
KPI Value KPI Value
Sensing accu acy N.A. End- o-end la ency < 1 s
De ice au onomy > 20 yea s Age o in o ma ion < 10 s
A g. powe consump ion < 1 µW Packe loss a e <1 %
Peak powe consump ion < 100 µW Message size <1 kbi
Communica ion ange < 10 m
Se ice a ea dimension < 200 m2
Max simul aneous de ices < 10 de ices
T ans e in e al On use eques
De ice speed S a ic
Use -expe ienced da a a e 1 kbps
Applica ion KPIs AI/ML KPIs
KPI Value KPI Value
Communica ion se ice a ailabili y > 99.9 % T aining complexi y N.A.
Posi ioning/Loca ion se ice a ailabili y > 99.9 % In e encing accu acy N.A.
Posi ioning/Loca ion se ice accu acy < 1 m In e encing la ency N.A.
Command se ice a ailabili y No equi ed
AI/ML capabili ies No equi ed
4.5.4 Key Value Indica o s
En i onmen al KVIs
Impac Sco e: 4 The "Pe sonal Belongings Finding" use case has a high en i onmen al im-
pac . Ha ing a me hod o loca ing los objec s educes he need o manu ac u e eplacemen
uni s. This educ ion in he eplacemen s means less ex ac ion o aw ma e ials and a conse-
quen educ ion in ene gy consump ion associa ed wi h he design, manu ac u ing and ans-
po a ion s ages.
Reco e ing los i ems also p e en s p ema u e disposal, p olonging he li e o he goods and
educing he amoun o was e and e-was e. This e ec ansla es in o a educ ion in he
ma e ials des ined o land ills, educing he elease o ha m ul subs ances ha esul om he
deg ada ion o plas ic and elec onic componen s.
Mo eo e , he use o A-IoT o his use case compa ed o o he echnologies has ad an ages
om an en i onmen al poin o iew. END ags ha e educed elec onic complexi y and use
low-en i onmen al-impac ma e ials. Thei manu ac u ing phase consumes ewe esou ces and
Page 44 o 94
D5.1 - Use cases, KPIs and KVIs
p oduces lowe CO2equi alen emissions, especially in he absence o ba e ies [60]. Du ing use,
A-IoT gene a es no ope a ional emissions om echa ging and main enance, as no ba e ies a e
used in END ags. Finally, a he end o li e, he absence o ba e ies and hea y componen s in
A-IoT acili a es was e managemen and educes e-was e gene a ion.
Social KVIs
Impac Sco e: 4 Reco e ing los objec s h ough emo e loca ion educes he anxie y and
us a ion associa ed wi h los belongings. By educing he ime spen sea ching o los i ems,
he s ess caused is mi iga ed. Mo eo e , he ce ain y o being able o loca e pe sonal i ems
in case o loss inc eases he sense o con ol and educes pe cei ed ulne abili y. This e ec
con ibu es o a be e social en i onmen in which indi iduals eel p o ec ed agains acciden al
losses.
Usage o END ags also con ibu es o he au onomy o use s wi h limi ed mobili y o cogni i e
impai men by acili a ing access o hei belongings wi hou equi ing assis ance. This inc eased
sel -su iciency imp o es sel -es eem and educes hei dependence.
Economic KVIs
Impac Sco e: 4 F om he owne ’s poin o iew, eco e ing los i ems di ec ly educes he
impac associa ed wi h pu chasing eplacemen uni s. Fo example, his sa ings is pa icula ly
signi ican o high- alue i ems, whe e a oiding a single eplacemen can o se he in es men in
emo e loca ion ools. Fo objec s suscep ible o duplica ion (keys, access ca ds, c edi ca ds...),
loca ing he o iginal elimina es he need o pay o copying se ices.
F om he manu ac u e ’s poin o iew, by elimina ing expensi e and powe -hung y ac i e com-
ponen s, A-IoT enables olume p oduc ion o END ags, so he economic cos o each uni
becomes e y low.
Inno a ion KVIs
Impac Sco e: 3 The "Pe sonal Belongings Finding" use case does no o e signi ican in-
no a i e alue because exis ing de ices al eady pe o m his unc ion. Howe e , i is he use o
A-IoT echnology ha b ings inno a i e alue o his use case.
The in eg a ion o backsca e communica ion, WPT and ambien ene gy ha es ing allows
END ags o ope a e wi hou ba e ies o ex e nal powe supplies, opening he doo o de ices
wi h ex emely long li espans and no main enance. Mo eo e , he simplici y o END ags allows
o hei in eg a ion in e e yday objec s (clo hing, accesso ies, key ings, packaging...) wi hou
modi ying he aes he ics o e gonomics o he p oduc . This inc eases adop ion and enables new
p oduc s like disposable ex ile labels o senso s ha can be inse ed in o single-use packaging.
Page 45 o 94
D5.1 - Use cases, KPIs and KVIs
En i onmen al
Social
Economic
Inno a ion
12345
Figu e 4.8: Impac Sco es o Pe sonal Belongings Finding.
Page 46 o 94
D5.1 - Use cases, KPIs and KVIs
4.6 In-body o Wea able Medical Senso s
Table 4.15: Use Case Families o In-body o Wea able Medical Senso s use case.
Use Case Families
Ha d- o- each de ices Massi e-scale senso ne wo ks Ex eme-li e ime applica ions
Table 4.16: END classes o In-body o Wea able Medical Senso s use case.
END Classes
UP ENDs DB ENDs DP ENDs CO ENDs
4.6.1 Desc ip ion
In-body and wea able senso s a e e olu ionizing he way human ac i i ies and physiological
s a es a e moni o ed, expanding a beyond adi ional medical applica ions o include wellness,
spo s pe o mance, elde ly ca e, and li es yle op imiza ion. These senso s unc ion by de ec ing
and measu ing physical (e.g., mo ion, p essu e), biochemical (e.g., glucose, hyd a ion), o en i-
onmen al (e.g., empe a u e, humidi y) pa ame e s h ough a ious sensing elemen s. Da a is
hen p ocessed locally o ansmi ed wi elessly o in e p e a ion and ac ion.
Wea able senso s a e ypically wo n ex e nally in eg a ed in o ex iles, w is bands, o skin ad-
hesi es while in-body senso s a e ei he implan ed o inges ed. Ad ances in lexible elec onics,
sma ma e ials, minia u ized ci cui s, and wi eless communica ion ha e enabled hese senso s o
become inc easingly unob usi e, du able, and use - iendly. Fo example, i ness acke s mon-
i o ca dio ascula me ics, sma ab ics de ec hyd a ion le els, sma diape s ale mois u e
p esence [61], and implan able senso s p o ide eal- ime da a o mul iple medical applica ions
(e.g. glucome e s, pulse oxime e s, ehabili a ion eedback...).
An illus a i e example in ol es a hyb id wea able pla o m combining a sole-in eg a ed T ibo-
elec ic Nanogene a o (TENG) wi h a mic oneedle-based elec ochemical senso . The TENG,
embedded in oo wea , se es a dual pu pose: ha es ing mechanical ene gy om walking and
ac ing as a gai senso [62]. This ha es ed ene gy di ec ly powe s a low-ene gy mic oneedle
senso ha con inuously moni o s bioma ke s in in e s i ial luid, c ea ing a ba e y-less bio-
chemical sensing sys em. Addi ionally, in eg a ing ligh weigh AI amewo ks such as TinyML
enables on-de ice p ocessing o gai da a, acili a ing eal- ime ac i i y moni o ing [63]. An
o e iew o he en i e sys em is shown in Figu e 4.9.
Page 47 o 94
D5.1 - Use cases, KPIs and KVIs
Figu e 4.11: Illus a ion o Sma Ag icul u e use case.
4.7.2 Func ional Requi emen s
Communica ion
ENDs mus es ablish au onomous uplink connec ions and pe o m uplink ansmission wi hou
ex e nal wake-up commands. Suppo ed uplink ansmissions can be pe iodic o e en igge ed,
based on changes in en i onmen al pa ame e s (e.g., mois u e h eshold exceeded). Fu he -
mo e, ENDs mus ely on ambien backsca e ing communica ion o ul a-low-powe ac i e
ansmission o uplink da a ansmission. They mus suppo ansmission o small payloads
(e.g., 50–200 bi s) ep esen ing sensed me ics like humidi y, empe a u e, o CO2le el. ENDs
mus also be able o in e ace wi h nea by eade nodes (e.g., a pico-cell o mobile UE) placed
in g eenhouses o open ields. Coexis ence wi h housands o simila de ices ope a ing in he
same a m o g eenhouse, wi h minimal in e e ence and e icien channel access mechanisms,
including e o de ec ion/co ec ion, is also a e y impo an equi emen .
Posi ioning/Loca ion
Sys ems mus p o ide absolu e posi ion o ENDs in ou doo scena ios such as o unde g ound
soil sensing in open ields o suppo spa ially a ge ed ac ions (e.g., p ecision i iga ion, lo-
calized e iliza ion). They mus also suppo ela i e posi ioning be ween ENDs in p oximi y
o collabo a i e sensing o dis ibu ed decision making (e.g., pes zone de ec ion by neighbo -
ing senso s). In indoo ag icul u e se ings (e.g., sma g eenhouses), sys ems mus p o ide
oom- o ack-le el posi ioning o ENDs, enabling mic oclima e moni o ing and spa ial decision
suppo . De ices mus suppo mul i-sou ce posi ioning whe e easible o imp o e accu acy,
using combina ions o echnologies and signals, such as cellula and global na iga ion sa elli e
sys em (GNSS) (mo e applicable o ou doo ields), BLE,Wi-Fi, and ecei ed signal s eng h
indica o (RSSI) inge p in ing (mo e use ul o indoo g eenhouses), RFID ( o close-p oximi y
sensing), and use o local e e ences o mesh iangula ion (e.g., o he ENDs, s a ic eade s,
base s a ions).
Managemen
Rega ding managemen , sys ems mus p o ide mechanisms o ENDs so wa e/ i mwa e up-
da es, heal h checks o iden i y aul s o mal unc ions, and o empo a ily o pe manen ly dis-
able ENDs, including unc ionali y o emo e deac i a ion (e.g., a end-o -season o o secu i y
b eaches). Ac i a ion-on-demand mus also be suppo ed, allowing newly deployed o do man
ENDs o be ini ialized emo ely by a con olle o eade node.
Page 54 o 94

D5.1 - Use cases, KPIs and KVIs
Collec ed In o ma ion and Ne wo k Exposu e
Sys ems mus suppo access con ol policies o di e en oles (e.g., a me , ag onomis , en-
do ) ega ding ead/w i e pe missions o speci ic END g oups o da a s eams. On he o he
hand, ENDs mus p o ide eal- ime o nea eal- ime access o collec ed en i onmen al da a
(e.g., empe a u e, soil mois u e) when such da a collec ion is meaning ul o he a ming ask
p og ession, as well as when ene gy a ailabili y and communica ion condi ions pe mi . ENDs
mus also expose ligh weigh me ada a (e.g., de ice ID, loca ion, imes amp, senso ype) along
wi h senso eadings o acili a e au oma ed in e p e a ion and in eg a ion in o a m manage-
men sys ems.
