Abs ac —Minimally in asi e wi eless de ices, allowing he
sampling o gu ’s bac e ia, a e needed o a longi udinal
unde s anding o he ole o he mic obio a on he human
heal h. He ein, we p esen a no el magne ic ac ua ion sys em
i ing inside a 11.5 x 30.5 mm wi eless inges ible capsule.
Lacking any elec onic componen s, he capsule obo is
designed o he collec ion o mic obio a’s samples h ough
mechanical b ushing. Wi eless ac i a ion and in si u sampling
a e enabled by an ex e nal pe manen magne ic sou ce. This
componen , when app oaching he capsule, p og essi ely
allows: (1) he adhesion o he de ice o he mucosa, (2) he
exposu e o he b ushes, and (3) he sampling by mul iple
o a ions. Nume ical and analy ical models we e de eloped o
dimensioning he sys em, and we e alida ed by bench op
expe imen s.
I. INTRODUCTION
In he las decade, subs an ial esea ch has been
in es iga ing he ela ionship be ween he
mic obio a/mic obiome and he human heal h [1]. By
de ini ion, he mic obio a is he collec ion o 10-100 illion
mic obial cells li ing in symbiosis wi h he human body,
p ima ily, i.e. > 90%, in he gas oin es inal (GI) ac [2]. As
shown by mul iple s udies, an al e a ion o he gu mic obial
composi ion (i.e., dysbiosis) is associa ed wi h pa hogenesis
o se e al diseases, e.g. diabe es, obesi y, in lamma o y
bowel diseases, ca dio- ascula and also psychological
diso de s [1]. Howe e , whe he he dysbiosis is a cause, a
consequence, o a modula o o he disease has no been
ca e ully in es iga ed ye [3]. One o he main easons o
his esea ch gap, is due o he cu en in asi e echniques o
sampling he mic obio a (e.g., biopsy o mucosal b ushing
du ing endoscopic p ocedu es), hence limi ed only o
symp oma ic pa ien s. As a solu ion, he de elopmen o non-
in asi e de ices, such as inges ible wi eless capsules, would
make he sampling o mic obio a mo e accessible, including
* This wo k was suppo ed by he ATLAS p ojec . This p ojec has
ecei ed unding om he Eu opean Union’s Ho izon 2020 esea ch and
inno a ion p og amme unde he Ma ie Sklodowska-Cu ie g an ag eemen
No 813782.This wo k was also pa o he CISAS (MIUR-CNR) p ojec ,
unded by he I alian Minis y o Uni e si y and Resea ch.
M. Finocchia o, A. Menciassi and G. Ciu i a e wi h he BioRobo ics
Ins i u e, Scuola Supe io e San ’Anna, I aly and wi h he Depa men o
Excellence in Robo ics & AI, Scuola Supe io e San 'Anna, I aly. M.
Finocchia o is also wi h Cen e o Resea ch in Biomedical Enginee ing o
Uni e si a Poli ècnica de Ca alunya, Spain (co esponding au ho ’s e-mail:
ma ina. inocchia [email protected]).
M. Sp o ie i and C. Giosuè a e wi h Na ional Resea ch Council (CNR),
Ins i u e o An h opic Impac s and Sus ainabili y in he Ma ine
En i onmen (IAS), I aly.
G. D ago and F. Cibella a e wi h Na ional Resea ch Council (CNR),
Ins i u e o Biomedical Resea ch and Inno a ion (IRIB), I aly.
he small bowel, which is so a no easible by commonly
used echnologies. As a ma e o ac , his al e na i e would
ha e a massi e impac in clinical esea ch, as well as in
u u e mul i-modali y and in si u diagnos ic echniques [4].