Sys ems mus implemen Role-Based Access Con ol (RBAC) o ensu e only au ho ized en-
i ies (e.g., a me , ag onomis , equipmen endo ) can access END da a, issue commands,
o econ igu e beha io . They mus also limi unnecessa y exposu e o me ada a (e.g., de ice
ype, loca ion, sensing ole) o p e en p o iling o in e ence a acks, especially when ENDs a e
deployed in sensi i e ag icul u al zones o wi h p op ie a y c ops. Secu i y measu es like da a
enc yp ion a sou ce and dec yp ion a des ina ion shall be used when such me ada a is equi ed,
e.g., o da a p epa a ion, model eeding, aining o in e ence pu poses. In addi ion, sys ems
mus suppo policy-based e en ion and dele ion o ag icul u al da a, ensu ing his o ical da a is
e ained only as needed and dele ed upon use eques o END e oca ion.
Ene gy
ENDs mus suppo mul i-sou ce ene gy ha es ing (e.g., combining sola and RF) o imp o e
ene gy a ailabili y and eliabili y in di e se ag icul u al en i onmen s. The de ices mus also
include a small ene gy s o age componen (e.g., capaci o , mic o-ba e y, o supe capaci o ) o
enable ope a ion du ing empo a y una ailabili y o ambien ene gy sou ces (e.g., nigh - ime o
cloudy condi ions). The ene gy s o age sys em mus suppo sa e, e icien , and apid cha ging
unde low inpu powe condi ions (µW o mW scale). Fu he mo e, ene gy ha es ing and
s o age sys ems mus be obus o ag icul u al condi ions, including mois u e, dus , empe a u e
a ia ion, and exposu e o di ec sunligh o e ilize chemicals.
Use -Sa e y and Robus ness
ENDs mus ope a e eliably wi hin a b oad empe a u e ange (e.g., –20°C o +60°C) and
high humidi y (e.g., > 90 %), which a e ypical in open- ield ag icul u e and g eenhouses. In
addi ion, in e nal componen s mus be selec ed o p e en ailu e due o condensa ion o he mal
cycling. Fu he mo e, ENDs mus be enclosed in wea he - esis an housings o wi hs and ha sh
ou doo ag icul u al condi ions such as ain, di ec sunligh , wind, humidi y, os , and dus .
Sys ems mus also include physical p o ec ion agains co osion, especially o ENDs exposed
o i iga ion wa e and soil, o o wi hs and disin ec an s in g eenhouses a e c op ha es .
In addi ion, ENDs mus ole a e mechanical shocks and ib a ions, including hose caused by
passing a m machine y, animals, o wind-blown deb is, and mus be UV- esis an o long- e m
ou doo exposu e and chemically esis an o pes icides, e ilize s, and cleaning agen s used
in ag icul u al ope a ions. Ma e ials used in ENDs mus be non- oxic and ag icul u ally sa e,
ensu ing no ha m ul leacha es en e he soil o wa e supply, especially o c ops in ended o
human consump ion. ENDs designed o deploymen in di ec con ac wi h soil o c ops mus
comply wi h en i onmen al sa e y and biodeg adabili y guidelines. Finally, manual o au onomous
Page 55 o 94
D5.1 - Use cases, KPIs and KVIs
aul eco e y mechanisms mus be in place in case o END mal unc ion.
AI/ML
Wi h u he ad ancemen s in ene gy ha es ing ci cui s and AI-powe ed edge compu ing (e.g.,
TinyML), ENDs a e expec ed o e ol e in o mo e in elligen agen s capable o no jus sensing
bu also esponding o en i onmen al changes, which a e ypical in ou doo s ag icul u e en i-
onmen s, au onomously. AI/ML models can be applied o imp o e ag icul u al asks (c op
con amina ion de ec ion, op imiza ion o wa e consump ion in c ops wa e ing, e c.) as well as
o imp o e communica ion- ela ed asks ( il e ing, in e e ence supp ession, channel es ima ion,
e c.). AI models mus be obus o noisy senso inpu and long- e m d i , especially in ha sh
ag icul u al en i onmen s wi h a iable wea he , soil, and senso aging e ec s. Sys ems should
enable inc emen al o online lea ning models, allowing ENDs o adjus o seasonal changes (e.g.,
adap ing ligh -ha es ing p edic ion du ing cloudy mon hs) wi hou comple e e aining.
4.7.3 Key Pe o mance Indica o s
Table 4.20: KPIs o Sma Ag icul u e use case.
De ice KPIs Ne wo k KPIs
KPI Value KPI Value
Sensing accu acy > 95 % End- o-end la ency < 100 ms
De ice au onomy > 10 yea s Age o in o ma ion < 1 min
A g. powe consump ion < 100 µW Packe loss a e <1 %
Peak powe consump ion < 1 mW Message size < 1000 bi
Communica ion ange < 100 m; < 500 m 3
Se ice a ea dimension < 70000 m2
Max simul aneous de ices < 70000 de ices
T ans e in e al 60 min
De ice speed S a ic
Use -expe ienced da a a e 1 kbps
Applica ion KPIs AI/ML KPIs
KPI Value KPI Value
Communica ion se ice a ailabili y > 99.9 % T aining complexi y > 50 MFLOPS
Posi ioning/Loca ion se ice a ailabili y > 95 % In e encing accu acy > 90 %
Posi ioning/Loca ion se ice accu acy < 1 m In e encing la ency < 1 s
Command se ice a ailabili y > 99 %
AI/ML capabili ies Yes
3< 100 m in indoo scena ios; < 500 m in ou doo scena ios.
Page 56 o 94
D5.1 - Use cases, KPIs and KVIs
4.7.4 Key Value Indica o s
En i onmen al KVIs
Impac Sco e: 5 The Sma Ag icul u e use case con ibu es signi ican ly o en i onmen al
sus ainabili y by enabling p ecise, low- esou ce a ming p ac ices. ENDs educe he need o
equen human o machine y p esence in he ield, minimizing uel use and soil compac ion.
These de ices p omo e esou ce-e icien i iga ion, e iliza ion, and pes icide applica ion by de-
li e ing g anula , eal- ime en i onmen al da a. This educes uno and con amina ion o nea by
ecosys ems. The ENDs ambien ene gy ha es ing na u e elimina es ba e y- ela ed was e and
main enance ips. Thei minimal ma e ial complexi y and long ope a ional li espan educe li e-
cycle emissions and suppo sus ainable de ice deploymen in la ge-scale, emo e ag icul u al
en i onmen s.
Social KVIs
Impac Sco e: 4 Sma ag icul u e enhances ood secu i y and quali y by enabling be e c op
managemen , yield op imiza ion, and esilience o en i onmen al a iabili y. The use o ENDs
enables da a-d i en a ming e en in u al and unde se ed a eas whe e con en ional IoT in as-
uc u e may be una ailable o un eliable. Fa me s bene i om imp o ed wo king condi ions,
educed manual moni o ing, and ewe ha m ul exposu es (e.g., o pes icides). Fu he mo e,
he au onomy and simplici y o ENDs empowe small-holde a me s, suppo ing inclusi e ech-
nology adop ion ega dless o digi al li e acy o capi al in es men . O e ime, hese sys ems
con ibu e o mo e equi able and sus ainable ood sys ems.
Economic KVIs
Impac Sco e: 4 END-based sma ag icul u e sys ems o e cos sa ings h ough educed
inpu was e (e.g., wa e , e ilize ), minimized c op loss om s ess o disease, and lowe ed labo
demands o moni o ing. Thei ul a-low-cos and main enance- ee ope a ion enables massi e-
scale deploymen s e en in esou ce-cons ained se ings, p o iding a s ong e u n on in es men
o e ime. By enabling ea ly anomaly de ec ion and condi ion-based ac ions (e.g., i iga ion
on demand), hey educe expensi e eac i e in e en ions. ENDs also open new economic
oppo uni ies in he ag i- ech ma ke , including senso manu ac u ing, da a analy ics, and AI-
powe ed a m managemen pla o ms.
Inno a ion KVIs
Impac Sco e: 4 Sma ag icul u e wi h Ambien IoT in oduces a high deg ee o inno a ion
by combining ba e yless sensing, ene gy ha es ing, and AI-powe ed edge analy ics o massi e
scale and ha d- o- each deploymen s. Inno a ions such as ene gy-awa e sensing, TinyML-based
local anomaly de ec ion, and adap i e da a p io i iza ion enable ENDs no jus o sense, bu
o in e p e and ac on ag icul u e-speci ic condi ions au onomously. These ad ances ma k a
shi om con en ional da a collec ion o in elligen , sus ainable, and sel -managed ag icul u al
sys ems.
Page 57 o 94
D5.1 - Use cases, KPIs and KVIs
En i onmen al
Social
Economic
Inno a ion
12345
Figu e 4.12: Impac Fac o s o Sma Ag icul u e.
Page 58 o 94
D5.1 - Use cases, KPIs and KVIs
4.8 Asse , P oduc , Tool, and I em T acking
Table 4.21: Use Case Families o Asse , P oduc , Tool, and I em T acking use case.
Use Case Families
Ha d- o- each de ices Massi e-scale senso ne wo ks Ex eme-li e ime applica ions
Table 4.22: END classes o Asse , P oduc , Tool, and I em T acking use case.
END Classes
UP ENDs DB ENDs DP ENDs CO ENDs
4.8.1 Desc ip ion
This use case ocuses on p o iding eal- ime acking o c i ical asse s, p oduc s, ools, and
indi idual i ems wi hin con ined, high-p ecision ope a ional en i onmen s such as hospi als, lab-
o a o ies, ad anced manu ac u ing acili ies, and pha maceu ical clean ooms. I s objec i e is
o enable ad anced p ocess moni o ing, g anula quali y con ol, obus he p e en ion, and
op imized ope a ional e iciency.
Di e en ia ing om gene ic end- o-end logis ics, his acking p o ides con inuous, p ecise mon-
i o ing wi h signi ican ly highe upda e a es and lowe la ency. I o e s obus and accu a e
posi ioning, achie ing sub-me e le el p ecision e en in challenging and dense indoo en i on-
men s, while handling la ge olumes o con ex ual da a pe i em ( o example, en i onmen al
condi ions, usage and posi ion his o y, in eg i y s a us). These capabili ies enable c i ical appli-
ca ions such as p ocess moni o ing and digi al winning o wo k low alida ion and e iciency;
quali y con ol and compliance h ough au oma ed e i ica ion o i em placemen , posi ion his-
o y and en i onmen al condi ions; he con ol and loss p e en ion ia eal- ime geo encing
and anomali y de ec ion; and ope a ional op imiza ion by p o iding insigh s in o asse u iliza ion
and esou ce alloca ion.
In o de o allow o acking o objec s, agging is pe o med, simila ly o cu en RFID sce-
na ios. Tags ha e o ely on ene gy ha es ing o ul a-low powe consump ion (ENDs), which
is essen ial o he pe asi e, long- e m, and main enance- ee deploymen equi ed by his kind
o acking ac oss coun less i ems. In eg a ion app oaches include ex e nal agging; packaging-
in eg a ed agging ( o sma consumables/ki s); design-in eg a ed agging ( o ins umen / ool
in elligence); and embedded i em agging ( o componen /p oduc -le el digi al iden i y).
Fo all he objec i es o his use case, p ecise posi ion and con inuous acking o ENDs is
c ucial, equi ing ad anced signal p ocessing and in e ence algo i hms o map cons uc ion and
loca ion. A schema ic o e iew o he use case is gi en in Figu e 4.13.
4.8.2 Func ional Requi emen s
Page 59 o 94

D5.1 - Use cases, KPIs and KVIs
Cloud Da acen e
AP
Edge Se e
END
END
END
AP
WPT and Backsca e
P ocess Moni o ing
The P o ec ion
Se ice A ea
Con ex ual Da a
END
AI/ML
AI/ML
Quali y Con ol
P ocess Efficiency
Inc ease
Figu e 4.13: Illus a ion o Asse , P oduc , Tool, and I em T acking use case.
Communica ion
Passi e backsca e communica ion is he mos desi able o m o wi eless communica ions o
his use case due o i s suppo in ul a-low powe ENDs [33], [66]. Ac i e communica ion may be
possible in u u e applica ions i ecei e-side powe budge s a e inc eased om he mic owa -
le el o he milliwa le el [27]. Ne e heless, ac i e communica ion can be cos -p ohibi i e
and ene gy-in ensi e, pa icula ly in la ge o complex indoo en i onmen s. The wo exempla y
solu ions in [33], [66] add ess hese limi a ions by employing passi e communica ion h ough
ambien backsca e , wi h signal de ec ion ca ied ou ei he a he use equipmen o a he
ne wo k side (e.g., he APs), depending on he deploymen scena io.
To enable join backsca e communica ion and channel es ima ion ( o allow o obus posi-
ioning/localiza ion), ENDs mus employ ambien backsca e communica ion o ansmi uplink
signals owa ds he in as uc u e. Fo backsca e communica ion, he in as uc u e may com-
p ise sepa a e ansmi ing and ecei ing APs o a oid ull-duplex ope a ion o APs. The in as-
uc u e mus de ec weak and po en ially hea ily in e e ed backsca e signals om he END,
whe e in e e ence p ima ily a ises om he di ec ansmission pa hs be ween he APs. The
in as uc u e mus ensu e high dynamic ange a he ecei ing APs o enable eliable de ec ion
o low-powe backsca e ed signals. The in as uc u e may ansmi wideband exci a ion signals
o allow o wideband channel es ima ion, which is essen ial o ToF-based localiza ion. The in-
as uc u e may p o ide su icien compu a ional esou ces a he APs o de ec ing backsca e
signals om he END, cancelling in e e ence, and demodula ing backsca e symbols, alongside
s anda d p ocessing asks equi ed o he legacy signals such as o hogonal equency-di ision
mul iplexing (OFDM). In mul i-AP deploymen s, he in as uc u e may main ain synch oniza-
ion o ime, equency, and phase ac oss APs o suppo cohe en signal p ocessing and enhance
posi ioning accu acy and obus ness. The in as uc u e and he ENDs mus suppo mul iple
access p o ocols o massi e deploymen s o ENDs.
Posi ioning/Loca ion
To enable posi ioning, mapping, o sensing in a b oade sense, he wi eless in as uc u e mus
be capable o acqui ing channel obse a ions, i.e., noisy CSI. Se e al “channel pa ame e s” can
be es ima ed om noisy CSI and con ain END-posi ion in o ma ion.
Page 60 o 94
D5.1 - Use cases, KPIs and KVIs
RSS-based posi ioning. The sys em in [66] is based on he use o ENDs deployed a known
posi ions h oughou he building. Each END ag pe iodically ansmi s a unique iden i ie by
modula ing and e lec ing an inciden RF signal, a he han ac i ely gene a ing i s own ansmis-
sions. Two complemen a y con igu a ions a e conside ed. In he downlink-based con igu a ion,
ENDs modula e ambien cellula pilo signals (e.g., om a ou h-gene a ion (4G) o 5G base
s a ion (BS)), and hese modula ions a e de ec ed by he UE h ough analysis o pe u ba ions
in he CSI [67]. In he uplink-based con igu a ion, ENDs e lec he uplink signals emi ed by
he UE, and modula ed backsca e is de ec ed a he BS o a nea by AP [68]. In bo h con igu-
a ions, he obse ed signal a ia ions a e used o de e mine he END wi h he s onges signal,
whose known loca ion se es as a p oxy o he UE’s posi ion.
Angle-Delay-based posi ioning. The sys ems in es iga ed in [69] and [70] belong o he classes
o a Radio S ipes and a RadioWea es in as uc u e, espec i ely. We ha e shown in [71, Fig. 2]
ha in dis ibu ed MIMO (D-MIMO) in as uc u es wi h dis ibu ed APs, angula -domain-based
posi ioning supe sedes delay domain-based posi ioning ypically o bandwid hs below 100 MHz
in s ong line-o -sigh (LoS) channels. We ound ha he pa h o e lap in s ong mul ipa h
channels nega i ely impac s he pe o mance in such a egime [69, Fig. 4], bu la ge bandwid h
can help esol e he mul ipa h componen s. Based on measu ed channels, we ha e shown ha
simul aneous localiza ion and mapping (SLAM) can be used o bo h in e he END loca ion and
a geome ic map o he p opaga ion en i onmen , whe e cen ime e -le el posi ioning accu acy
has been demons a ed [70].
Delay-Dopple -based posi ioning. Dopple -based posi ioning may supe sede delay-based posi-
ioning in D-MIMOs in as uc u es [72, p. 20]. While he accu acy o Dopple -based eloci y
es ima ion is independen o he de ice mo ion, he posi ioning accu acy depends on he de ice
eloci y. Ne e heless, in combina ion wi h he delay-domain, cos -e ec i e ye accu a e wi eless
posi ioning, e en wi h a ew APs is possible in ealis ic scena ios wi h ha sh channel condi ions
such as pa ial obs uc ed-line-o -sigh (OLoS) and s ong mul ipa h p opaga ion [73]. Close
o cen ime e -le el posi ioning accu acy was demons a ed on he measu emen s om [72].
Dopple -based o ca ie -phase-based posi ioning [74] may be pa icula ly sui able o localizing
ENDs in backsca e communica ion because hese channels inhe en ly a oid he ca ie -phase
calib a ion p oblem.
Managemen
The in as uc u e mus p o ide mechanisms o he li ecycle managemen o ENDs. This in-
cludes disco e y, iden i ica ion, and logical onboa ding o new ENDs in o he sys em. Con e sely,
i mus also suppo e ec i e mechanisms o disabling and o boa ding ENDs when hey a e
no longe equi ed o become inope able. Fu he mo e, o ensu e consis en and scalable de-
ploymen , he sys ems mus acili a e cen alized, au oma ed p o isioning and con igu a ion o
APs. Las ly, he in as uc u e needs o o e in e aces o au oma ed i mwa e upda es o APs.
These upda es a e c ucial o implemen ing imp o ed ea u es, such as ad anced localiza ion and
es ima ion algo i hms, and o add essing c i ical secu i y ulne abili ies.
Collec ed In o ma ion and Ne wo k Exposu e
APs wi hin he ne wo k a e equi ed o p o ide in e aces o he exchange o channel es ima es o
in e media e posi ioning/localiza ion esul s. This da a exchange can occu ei he ia backhaul
Page 61 o 94
D5.1 - Use cases, KPIs and KVIs
links o h ough o e - he-ai in e aces. While sys ems may collec hese channel es ima es o
in e media e esul s, he subsequen p ocessing o his collec ed in o ma ion can be pe o med
ei he a he ne wo k edge o on cen alized sys ems. A e p ocessing, he sys ems mus p o ide
he posi ioning/localiza ion esul s o au ho ized en i ies, ensu ing da a access is con olled and
secu e.
Ene gy
RF WPT echnology is he p e e ed me hod o powe ENDs in his use case, equi ing an
in as uc u e capable o p o iding e icien WPT as a se ice on a massi e scale. An enna
a ays can p o ide bo h he ansmission e iciencies equi ed o powe ENDs and he egula o y
compliance o ensu e exposu e-sa e ope a ion [27].
Use -Sa e y and Robus ness
Robus ope a ion e en in ha sh channel condi ions such as low-signal- o-noise a io (SNR) and
unde high END mobili y can, o ins ance, be accomplished by join ly in e ing he END loca ion
and a geome ic map o i s p opaga ion en i onmen , as has been shown in [70].
AI/ML
AI and ML a e c ucial o enhancing he capabili ies o his use case. They enable p edic i e
main enance by analysing usage pa e ns and senso da a om ools and asse s, which helps
an icipa e po en ial ailu es, op imize main enance schedules, and p e en down ime. These
echnologies also acili a e anomaly de ec ion, iden i ying unusual mo emen s, en i onmen al
de ia ions, o depa u es om s anda d ope a ing p ocedu es. This igge s immedia e ale s
o issues like he a emp s, quali y excu sions, o p ocess e o s.
Fu he mo e, AI and ML con ibu e o wo k low op imiza ion by analysing his o ical acking
da a o pinpoin bo lenecks, s eamline i em low, and enhance o e all ope a ional e iciency in
complex en i onmen s. Thei ole ex ends o enhanced localiza ion and mapping, whe e neu al
ne wo ks and da a-d i en app oaches a e expec ed o imp o e upon model-based me hods, such
as in hyb id ac o -g aph-based sys ems, inc easing accu acy and obus ness in si ua ions whe e
adi ional models a e in ac able o complex. Finally, AI and ML a e ins umen al in au oma ed
in en o y and audi , s eamlining hese p ocesses by au onomously econciling physical i ems
wi h hei digi al wins, he eby educing manual e o and e o s.
Page 62 o 94
D5.1 - Use cases, KPIs and KVIs
4.8.3 Key Pe o mance Indica o s
Table 4.23: KPIs o Asse , P oduc , Tool, and I em T acking.
De ice KPIs Ne wo k KPIs
KPI Value KPI Value
Sensing accu acy > 95 % End- o-end la ency < 1 s
De ice au onomy > 10 yea s Age o in o ma ion < 1 s
A g. powe consump ion < 10 µW; < 1 mW 4Packe loss a e <1 %
Peak powe consump ion < 100 µW; < 5 mW 5Message size < 1 kbi
Communica ion ange < 100 m
Se ice a ea dimension < 10000 m2
Max simul aneous de ices < 100 de ices
T ans e in e al 1 s
De ice speed < 5 m·s-1
Use -expe ienced da a a e 10 kbps
Applica ion KPIs AI/ML KPIs
KPI Value KPI Value
Communica ion se ice a ailabili y > 99 % T aining complexi y 6< 2.5 GFLOPS
Posi ioning/Loca ion se ice a ailabili y > 99 % In e encing accu acy > 95 %
Posi ioning/Loca ion se ice accu acy < 1 m In e encing la ency < 100 ms
Command se ice a ailabili y N.A.
AI/ML capabili ies Yes
4.8.4 Key Value Indica o s
En i onmen al KVIs
Impac Sco e: 3 This use case signi ican ly con ibu es o en i onmen al sus ainabili y by
enabling p ecise esou ce managemen and was e educ ion. Th ough g anula acking and
eal- ime insigh s, i acili a es a educ ion in spoilage, pa icula ly o pe ishable goods o sensi-
i e ma e ials in labo a o ies and pha maceu ical se ings, he eby minimizing was e gene a ion.
The op imized u iliza ion o asse s and esou ces, d i en by da a-in o med decisions, educes
o e all consump ion. Fu he mo e, he implemen a ion o p edic i e main enance, enabled by
con inuous moni o ing o c i ical ools and equipmen , ex ends asse li ecycles and educes he
need o p ema u e eplacemen s, leading o a dec ease in manu ac u ing and disposal- ela ed
en i onmen al bu dens.
Social KVIs
Impac Sco e: 3 The social impac o his ambien IoT use case is subs an ial, p ima ily
h ough he enhancemen o sa e y and he imp o emen o wo king condi ions. Fo ins ance,
4< 10 µW o passi e ENDs; < 1 mW o ac i e ENDs
5< 100 µW o passi e ENDs; < 5 mW o ac i e ENDs
6Conside ing inge p in ing-based posi ioning [75].