In ecen yea s, ew endoscopic capsules o mic obio a
collec ion ha e been in es iga ed. Salem e al. de eloped an
inges ible de ice, whose opening is egula ed by he
dissolu ion o a bio-g adable coa ing, allowing he en ance
o he luid inside he capsule. A he same ime, he closing
is enabled by a passi e bis able mechanism, igge ed by
olume inc ease o a sponge. Al hough p omising, his
solu ion does no enable choosing he speci ic loca ion o
collec ion, nei he localizing he sample, and i does no
conside he quali y, quan i y and co ec s o age o he
ma e ial in o de o p ope ly analyze i s composi ion h ough
he bac e ial 16S RNA sequencing [5]. A mo e ad anced
solu ion is p esen ed by Rezaei e al., who collec ed he
bac e ia h ough an osmo ic pump, enabled by he dissolu ion
o an ex e nal coa ing. In his case, i was also possible o
success ully analyze he bac e ia DNA’ samples, collec ed in
in- i o condi ions. None heless, a spa ially a ge ed sampling
is s ill no easible in his scena io, and he au ho s exp essed
hei in en ion o use a magne ic sys em, in he u u e, o
co e his gap [6].
In his amewo k, he cu en pape p esen s a no el
magne ic ac ua ion sys em o a mic obio a collec ion
capsule. The de ice, ee o any elec onical componen s, is
designed o allow a sa e mucosa b ushing and sample
s o age. In si u wi eless ac i a ion is enabled by he use o an
ex e nal pe manen magne , which igge s he o a ion o
minia u ized magne s inside he capsule. S a ing om he
sys em equi emen s, he pape shows he design low o he
ac ua ion sys em. In his con ex , a modula amewo k o
he compu a ion o he magne ic o ques and o ces enabling
he ac ua ion is p esen ed. Finally, p elimina y alida ion o
he design, h ough nume ical and analy ical me hods and
wi h bench op expe imen s, is epo ed.
II. REQUIREMENTS AND CAPSULE CONCEPT
A. Sys em equi emen s
As an inges ible de ice, he endoscopic capsule needs o
sa is y se e al limi ing equi emen s, especially in e ms o
size and sa e y. Mo eo e , addi ional cons ain s a e posed by
he need o collec ing and sa ely s o ing mic obio a samples.
He e, we lis he main equi emen s ha need o be
add essed:
• Capsule dimensions lowe o equal o hose o he
PillCamTM (Med onics, Minneapolis, USA), FDA
app o ed inges ible de ice [7] – leng h o 32 mm
and diame e o 12 mm.
Ma ina Finocchia o, C is ina Giosuè, Gaspa e D ago, Fabio Cibella, A ianna Menciassi, Membe
IEEE, Ma io Sp o ie i, Gas one Ciu i, Membe IEEE
Design o a magne ic ac ua ion sys em o a mic obio a-collec ion
inges ible capsule
• Localiza ion and wi eless ac i a ion o he sampling
p ocess in speci ic and selec ed loca ions.
• Sa e y o he de ice, a oiding p o uding pa s, loss o
componen s o elec omagne ic-associa ed isks.
• Volume o he collec ed sample su icien o RNA
sequencing analysis – equal o highe o 18 mg [8].
• App op ia e s o age o he collec ed sample inside
he capsule. In cu en clinical p ac ise, he
mic obio a sample is s o ed in a RNAla e 
s abiliza ion solu ion o a oid deg ada ion, igh
a e he collec ion (1:5 = olumes o RNA la e :
issue). Pa allel esea ches a e in es iga ing he need
o he p ese ing solu ion igh a e collec ion o
he sample, o i i can be added ew hou s la e ,
when he capsule is expulsed [9].
• Collec ion o a leas wo simul aneous samples o
egion o in e es o enable a mo e obus analysis.