Page 63 o 94
D5.1 - Use cases, KPIs and KVIs
Social KVIs
Impac Sco e: 3 By o e ing an unob usi e, con ex -awa e guide, he sys em can enhance
isi o engagemen and accessibili y—especially o use s who p e e pe sonalized, on-demand
con en wi hou wea ing o handling ex a de ices. The ouchless in e ac ion model also bene i s
hygiene and com o in public spaces.
Con e sely, eliance on isi o s’ own sma phones o dedica ed RF eade s may exclude hose
wi hou compa ible de ices o wi h limi ed digi al li e acy, po en ially widening he digi al-access
di ide. Ensu ing uni e sal access— ia loane eade s o mul ilingual, low- ech allback op-
ions—will mi iga e hese social isks, bu educe he economical bene i .
Economic KVIs
Impac Sco e: 3 Mos con empo a y museums ely on ba e y-powe ed handheld guides o
badges, which incu signi ican ongoing cos s o ba e y pu chase, cha ging in as uc u e,
labou o swap o cha ge uni s, and e en ual ba e y disposal. By con as , a ully passi e
backsca e -based guide elimina es all ba e y- ela ed expenses o e i s ope a ional li e. The
ul a-long ag li e ime (>20 yea s) means i ually ze o main enance labou and no ecu ing
consumables, deli e ing subs an ial o al cos o owne ship sa ings.
Howe e , ini ial ou lays o RF emi e s, eade in eg a ion, and backed so wa e may be highe
han o simple ba e y-powe ed handse s. Achie ing e u n o in es men (ROI) depends on
isi o olume and exhibi scale; smalle o seasonal ins alla ions migh need phased deploymen
o sha ed in as uc u e models o amo ize capi al in es men e ec i ely.
Inno a ion KVIs
Impac Sco e: 4 Deploying ully passi e RF backsca e o con ex -awa e museum guides
ep esen s a no el usion o A-IoT and cul u al-he i age engagemen . The app oach sides eps
ligh ing cons ain s and unlocks long- e m, main enance- ee ins alla ions in sensi i e exhibi ion
en i onmen s, ad ancing he s a e o he a in low-powe , loca ion-based se ices.
None heless, he echnology emains eme gen : in eg a ion wi h unmodi ied COTS de ices (e.g.,
sma phones), AI-enhanced ag de ec ion, and scalable managemen ools a e s ill ma u ing.
Con inued inno a ion in backsca e modula ion schemes, AI/ML de ec ion algo i hms, and
s anda dized p o ocols will de e mine how b oadly his concep can be adop ed.
Page 70 o 94

D5.1 - Use cases, KPIs and KVIs
En i onmen al
Social
Economic
Inno a ion
12345
Figu e 4.16: Impac Fac o s o Museum Guide.
Page 71 o 94
D5.1 - Use cases, KPIs and KVIs
4.10 End- o-end Logis ics
Table 4.27: Use Case Families o End- o-end Logis ics use case.
Use Case Families
Ha d- o- each de ices Massi e-scale senso ne wo ks Ex eme-li e ime applica ions
Table 4.28: END classes o End- o-end Logis ics use case.
END Classes
UP ENDs DB ENDs DP ENDs CO ENDs
4.10.1 Desc ip ion
E icien logis ics managemen is a co ne s one o mode n supply chains, ensu ing ha goods a e
deli e ed accu a ely, sa ely, and on ime. End- o-end logis ics encompasses mos o he li ecycle
o a p oduc — om i s o igin a he manu ac u ing si e o i s inal deli e y o he consume .
This p ocess in ol es mul iple s ages, including p oduc ion, wa ehousing, anspo a ion, and
dis ibu ion, each o which can bene i om in elligen moni o ing and au oma ion.
The eme gence o he A-IoT has in oduced a new pa adigm in logis ics. By embedding in el-
ligence in o e e yday objec s and en i onmen s, A-IoT enables seamless da a collec ion, com-
munica ion, and decision-making ac oss he supply chain. RFID echnology is a ounda ional
elemen in his con ex . Passi e RFID ags, a ached o p oduc s o packaging du ing manu ac-
u ing, s o e unique iden i ie s and essen ial me ada a. As goods mo e h ough he supply chain,
RFID eade s—s a egically placed a checkpoin s such as wa ehouses, loading docks, and e ail
ou le s—cap u e da a om hese ags. This enables eal- ime acking o p oduc loca ion and
s a us. Al hough passi e ags do no con ain hei own powe sou ce, hey can be ene gized ia
WPT om nea by eade s o mobile de ices, ensu ing con inuous ope a ion wi hou he need
o ba e ies. Beyond adi ional RFID,A-IoT expands he ange o capabili ies. Fo example,
empe a u e-sensi i e goods can be moni o ed using semi-passi e ags o e a longe pe iod o
ime wi h in e nal supe capaci o s o expand he ope a ion.
In addi ion, on-boa d uni s (OBUs) ins alled in deli e y ehicles p o ide aluable eleme y da a,
including ehicle loca ion, speed, uel consump ion, and main enance s a us. This in o ma ion
suppo s dynamic ou e op imiza ion, p edic i e main enance, and imp o ed lee managemen .
Toge he , sma ags, RFID, and o he A-IoT de ices o m a dis ibu ed, in elligen in as uc-
u e ha enhances isibili y, aceabili y, and e iciency ac oss he en i e logis ics chain. By
le e aging he capabili ies o A-IoT and ENDs, businesses can achie e mo e esilien , adap i e,
and sus ainable logis ics ope a ions.
A-IoT in eg a ion in o logis ics wo k lows, in his example RFID as enabling echnology, can
ollow se e al s a egic app oaches, each o e ing a ying le els o aceabili y, du abili y, and
cos -e iciency:
Page 72 o 94
D5.1 - Use cases, KPIs and KVIs
Re u ns
Diagnos ics
P oduc ion
Wa ehouse T anspo Deli e y
Secu e Cloud Da a
Re u n I ems
Figu e 4.17: Illus a ion o End- o-end Logis ics use case
•Ex e nal Tagging: ags a e a ixed o he ex e io o i ems, con aine s, o palle s. This
me hod is widely used o in en o y acking and shipmen e i ica ion due o i s simplici y
and low implemen a ion cos .
•Packaging-In eg a ed Tagging: Tags a e embedded di ec ly in o p oduc packaging o a
eusable anspo i em (RTI). This app oach enhances du abili y and educes he isk o
ag de achmen du ing handling and ansi .
•Design-In eg a ed Tagging: ag unc ionali y is inco po a ed in o he p oduc ’s design, o -
en in a de achable o semi-pe manen o m. This s a egy suppo s aceabili y h oughou
he p oduc li ecycle while allowing o ag euse o emo al a e sale.
•Embedded I em Tagging: ag componen s a e ully in eg a ed in o he p oduc i sel du ing
manu ac u ing. This me hod enables seamless acking and au hen ica ion, pa icula ly
aluable o high- alue goods. This agging me hod can be u he enhanced by combining
he de ice unc ion wi h he ag, o example, o ead ou diagnos ic in o ma ion o he
de ice o e i s li e cycle by he end cus ome .
4.10.2 Func ional Requi emen s
Communica ion
RFID ags mus suppo globally unique iden i ica ion and be eadable in dense ag en i on-
men s, such as wa ehouses o shipping con aine s, whe e hund eds o housands o ags may
be p esen simul aneously. Bulk iden i ica ion wi h a ead accu acy exceeding 99% is essen ial
o ensu e eliabili y. Communica ion p o ocols mus suppo an i-collision and e o co ec ion
mechanisms, and ope a e ac oss a ious equency bands (e.g., ul a-high equency (UHF),high
equency (HF)) depending on egional egula ions and applica ion needs.
Posi ioning/Loca ion
While RFID passi e ags do no possess ac i e ansmission capabili ies, hey can suppo posi-
ioning when deployed wi hin a dense in as uc u e o ixed eade . Same wi h A-IoT de ices i
hey ha e no ac i e communica ion capabili ies. These sys ems es ima e he posi ion o agged
i ems using echniques such as signal iangula ion, ToF,o RSSI. Howe e , achie ing sub-me e
Page 73 o 94
D5.1 - Use cases, KPIs and KVIs
accu acy wi h passi e sys ems alone emains challenging and o en equi es en i onmen al cali-
b a ion and edundancy.
To enhance bo h posi ioning p ecision and con ex ual loca ion awa eness, hyb id sys ems a e
inc easingly employed. These combine RFID wi h complemen a y echnologies such as UWB
o high-accu acy indoo posi ioning, o GPS o ou doo acking. A-IoT de ices, which o en
include mul iple ha es ing sou ces, senso s and connec i i y modules, u he en ich his ecosys-
em by p o iding con inuous da a s eams ha suppo eal- ime localiza ion and en i onmen al
moni o ing.
Managemen
E ec i e logis ics managemen equi es seamless in eg a ion o A-IoT da a in o esou ce plan-
ning and wa ehouse managemen sys ems. This includes eal- ime in en o y upda es, au oma ed
s ock le el ale s, and excep ion handling (e.g., damaged o missing i ems). Middlewa e mus
suppo da a il e ing, agg ega ion, and e en -based igge s o educe ne wo k load and imp o e
esponsi eness. Fu he , he en i onmen , ag numbe and ype is cons an ly changing. The e-
o e, lexible ne wo k s uc u es mus be implemen ed o be able o add, emo e, o econ igu e
new de ices.
Collec ed In o ma ion and Ne wo k Exposu e
END ags enhance logis ics by moni o ing en i onmen al ac o s such as empe a u e, humidi y,
ib a ion, and shock. These pa ame e s a e essen ial o main aining p oduc quali y, especially
in cold chain and agile goods anspo . Collec ed da a mus be secu ely ansmi ed o cloud
o edge pla o ms using enc yp ed and au hen ica ed channels. This p o ec s agains ampe ing
and unau ho ized access, ensu ing da a in eg i y h oughou he supply chain.
Ene gy
END ags ope a e wi hou in e nal ba e ies, elying solely on ha es ed ene gy om ambien
sou ces such as RF WPT and/o o he s. Ene gy ha es ing mus be su icien o powe sensing,
da a s o age, and communica ion unc ions. RFID eade s (bo h ixed and mobile) mus ensu e
eliable ag ac i a ion a e e y s age o he logis ics p ocess, including du ing ansi and a
deli e y checkpoin s.
Use -Sa e y and Robus ness
The sys em mus be obus agains en i onmen al s esso s such as dus , mois u e, empe a u e
luc ua ions, and elec omagne ic in e e ence. Tags and eade s should comply wi h ele an
sa e y s anda ds and be designed o long- e m du abili y, bu should also conside hei end o
li e (EoL) wi h op ion o ecycle o euse he ags. Scalabili y is also essen ial, wi h sys ems ex-
pec ed o handle housands o ags simul aneously wi hou deg ada ion in pe o mance. Secu e
digi al communica ion o END de ices in end- o-end logis ics also imp o es use p i acy. Unlike
p in ed labels, which expose s a ic in o ma ion, enc yp ed ansmissions allow con olled access,
suppo ing compliance and us .
Page 74 o 94
D5.1 - Use cases, KPIs and KVIs
AI/ML
AI/ML capabili ies a e p ima ily implemen ed on he in as uc u e side (e.g., edge se e s, cloud
pla o ms) a he han on he passi e RFID ags hemsel es. Due o hei ul a-low powe design,
passi e ags lack he compu a ional esou ces o onboa d p ocessing. Howe e , AI/ML can be
used o analyse agg ega ed RFID da a o anomaly de ec ion, p edic i e main enance, and ou e
op imiza ion. These insigh s can signi ican ly enhance ope a ional e iciency and decision-making
ac oss he logis ics chain.
4.10.3 Key Pe o mance Indica o s
Table 4.29: KPIs o End- o-end Logis ics use case.
De ice KPIs Ne wo k KPIs
KPI Value KPI Value
Sensing accu acy > 95 % End- o-end la ency < 1 s
De ice au onomy > 30 days Age o in o ma ion < 1 s
A g. powe consump ion < 1 µW Packe loss a e <0.1 %
Peak powe consump ion < 10 µW Message size < 10 kbi s
Communica ion ange < 10 m
Se ice a ea dimension < 100 m2
Max simul aneous de ices < 1000 de ices
T ans e in e al < 1 s
De ice speed < 5 m·s-1
Use -expe ienced da a a e 100 kbps
Applica ion KPIs AI/ML KPIs
KPI Value KPI Value
Communica ion se ice a ailabili y > 99 % T aining complexi y N.A.
Posi ioning/Loca ion se ice a ailabili y > 99 % In e encing accu acy N.A.
Posi ioning/Loca ion se ice accu acy < 0.5 m In e encing la ency N.A.
Command se ice a ailabili y No equi ed
AI/ML capabili ies No equi ed
4.10.4 Key Value Indica o s
En i onmen al KVIs
Impac Sco e: 3 The use o RFID echnology in end- o-end logis ics highly depends on he
exac implemen a ion o he echnology. I no mal ba code labels a e eplaced one- o-one wi h
RFID ags, he e will no be a huge en i onmen al impac . Bu i he en i e supply chain is
adap ed o he new echnology, and ecyclable RTI a e implemen ed, o ags a e di ec ly in e-
g a ed in o he de ices, he e o e educing he need o u he package labelling, and enabling
addi ional in o ma ion h oughou he li espan o he de ice, en i onmen al impac s can be
subs an ial.
Page 75 o 94

D5.1 - Use cases, KPIs and KVIs
Social KVIs
Impac Sco e: 4 By implemen ing RFID echnology in end- o-end logis ics, imp o emen s o
he e iciency and accu acy o supply chain ope a ions can be made. This leads o as e deli e y
imes, educed s ock-ou s, and imp o ed cus ome sa is ac ion. The eal- ime isibili y p o ided
by RFID echnology also enhances he anspa ency o he supply chain, allowing consume s o
ack he jou ney o hei p oduc s om he manu ac u e o hei doo s ep. De ice implemen ed
ags u he allow he end use o de ec he au hen ici y o he de ice, o enable o ga he mo e
in o ma ion on he de ice h oughou i s li espan.
Economic KVIs
Impac Sco e: 4 By educing he need o manual in en o y checks and imp o ing he accu acy
o in en o y eco ds, RFID echnology helps o educe labo cos s and minimize losses due o
s ock disc epancies, leading o cos sa ings o businesses. The imp o ed e iciency o logis ics
ope a ions also leads o cos sa ings in anspo a ion and wa ehousing, u he enhancing he
economic bene i s o RFID echnology.
Inno a ion KVIs
Impac Sco e: 3 The use o RFID echnology in end- o-end logis ics ep esen s a mode a e
inno a ion in supply chain managemen . By p o iding eal- ime isibili y and accu a e acking o
p oduc s, RFID echnology enables businesses o op imize hei logis ics ope a ions and imp o e
o e all supply chain e iciency. Mos ly he inno a i e aspec s will be p esen in he abili y o he
end use in e acing wi h he de ice ags, o example, o check au hen ici y, ead ou de ice
in o ma ion e y easily, like manuals o da a-shee s, o e en check de ice s a us, i he ag is
in eg a ed in o he de ice unc ionali y.
En i onmen al
Social
Economic
Inno a ion
12345
Figu e 4.18: End- o-end Logis ics.
Page 76 o 94
D5.1 - Use cases, KPIs and KVIs
4.11 Indus ial P edic i e Main enance
Table 4.30: Use Case Families o Indus ial P edic i e Main enance use case.
Use Case Families
Ha d- o- each de ices Massi e-scale senso ne wo ks Ex eme-li e ime applica ions
Table 4.31: END classes o Indus ial P edic i e Main enance use case.
END Classes
UP ENDs DB ENDs DP ENDs CO ENDs
4.11.1 Desc ip ion
P edic i e main enance suppo s he iden i ica ion o ea ly signs o equipmen wea , educing
unplanned down ime, and maximizing se ice schedules. In indus ial applica ions, ene gy-neu al
sensing enables con inuous, wi eless moni o ing o ha d- o- each o dis ibu ed asse s wi hou
he hassle o ba e y main enance.
Figu e 4.19: Illus a ion o P edic i e Main enance use case.
Typical applica ions a e as ollows.
•S agnan wa e de ec ion on ac o y o wa ehouse la oo s, whe e s agnan wa e can
cause s uc u al damage i no de ec ed.
•Machine hou coun e s, wi h un ime moni o ing de e mining main enance planning
based on ac ual usage ins ead o ixed in e als.
•Vib a ion analysis o mo o s, ans, o comp esso s, enabling ea ly de ec ion o misalign-
men , imbalance, o mechanical wea h ough spec al p ocessing.
•The mal moni o ing, wi h g adual empe a u e d i in bea ings o elec ical componen s
indica ing impending aul s.
Page 77 o 94
D5.1 - Use cases, KPIs and KVIs
These senso s o en ope a e wi h pe iodic powe supplied by ha es ing sou ces such as ib a ion,
indoo ligh , o empe a u e g adien s. E en -d i en upda es and local p ocessing educe ene gy
demands and enable ba e yless o ba e y-assis ed ope a ion o yea s.
4.11.2 Func ional Requi emen s
Communica ion
To conse e ene gy, de ices communica e in equen ly o on e en de ec ion. Low- h oughpu
sho - ange wi eless p o ocols (e.g., sub-GHz, BLE, o indus ial LoRa a ian s) a e su icien
o ansmi ale s o s a us messages. Whe e backsca e o wake-up adios a e used, commu-
nica ion is asynch onous and low-powe . Ga eways ga he he da a and o wa d hem o he
cloud o local con ol sys ems o diagnos ics o isualiza ion. The employed Medium Access
Con ol (MAC) s a egies a ou ene gy e iciency o e complexi y. Simple con en ion-based ap-
p oaches (e.g., ALOHA a ian s) a e common bu o e limi ed obus ness, especially in dense
deploymen s, due o hei " i e-and- o ge " na u e and lack o collision a oidance. Mo e eliable
al e na i es, such as du y-cycled ime di ision-mul iple access (TDMA), can p o ide p edic able
access wi h ewe collisions, bu equi e synch oniza ion and coo dina ion.
Posi ioning/Loca ion
P ecise posi ioning is usually no equi ed. The de ice ins alla ion is associa ed wi h speci ic
machines o loca ions, and he me ada a can be manually agged du ing he ins alla ion. In
dis ibu ed o mobile asse s, he g anula i y o he zone o oom le el can be su icien .
Managemen
Each senso mus ha e a unique ID and be associa ed wi h i s in ended asse in an asse man-
agemen sys em o digi al win. Up ime, signal quali y, and powe a ailabili y as minimum heal h
indica o s can be checked om ime o ime. De ices mus suppo aul de ec ion and ma king
o allow scheduled in e en ion o eplacemen when necessa y. In addi ion, suppo o OTA up-
da es and emo e pa ame e con igu a ion is essen ial, enabling i mwa e upda es o ope a ional
adjus men s o be applied wi hou physical access. These upda es can be igge ed manually by
ope a o s in esponse o diagnos ics, o au oma ically pushed by cloud-based sys ems as pa o
policy-d i en wo k lows. Fo example, i he cloud de ec s ha a senso is sending da a mo e
equen ly han necessa y, causing unnecessa y ene gy d ain, i can au oma ically issue a down-
link o adjus he epo ing in e al o de ec ion h eshold. Beyond de ice-le el managemen ,
he ne wo k mus suppo e ec i e commissioning o new de ices and o de ly decommissioning
o obsole e o mal unc ioning ones.
Collec ed In o ma ion and Ne wo k Exposu e
The in o ma ion ga he ed by ENDs used in indus ial p edic i e main enance con ex ypically
includes un ime logs such as hou s ac i e, e en coun s like he p esence o wa e o ib a ion
anomalies, and condi ion me ics including empe a u e and ib a ion spec um analysis. A key
challenge in his con ex is balancing he demand o nea eal- ime access o da a wi h he
need o conse e de ice ene gy, as mo e equen ansmissions signi ican ly inc ease ene gy
consump ion and hus educe de ice li e ime.
Page 78 o 94
D5.1 - Use cases, KPIs and KVIs
Indus ial eleme y is no equi ed o be sensi i e, bu mus be p o ec ed wi h enc yp ion and se-
cu e au hen ica ion. Ga eways and s o age need o o e in eg i y, pa icula ly in hose ins ances
whe e da a a ec main enance choices o egula o y compliance.
Ene gy
Ene gy-neu al ope a ion can ely on a a ie y o ene gy ha es ing sou ces: mos commonly
ambien ligh , possibly combined wi h o he s such as ib a ion o he mal g adien s o imp o e
obus ness by compensa ing o he a iabili y o indi idual sou ces. De ices bu e ha es ed
ene gy using (a combina ion o ) ene gy s o age elemen s such as supe capaci o s (enabling apid
discha ge o in e mi en sensing, p ocessing, and ansmission asks) and ba e ies (le e aging
high ene gy densi y).
E en -d i en a chi ec u es a oid unnecessa y ene gy consump ion. Ba e y-assis ed nodes can
be accep able o mul i-yea deploymen bu should be a oided whe e eplacemen is no possible.
Use -Sa e y and Robus ness
De ices mus be indus ial-g ade: esis an o ex emes o dus , mois u e, ib a ion, and em-
pe a u e o ensu e eliable ope a ion unde ha sh condi ions. They mus also unc ion eliably
amid elec ical noise, me allic obs uc ions, and ha sh elec omagne ic en i onmen s ypical o
indus ial se ings, whe e wi eless communica ion aces e lec ions, in e e ence, and signal a -
enua ion. To p e en dis up ion, de ices should be designed so ha hei ope a ion does no
in e e e wi h he asse s hey moni o . The physical size o he de ice mus be app op ia e
o he use case: gene ally less es ic i e in open deploymen s such as oo op moni o ing bu
po en ially c i ical in con ined spaces such as size-cons ained machine y, whe e compac ness
is necessa y o a oid impeding ope a ion o main enance. Finally, de ices should inco po a e
ail-sa es such as wa chdog ime s and allback logic o handle communica ion loss o in e nal
ailu es g ace ully.
AI/ML
Basic signal p ocessing echniques, such as peak de ec ion o as Fou ie ans o m (FFT), can
ypically be pe o med locally a he edge o p e- il e o classi y da a, hus educing he olume o
da a ha needs o be ansmi ed. Local ML on he senso is cons ained by ene gy a ailabili y,
which leads o a ade-o be ween pe o ming ligh weigh local in e ence wi h po en ially lowe
accu acy and o loading da a o cloud-based ML models, which may o e highe accu acy,
bu equi e mo e ene gy o communica ion. Cloud-side AI models can agg ega e da a ac oss
mul iple asse s, iden i y long- e m ends, and p o ide mo e accu a e ailu e o ecas s.
Page 79 o 94
D5.1 - Use cases, KPIs and KVIs
and main enance ope a ions. T ans e lea ning echniques may be applied o enable obo s and
s andalone ENDs o e ine AI models using sha ed en i onmen al con ex , educing edundan
e aining. Fu he , apa om being used o si ua ional awa eness and obo con ol pu poses
(such as manoeu ing), AI/ML models can be used o assis wi h communica ion asks, o
example, by ecommending di e en END du y cycling o signal ansmission/ ecep ion policies
o in e e ence mi iga ion o a oidance. The sys em mus also enable issuance and ans e
o p edic ed END ene gy consump ion along in ended obo pa hs, le e aging eal- ime and
his o ical da a o op imize ou ing and esou ce alloca ion.
4.12.3 Key Pe o mance Indica o s
Table 4.35: KPIs o Coope a i e Mobile Robo s use case.
De ice KPIs Ne wo k KPIs
KPI Value KPI Value
Sensing accu acy > 95 % End- o-end la ency < 0.5 s
De ice au onomy > 10 yea s Age o in o ma ion < 0.5 s
A g. powe consump ion < 1 mW Packe loss a e <0.1 %
Peak powe consump ion < 50 mW Message size < 1 kbi
Communica ion ange < 20 m
Se ice a ea dimension < 10000 m2
Max simul aneous de ices < 1000 de ices
T ans e in e al On eques
De ice speed < 5 m·s-1
Use -expe ienced da a a e 100 kbps
Applica ion KPIs AI/ML KPIs
KPI Value KPI Value
Communica ion se ice a ailabili y > 99.9 % T aining complexi y < 1 GFLOP
Posi ioning/Loca ion se ice a ailabili y > 99 % In e encing accu acy > 90 %
Posi ioning/Loca ion se ice accu acy < 1 m In e encing la ency < 10 s
Command se ice a ailabili y No equi ed
AI/ML capabili ies Yes
4.12.4 Key Value Indica o s
En i onmen al KVIs
Impac Sco e: 3 ENDs equipped wi h ene gy ha es ing and wi eless powe ans e can
educe eliance on con en ional ba e ies, ensu ing sus ainable ope a ions. Howe e , obo ic
deploymen s ill equi es ene gy o communica ion, p ocessing, and obo cha ging, con ibu ing
o en i onmen al impac . Mi iga ion s a egies mus inco po a e enewable ene gy sou cing,
ene gy-e icien ha dwa e/so wa e design, and ecyclable componen s, educing li ecycle was e.
Page 86 o 94

D5.1 - Use cases, KPIs and KVIs
Social KVIs
Impac Sco e: 2 A-IoT migh con ibu e o longe ope a ional up ime and educed main-
enance e o s in a coope a i e obo ic sys em, which could sligh ly in luence wo k o ce oles
(e.g., ewe asks ela ed o ba e y eplacemen o de ice moni o ing). Howe e , his ea u e
would no signi ican ly al e he b oade social conside a ions, as compa ed o a coope a i e
mobile obo sys em inco po a ing con en ional IoT de ices.
Economic KVIs
Impac Sco e: 4 Unlike con en ional IoT-equipped obo s, END-enabled sys ems le e age
ene gy ha es ing WPT, he eby educing dependency on ba e y eplacemen s and lowe ing
long- e m ope a ional cos s. This shi enhances cos -e ec i eness, especially o smalle busi-
nesses ha may s uggle wi h high main enance expenses. F om a ma ke pe spec i e, A-IoT
adop ion could lowe en y ba ie s o businesses by educing ha dwa e cos s, bu i may also
in ensi y monopoliza ion isks i only la ge co po a ions can a o d ad anced A-IoT-powe ed
obo ics. In addi ion, egula o y amewo ks mus e ol e o add ess da a secu i y, ene gy ha -
es ing policies, and A-IoT-d i en au oma ion go e nance, ensu ing ai compe i ion and e hical
deploymen .
Inno a ion KVIs
Impac Sco e: 4 The he mal ene gy dissipa ed om AI/ML wo kload p ocessing can be
pa ially cap u ed h ough he moelec ic ha es ing and ealloca ed o c i ical obo ic asks,
such as con inuing model aining o ensu ing unin e up ed senso da a ansmission, hence,
educing dependency on ex e nal powe sou ces. The e o e, a KVI in his con ex would be on
he “ene gy euse po en ial” aining o in e ence da a may ca y.
Also, be o e ini ia ing T ans e Lea ning, he ne wo k mus assess he use ulness o he AI/ML
model in i s cu en con ex , weighing his decision agains an al e na i e decision on wi elessly
ans e ing ene gy. A new KVI “AI/ML model use ulness” can be shaped o e alua e he mos
e ec i e decision in space and in ime.
Page 87 o 94
D5.1 - Use cases, KPIs and KVIs
En i onmen al
Social
Economic
Inno a ion
12345
Figu e 4.22: Coope a i e Mobile Robo s.
Page 88 o 94
D5.1 - Use cases, KPIs and KVIs
Chap e 5
Conclusions
In his documen , an analysis o use cases o A-IoT is conduc ed. Ini ially, A-IoT and he ech-
nologies suppo ing i a e in oduced, p o iding he con ex o unde s anding he es o he
documen . A e iew o he s a e o he a on he use case de ini ion o A-IoT is pe o med,
ocusing on ac i i ies ca ied ou by SDOs and Eu opean p ojec s. The co e o he documen
desc ibes se e al p oposed use cases. These a e examined h ough he iden i ica ion o unc-
ional equi emen s, KPIs, and KVIs in ou dimensions: en i onmen al, social, economic, and
inno a i e. Wi h his app oach, a iew o he po en ial o A-IoT is p o ided, and insigh s in o
i s p ac ical implemen a ion a e gene a ed. I is impo an o highligh ha he KVI analysis
pe o med by he expe s con ibu ing o his deli e able should be conside ed as p elimina y,
as he ac o s ha i conside s a e di icul o quan i y and an icipa e. Fu he s ages o his
analysis and mo e esea ch on END capabili ies a e needed o make i mo e accu a e.
Deli e able D5.1 uses D2.1 as one o i s p ima y e e ences, le e aging on he classi ica ion o
ENDs. Mo eo e , D5.1 analyses use cases ha will se e as he ounda ion o he upcoming
deli e ables D5.2 and D5.3, which will de elop he co esponding p oo s o concep and demon-
s a o s. D5.1 is also a e e ence o deli e ables D3.1, D3.2, D3.3, D4.1, D4.2, and D4.3, as
i p o ides ele an KPIs and KVIs o suppo he de elopmen o p o ocols, in as uc u es and
algo i hms o be desc ibed in hose documen s. D5.1 di ec ly con ibu es o he achie emen
o Miles one 2: "Technology ade-o s, use cases, KPIs, and KVIs de ined". Addi ionally, i
con ibu es indi ec ly o he ul ilmen o Speci ic Objec i e 4: "Implemen , es , alida e, and
demons a e end- o-end in elligen END capabili ies, and 6G ne wo k unc ionali y" by de ining
he use cases ha will guide he de elopmen o he AMBIENT-6G’s p oo s o concep and
demons a o s.
Page 89 o 94
D5.1 - Use cases, KPIs and KVIs
Re e ences
[1] AMBIENT-6G, D2.1 END a chi ec u e, echnology ade-o , and cha ac e iza ion, May 2025.
[2] T. Pi son, L. Gola d, and D. Bol, “E alua ing he (i ) ele ance o IoT solu ions wi h espec o en i onmen al
limi s based on LCA and backcas ing s udies,” in Nin h Compu ing wi hin Limi s 2023, LIMITS, Jun. 2023.
[3] In e na ional Telecommunica ion Union. “C ea ing a ci cula economy o ICT equipmen .” (2023), [Online].
A ailable: h ps://www.i u.in /en/mediacen e/backg ounde s/Pages/e-was e.aspx ( isi ed on
11/24/2023).
[4] W. Li, Q. Liu, Y. Zhang, C. Li, Z. He, W. C. H. Choy, P. J. Low, P. Sona , and A. K. K. Kyaw, “Biodeg adable
Ma e ials and G een P ocessing o G een Elec onics,” Ad anced Ma e ials, ol. 32, no. 33, p. 2 001 591,
2020. DOI: 10.1002/adma.202001591.
[5] D. Danninge , R. P uckne , L. Holzinge , R. Koeppe, and M. Kal enb unne , “MycelioT onics: Fungal
mycelium skin o sus ainable elec onics,” Science Ad ances, ol. 8, no. 45, eadd7118, 2022. DOI: 10.
1126/sciad .add7118.
[6] A. S. Weddell, M. Magno, G. V. Me e , D. B unelli, B. M. Al-Hashimi, and L. Benini, “A su ey o mul i-
sou ce ene gy ha es ing sys ems,” in 2013 Design, Au oma ion & Tes in Eu ope Con e ence & Exhibi ion
(DATE), 2013, pp. 905–908. DOI: 10.7873/DATE.2013.190.
[7] H. Yang, M. Huang, M. Ren, and X. Li, “Piezoelec ic ene gy ha es ing in e ace ci cui o small a ea
and low powe consump ion‚ a e iew,” Measu emen , ol. 242, p. 116 051, 2025, ISSN: 0263-2241. DOI:
h ps : / / doi . o g / 10 . 1016 / j . measu emen . 2024 . 116051. [Online]. A ailable: h ps : / / www .
sciencedi ec .com/science/a icle/pii/S0263224124019365.
[8] S.-E. Kim, T. Kang, K.-I. Oh, M. J. Pa k, H.-I. Pa k, I. G. Lim, and J.-J. Lee, “Ene gy Managemen
In eg a ed Ci cui o Mul i-Sou ce Ene gy Ha es e s in WBAN Applica ions,” Applied Sciences, ol. 8,
no. 8, 2018, ISSN: 2076-3417. DOI: 10.3390/app8081262. [Online]. A ailable: h ps://www.mdpi.com/
2076-3417/8/8/1262.
[9] C. Delgado, J. M. Sanz, and J. Famaey, “On he Feasibili y o Ba e y-Less LoRaWAN Communica ions
Using Ene gy Ha es ing,” in 2019 IEEE Global Communica ions Con e ence (GLOBECOM), 2019.
[10] C. Delgado, J. M. Sanz, C. Blondia, and J. Famaey, “Ba e yless LoRaWAN Communica ions Using Ene gy
Ha es ing: Modeling and Cha ac e iza ion,” IEEE In e ne o Things Jou nal, ol. 8, no. 4, pp. 2694–2711,
2021.
[11] M. Capuzzo, C. Delgado, A. K. Sul ania, J. Famaey, and A. Zanella, “Enabling G een IoT: Ene gy-Awa e
Communica ion P o ocols o Ba e y-less LoRaWAN De ices,” in P oceedings o he 24 h In e na ional
ACM Con e ence on Modeling, Analysis and Simula ion o Wi eless and Mobile Sys ems, 2021, pp. 95–98.
[12] A. K. Sul ania, C. Delgado, and J. Famaey, “Enabling Low-La ency Blue oo h Low Ene gy on Ene gy
Ha es ing Ba e yless De ices Using Wake-Up Radios,” Senso s, ol. 20, no. 18, 2020.
[13] A. K. Sul ania, C. Delgado, C. Blondia, and J. Famaey, “Downlink Pe o mance Modeling and E alua ion
o Ba e yless Low Powe BLE Node,” Senso s, ol. 22, no. 8, 2022.
[14] A. K. Sul ania and J. Famaey, “Ba e yless Blue oo h Low Ene gy P o o ype Wi h Ene gy-Awa e Bidi ec-
ional Communica ion Powe ed by Ambien Ligh ,” IEEE Senso s Jou nal, ol. 22, no. 7, pp. 6685–6697,
2022.
[15] A. K. Sul ania and J. Famaey, “Ba e yless NB-IoT p o o ype o bidi ec ional communica ion powe ed by
ambien ligh ,” Ad Hoc Ne wo ks, ol. 142, p. 103 100, 2023.
[16] L. Moons, S. Nasse , A. Sabo ic, R. K. Singh, and J. Famaey, “E alua ing Fas and G an -F ee Uplink
Access in Nex -Gene a ion Cellula IoT Ne wo ks,” in P oceedings o he 2024 3 d In e na ional Con e ence
on 6G Ne wo king (6GNe ), IEEE, 2024, pp. 19–24. DOI: 10.1109/6GNe 63182.2024.10765754. [Online].
A ailable: h ps://doi.o g/10.1109/6GNe 63182.2024.10765754.
Page 90 o 94
D5.1 - Use cases, KPIs and KVIs
[17] N. A. Mi siou, S. A. Tegos, P. D. Diaman oulakis, P. G. Sa igiannidis, and G. K. Ka agiannidis, “P oac i e
Scheduling o Ze o-Ene gy De ice Ne wo ks Wi h Fas Uplink G an ,” IEEE Wi eless Communica ions
Le e s, ol. 12, no. 7, pp. 1184–1188, 2023.
[18] “3GPP TR 38.769 V19.0.0 S udy on solu ions o Ambien IoT (In e ne o Things) in NR.” [Online].
A ailable: h ps://www.3gpp.o g/ p/Specs/a chi e/38_se ies/38.769/38769-j00.zip.
[19] R. Mish a and A. Mish a, “Cu en esea ch on In e ne o Things (IoT) secu i y p o ocols: A su ey,”
Compu e s & Secu i y, p. 104 310, 2025.
[20] G. Callebau , G. Leende s, J. Van Mulde s, G. O oy, L. De S ycke , and L. Van de Pe e, “The A
o Designing Remo e IoT De ices–Technologies and S a egies o a Long Ba e y Li e,” Senso s, ol. 21,
no. 3, Jan. 2021, ISSN: 1424-8220. DOI: 10.3390/s21030913. [Online]. A ailable: h ps://www.mdpi.
com/1424-8220/21/3/913.
[21] P. Mi anda, M. Siekkinen, and H. Wa is, “TLS and ene gy consump ion on a mobile de ice: A measu emen
s udy,” in 2011 IEEE Symposium on Compu e s and Communica ions (ISCC), IEEE, 2011, pp. 983–989.
[22] K. O. Ade emi Alimi, K. Ouahada, A. M. Abu-Mah ouz, and S. Rime , “A su ey on he secu i y o low
powe wide a ea ne wo ks: Th ea s, challenges, and po en ial solu ions,” Senso s, ol. 20, no. 20, p. 5800,
2020.
[23] X. Yang, L. Shu, Y. Liu, G. P. Hancke, M. A. Fe ag, and K. Huang, “Physical secu i y and sa e y o IoT
equipmen : A su ey o ecen ad ances and oppo uni ies,” IEEE T ansac ions on Indus ial In o ma ics,
ol. 18, no. 7, pp. 4319–4330, 2022.
[24] M.-K. Oh, S. Lee, and Y. Kang, “Wi-sun de ice au hen ica ion using physical laye inge p in ,” in 2021
In e na ional Con e ence on In o ma ion and Communica ion Technology Con e gence (ICTC), IEEE, 2021,
pp. 160–162.
[25] F. Alwahedi, A. Aldhahe i, M. A. Fe ag, A. Ba ah, and N. Tihanyi, “Machine lea ning echniques o
IoT secu i y: Cu en esea ch and u u e ision wi h gene a i e AI and la ge language models,” In e ne o
Things and Cybe -Physical Sys ems, ol. 4, pp. 167–185, 2024.
[26] J. Van Mulde s, S. Boeckx, J. Cappelle, L. Van de Pe e, and L. De S ycke , “UAV-based solu ion o
ex ending he li e ime o IoT de ices: e iciency, design and sus ainabili y,” F on ie s in Communica ions and
Ne wo ks, ol. Volume 5 - 2024, 2024, ISSN: 2673-530X. DOI: 10.3389/ cmn.2024.1341081.
[27] B. J. B. Deu schmann, U. Muehlmann, A. Kaplan, G. Callebau , T. Wilding, B. Cox, L. V. de Pe e, F.
Tu esson, E. G. La sson, and K. Wi isal, “Physically La ge Ape u es o Wi eless Powe T ans e : Pe o -
mance and Regula o y Aspec s,” IEEE Wi eless Communica ions Magazine, 2025, Accep ed o publica ion.
[Online]. A ailable: h ps://a xi .o g/abs/2503.06807.
[28] U. Kunna h Ganesan, R. Sa end ana h, and E. G. La sson, “BeamSync: O e - he-Ai Synch oniza ion
o Dis ibu ed Massi e MIMO Sys ems,” IEEE T ansac ions on Wi eless Communica ions, ol. 23, no. 7,
pp. 6824–6837, 2024. DOI: 10.1109/TWC.2023.3335089.
[29] B. J. B. Deu schmann and P. Vou as, Spa io empo al Synch oniza ion o Dis ibu ed A ays using Pa icle-
Based Loopy Belie P opaga ion, 2025. a Xi : 2504 .00744 [eess.SP]. [Online]. A ailable: h ps :/ /
a xi .o g/abs/2504.00744.
[30] K. Wu, J. A. Zhang, X. Huang, R. W. Hea h, and Y. J. Guo, “G een Join Communica ions and Sensing
Employing Analog Mul i-Beam An enna A ays,” IEEE Communica ions Magazine, ol. 61, no. 7, pp. 172–
178, 2023. DOI: 10.1109/MCOM.007.2200495.
[31] A. Aza bah am, O. L. A. López, and M. La a-Aho, “Wa e o m Op imiza ion and Beam Focusing o Nea -
Field Wi eless Powe T ans e Wi h Dynamic Me asu ace An ennas and Non-Linea Ene gy Ha es e s,”
IEEE T ansac ions on Wi eless Communica ions, ol. 24, no. 2, pp. 1031–1045, 2025. DOI: 10.1109/TWC.
2024.3503908.
[32] V. Jamali, A. M. Tulino, G. Fische , R. R. Mülle , and R. Schobe , “In elligen Su ace-Aided T ansmi e A -
chi ec u es o Millime e -Wa e Ul a Massi e MIMO Sys ems,” IEEE Open Jou nal o he Communica ions
Socie y, ol. 2, pp. 144–167, 2021. DOI: 10.1109/OJCOMS.2020.3048063.
[33] L. D’Angelo, B. J. Deu schmann, and K. Wi isal, “Combined Wideband Channel Es ima ion and Di ec
Link In e e ence Mi iga ion in Bis a ic Backsca e Sys ems,” in 2024 IEEE 25 h In e na ional Wo kshop
on Signal P ocessing Ad ances in Wi eless Communica ions (SPAWC), 2024, pp. 506–510. DOI: 10.1109/
SPAWC60668.2024.10694033.
[34] A. Kaplan, J. Viei a, and E. G. La sson, “Di ec Link In e e ence Supp ession o Bis a ic Backsca e
Communica ion in Dis ibu ed MIMO,” IEEE T ans. Wi eless Commun., ol. 23, no. 2, pp. 1024–1036,
2024.
Page 91 o 94

D5.1 - Use cases, KPIs and KVIs
[35] O. L. A. López, H. Al es, R. D. Souza, and S. Mon ejo-Sánchez, “S a is ical Analysis o Mul iple An enna
S a egies o Wi eless Ene gy T ans e ,” ol. 67, no. 10, pp. 7245–7262, 2019. DOI: 10.1109/TCOMM.
2019.2928542.
[36] R. M. D ei ue s and R. W. Hea h, “Massi e MIMO in 5G: How Beam o ming, Codebooks, and Feedback
Enable La ge A ays,” IEEE Communica ions Magazine, ol. 61, no. 12, pp. 18–23, 2023. DOI: 10.1109/
MCOM.001.2300064.
[37] B. J. B. Deu schmann, T. Wilding, M. G abe , and K. Wi isal, “XL-MIMO Channel Modeling and P edic ion
o Wi eless Powe T ans e ,” in 2023 IEEE In e na ional Con e ence on Communica ions Wo kshops (ICC
Wo kshops), 2023, pp. 1355–1361. DOI: 10.1109/ICCWo kshops57953.2023.10283480.
[38] Tenso Flow, Tenso Flow Li e o Mic ocon olle s,h ps://www. enso low.o g/li e/mic ocon olle s,
2024.
[39] A m L d., CMSIS-NN: E icien Neu al Ne wo k Ke nels o A m Co ex-M CPUs,h ps://de elope .
a m.com/solu ions/machine-lea ning-on-a m/de elope -ma e ial/how- o-guides/cmsis-nn,
2024.
[40] A. Adi e al., “Ene gy-Awa e Machine Lea ning o IoT De ices: A Su ey,” IEEE In e ne o Things Jou nal,
2020. [Online]. A ailable: h ps://a xi .o g/pd /2309.00022.
[41] A. Dha e al., “On-De ice Machine Lea ning: A Su ey om he Algo i hmic and Lea ning Theo y Pe spec-
i e,” a Xi p ep in a Xi :2301.12345, 2023. [Online]. A ailable: h ps://a xi .o g/abs/2301.12345.
[42] Y. Zhang e al., “Tiny Con inual Lea ning on Ene gy-Ha es ing De ices,” P oceedings o he ACM on
In e ac i e, Mobile, Wea able and Ubiqui ous Technologies, ol. 6, no. 3, pp. 1–25, 2022.
[43] B. Sudha san, J. G. B eslin, M. Tahi , M. In iza Ali, O. Rana, S. Dus da , and R. Ranjan, “OTA-TinyML:
O e he Ai Deploymen o TinyML Models and Execu ion on IoT De ices,” en, IEEE In e ne Compu ing,
ol. 26, no. 3, pp. 69–78, May 2022, ISSN: 1089-7801, 1941-0131. DOI: 10.1109/MIC.2021.3133552.
[Online]. A ailable: h ps://ieeexplo e.ieee.o g/documen /9811289/.
[44] 3GPP, Speci ica ion # 38.848; S udy on Ambien IoT (In e ne o Things) in RAN; Release 18, Sep. 2023.
[Online]. A ailable: h ps://po al.3gpp.o g/desk opmodules/Speci ica ions/Speci ica ionDe ails.
aspx?speci ica ionId=4146.
[45] 3GPP, Speci ica ion # 22.840; S udy on Ambien powe -enabled In e ne o Things, Dec. 2023. [Online].
A ailable: h ps://po al.3gpp.o g/desk opmodules/Speci ica ions/Speci ica ionDe ails.
aspx?speci ica ionId=4045.
[46] 3GPP, Speci ica ion # 22.369; Se ice equi emen s o Ambien powe -enabled IoT, Sep. 2024. [Online].
A ailable: h ps://po al.3gpp.o g/desk opmodules/Speci ica ions/Speci ica ionDe ails.
aspx?speci ica ionId=4232.
[47] IEEE, IEEE 802.11-23/0436 0; Technical Repo on suppo o AMP IoT de ices in WLAN, Ma . 2023.
[Online]. A ailable: h ps://men o .ieee.o g/802.11/dcn/23/11-23-0436-00-0amp- echnical-
epo -on-suppo -o -amp-io -de ices-in-wlan.docx.
[48] ADROIT6G, D2.1 Requi emen s, use cases and sys em a chi ec u e – Ini ial, Jul. 2023. [Online]. A ailable:
h ps://ad oi 6g.eu/wp-con en /uploads/2023/12/ADROIT6G_D2.1_Requi emen s-Use-cases-
A chi ec u e_Final.pd .
[49] ADROIT6G, D2.3 Requi emen s, use cases and sys em a chi ec u e, Jan. 2024. [Online]. A ailable: h ps:
//ad oi 6g.eu/wp-con en /uploads/2025/01/ADROIT6G_D2.3_Final.pd .
[50] HEXA-X, Deli e able D1.1 6G Vision, use cases and key socie al alues, Feb. 2021. [Online]. A ailable:
h ps://hexa-x.eu/wp-con en /uploads/2021/02/Hexa-X_D1.1.pd .
[51] HEXA-X-II, Deli e able D1.2 6G Use Cases and Requi emen s, Dec. 2023. [Online]. A ailable: h ps:
//hexa-x-ii.eu/wp-con en /uploads/2024/01/Hexa-X-II_D1.2.pd .
[52] In ellIoT, Deli e able D2.1 Use case speci ica ion & Open Call de ini ion ( i s e sion), Ma . 2021. [Online].
A ailable: h ps://in ellio .eu/wp-con en /uploads/2022/09/Deli e able-D2.1-Use-case-
speci ica ion-and-Open-Call-de ini ion-1.pd .
[53] In ellIoT, Deli e able D2.4 Use Case speci ica ion & Open Call de ini ion ( inal e sion), Ap . 2022. [Online].
A ailable: h ps://in ellio .eu/wp-con en /uploads/2022/10/In ellIoT_deli e able_D2.4.
pd .
[54] REINDEER, D1.1. Use case-d i en speci ica ions and echnical equi emen s and ini ial channel model, Oc .
2021. [Online]. A ailable: h ps://zenodo.o g/ eco ds/5561844.
[55] SUPERIOT, D1.4 Desc ip ion o selec ed scena ios, applica ions, and hei equi emen s (2), Dec. 2024.
[56] “How he Blue oo h®Elec onic Shel Label S anda d Will Impac he Sma Re ail Ma ke ,” ABI Resea ch,
Tech. Rep., Jun. 2023.
Page 92 o 94
D5.1 - Use cases, KPIs and KVIs
[57] J.-S. Pa k and B.-J. Jang, “Elec onic shel label sys em employing a isible ligh iden i ica ion link,” in 2016
IEEE Radio and Wi eless Symposium (RWS), 2016, pp. 219–222.
[58] E. E. Dikel, Y. E. Li, M. Vuo a i, and S. Mancini, “E alua ing he s andby powe consump ion o sma
LED bulbs,” Ene gy and Buildings, ol. 186, pp. 71–79, 2019, ISSN: 0378-7788. DOI: h ps://doi.o g/
10.1016/j.enbuild.2019.01.019.
[59] J. Cappelle, L. V. de Pe e, E. Fi zge ald, S. Ra y s, W. Gajda, V. D. Smed , B. Cox, and G. Callebau , IoT
on he Road o Sus ainabili y: Vehicle o Bandi ? 2024. a Xi : 2405.20706 [eess.SP]. [Online]. A ailable:
h ps://a xi .o g/abs/2405.20706.
[60] B. Aliakba ian, S. Ghi landi, A. Rizzi, R. S e anini, and G. Vignali, “Li e cycle assessmen o plas ic and
pape -based ul a high equency RFID ags,” In e na ional Jou nal o RF Technologies, ol. 14, no. 1,
pp. 17–32, Jun. 2024. DOI: 10.3233/RFT-230044.
[61] J. Sennesael, A. Moe man, H. Rogie , and P. Van To e, “Ca bon-Ink Sensing Pa e ns o a Con ac less
Sma Diape Sys em,” in 2024 IEEE In e na ional Con e ence on Flexible and P in able Senso s and Sys ems
(FLEPS), 2024, pp. 1–4. DOI: 10.1109/FLEPS61194.2024.10603972.
[62] J. Li, Z. Xie, Z. Wang, Z. Lin, C. Lu, Z. Zhao, Y. Jin, J. Yin, S. Mu, C. Zhang, W. Gui, X. Liang, J. Wang,
and W. Ding, “A iboelec ic gai senso sys em o human ac i i y ecogni ion and use iden i ica ion,”
Nano Ene gy, ol. 112, p. 108 473, 2023, ISSN: 2211-2855. DOI: h ps : / / doi . o g / 10 . 1016 / j .
nanoen.2023.108473. [Online]. A ailable: h ps://www.sciencedi ec .com/science/a icle/pii/
S2211285523003105.
[63] H. Hassan, M. Abdullah, M. Z. Hussain, M. Zulki l Hasan, M. Mus a a, A. Khalid, R. Awan, U. Hussain,
Z. A. Khan, and A. Ja aid, “TinyML-Based Gai Recogni ion o Use De ec ion on an STM32L4 IoT
Node,” in 2024 IEEE 9 h In e na ional Con e ence o Con e gence in Technology (I2CT), 2024, pp. 1–6.
DOI: 10.1109/I2CT61223.2024.10543752.
[64] W. Chan aweesomboon, “Blue oo h Geo-Fence o Elde ly and Pa ien Ca e,” in 2021 25 h In e na ional
Compu e Science and Enginee ing Con e ence (ICSEC), 2021, pp. 252–255. DOI: 10.1109/ICSEC53205.
2021.9684648.
[65] A. N. A un, B. Lee, F. A. Cas iblanco, D. R. Buckmas e , C.-C. Wang, D. J. Lo e, J. V. K ogmeie , M. M.
Bu , and A. Ghosh, “Ambien IoT: Communica ions Enabling P ecision Ag icul u e,” IEEE Communica ions
Magazine, ol. 63, no. 4, pp. 137–143, 2025. DOI: 10.1109/MCOM.003.2400508.
[66] S. Yang, Y. Bénédic, D.-T. Phan-Huy, J.-M. Go ce, and G. Villemaud, “Indoo Localiza ion o Sma phones
Thanks o Ze o-Ene gy-De ices Beacons,” in 2024 18 h Eu opean Con e ence on An ennas and P opaga ion
(EuCAP), 2024, pp. 1–5. DOI: 10.23919/EuCAP60739.2024.10501308.
[67] J. Liao, X. Wang, K. Ru ik, R. Jän i, and D.-T. Phan-Huy, “In-Band Ambien FSK Backsca e Commu-
nica ions Le e aging LTE Cell-Speci ic Re e ence Signals,” IEEE Jou nal o Radio F equency Iden i ica ion,
ol. 7, pp. 267–277, 2023. DOI: 10.1109/JRFID.2023.3280108.
[68] J. Liao, T. Zhang, K. Ru ik, R. Jän i, and D.-T. Phan-Huy, “Ambien Backsca e Communica ion in
LTE Uplink Sounding Re e ence Signal,” a Xi p ep in a Xi :2501.10952, 2025.
[69] A. Fascis a, B. J. B. Deu schmann, M. F. Keskin, T. Wilding, A. Coluccia, K. Wi isal, E. Lei inge , G.
Seco-G anados, and H. Wymee sch, “Join Localiza ion, Synch oniza ion and Mapping ia Phase-Cohe en
Dis ibu ed A ays,” IEEE Jou nal o Selec ed Topics in Signal P ocessing, ol. 19, no. 2, pp. 412–429,
2025. DOI: 10.1109/JSTSP.2025.3533111.
[70] B. J. B. Deu schmann, E. Lei inge , and K. Wi isal, Geome y-Based Channel Es ima ion, P edic ion, and
Fusion, 2025. a Xi : 2503.17868 [eess.SP]. [Online]. A ailable: h ps://a xi .o g/abs/2503.17868.
[71] A. Fascis a, B. J. B. Deu schmann, M. F. Keskin, T. Wilding, A. Coluccia, K. Wi isal, E. Lei inge , G. Seco-
G anados, and H. Wymee sch, “Uplink Join Posi ioning and Synch oniza ion in Cell-F ee Deploymen s wi h
Radio S ipes,” in 2023 IEEE In e na ional Con e ence on Communica ions Wo kshops (ICC Wo kshops),
2023, pp. 1330–1336. DOI: 10.1109/ICCWo kshops57953.2023.10283650.
[72] C. Nelson, X. Li, A. Fedo o , B. Deu schmann, and F. Tu esson, “Dis ibu ed MIMO Measu emen s
o In eg a ed Communica ion and Sensing in an Indus ial En i onmen ,” Senso s, ol. 24, no. 5, 2024,
ISSN: 1424-8220. DOI: 10 . 3390 / s24051385. [Online]. A ailable: h ps : / / www . mdpi . com / 1424 -
8220/24/5/1385.
[73] B. Deu schmann, C. Nelson, M. Hen iksson, G. Ma i, A. Kosasih, N. Te o, E. Lei inge , and F. Tu esson,
“Accu a e Di ec Posi ioning in Dis ibu ed MIMO Using Delay-Dopple Channel Measu emen s,” in 2024
IEEE 25 h In e na ional Wo kshop on Signal P ocessing Ad ances in Wi eless Communica ions (SPAWC),
2024, pp. 606–610. DOI: 10.1109/SPAWC60668.2024.10694279.
Page 93 o 94
D5.1 - Use cases, KPIs and KVIs
[74] L. Gö schache , J. G osinge , H. N. Khan, and W. Bösch, “Fas wo dimensional posi ion upda e sys em
o UHF RFID ag acking,” in 2017 IEEE MTT-S In e na ional Mic owa e Symposium (IMS), 2017,
pp. 1331–1334. DOI: 10.1109/MWSYM.2017.8058857.
[75] A. Pi na , B. Be alanič, G. Ce a , M. Moho čič, M. Meža, and C. Fo una, “Towa ds Sus ainable Deep
Lea ning o Wi eless Finge p in ing Localiza ion,” in ICC 2022 - IEEE In e na ional Con e ence on Com-
munica ions, 2022, pp. 3208–3213. DOI: 10.1109/ICC45855.2022.9838464.
[76] A. Abedi, M. H. Mazahe i, O. Aba i, and T. B ech , “Wi ag: Re hinking backsca e communica ion o
wi i ne wo ks,” in P oceedings o he 17 h ACM wo kshop on ho opics in ne wo ks, 2018, pp. 148–154.
[77] E. Sol anaghaei, A. Donga e, A. P abhaka a, S. Kuma , A. Rowe, and K. Whi ehouse, “Tag i: Loca ing
ul a-low powe wi i ags using unmodi ied wi i in as uc u e,” P oceedings o he ACM on In e ac i e,
Mobile, Wea able and Ubiqui ous Technologies, ol. 5, no. 1, pp. 1–29, 2021.
[78] I. O. o S anda diza ion, Robo s and obo ic de ices — Sa e y equi emen s o indus ial obo s — Pa
1: Robo s, 2011. [Online]. A ailable: h ps://www.iso.o g/s anda d/51330.h ml.
[79] R. I. Associa ion and A. N. S. Ins i u e, Indus ial Robo s and Robo Sys ems — Sa e y Requi emen s,
2012. [Online]. A ailable: h ps://webs o e.ansi.o g/s anda ds/ ia/ansi ia 15062012.
Page 94 o 94