B. Capsule concep
Gi en he small olume a ailable inside a pill-size
inges ible capsule (i.e., a ound 2 mL), he use o s anda d
mecha onic sys ems (i.e., mo o s, d i e s, elec onic boa ds,
ba e ies) poses se e al challenges. Mo eo e , sa e y issues
ela ed o he inges ion o elec ical componen s impose
addi ional limi a ions. As a solu ion, ou design is buil upon
an ac ua ion sys em made o only passi e componen s,
ope a ed emo ely by an ex e nal pe manen magne . As
shown in Fig. 1, he mechanism i s inside a 11.5 x 30.5 mm
capsule and allows he mic obio a sampling h ough luminal
b ushing. This clinical echnique p o ides he bes p opo ion
o bac e ia o hos DNA wi h espec o he o he sampling
me hods, wi hou in ol ing any bleeding [9]. In ou case, he
specimen is collec ed wi h he capsule using wo o a ing
b ushes (Fig. 1.d-e). Each componen has an ex e nal su ace
wi h compa able dimension (i.e., 50 mm²) o hose used in
clinics (i.e., Cook Medical Cy hology b ush). Each b ush,
once collec ed he issue, s o es i inside a sepa a e ese oi ,
which is sized o house he co ec olume o RNAla e 
(i.e., a leas 5 imes he olume o he sample). In ac , e en
hough he manda o y use o a p ese a ion solu ion is unde
in es iga ion, au ho s decided o gua an ee he needed space
inside he capsule. Two ga es co e he b ushes; hei
opening, oge he wi h he sampling p ocess, a e con olled
emo ely (Fig. 1.d).
Finally, i e N52 NdFeB in e nal pe manen magne s
(IPMs) cons i u es he capsule’s ac ua ion sys em (Fig. 1.c-e).
In pa icula , wo o hem ha e a hemisphe ical shape, and
a e ixed a he wo ends o he capsule. Whe eas wo
diame ically-magne ized hollow cylinde s a e moun ed on a
sha , and igidly connec ed o he b ushes, hus allowing
hei o a ion. Finally, a hi d hin hollow diame ically-
magne ized cylinde , ee o o a e a ound he capsule in e nal
wall, d i es he opening and closing o he ga es. A summa y
o he capsule’s speci ica ions is p o ided by Table 1.
The IPMs o a ions a e d i en by he a ia ion o he
magne ic ield induced by an ex e nal pe manen magne
(EPM). The sys em has h ee main wo king phases: (1)
a ac ion o he capsule o he in es inal walls; (2) opening o
he de ice; and (3) b ushing o he mucosa. As Fig. 2 shows,
when he EPM is placed a a dis ance d1 om he capsule, he
de ice is a ac ed o he mucosa by a o ce F. Once he EPM
is mo ed o a dis ance d2 smalle han d1, he gene a ed
magne ic ield o ces he o a ion o he ga es, exposing he
b ushes. Las ly, when he EPM eaches a dis ance d3 lowe
han d2, he o a ion o he cylind ical magne s is enabled,
d i ing he b ushing. In his case, he EPM is mo ed
backwa d and o wa d o a Δd a ound d3 (such ha d3-Δd <
d2), allowing mul iple o a ions o he b ushes. Finally, when
he sampling ask has ended, he EPM is emo ed leading o
he closu e and eleasing o he capsule.
III. MAGNETIC ACTUATION SYSTEM
A. Magne ic p inciple
The mechanism allowing he o a ion o he ga es and he
b ushes, is based on wo non-linea magne ic o sional
sp ings, i s ly in oduced by Simi e al. [10]. In ac , when
placing wo diame ically-magne ized hollow cylinde s on a
sha , a ixed sho dis ance, and ee o o a e, hey o ien
hemsel es o ha e pa allel and opposi e magne iza ion
Speci ica ions
Values
Capsule dimension
11.5 x 30.5 mm
B ush ex e nal su ace
50 mm² each
Rese oi size
261 mm3 each
Numbe o samples
2
IPMs dimensions [mm]
IPM1,5: R=5.25
IPM3: D=10, d=9, L= 7
IPM2,4: D=7, d=1.25, L=2
Figu e 1. 3D model o he endoscopic capsule: a) ex e nal iew wi h closed
ga es; b) he capsule wall is emo ed o see he in e nal componen s; c)
in e nal magne s a anged inside he capsule; d) he ga es (in g een), when
opened, exposes he b ushes (in g ey) o he ex e nal en i onmen ; e) 3D
CAD o he i e in e nal magne s and one o he wo b ushes.
Figu e 2. Ope a ional low: he EPM posi ion igge s i s he adhesion o
he capsule o he in es inal walls, second he opening o he ga es, and hi d
he b ushing o he mucosa.
TABLE 1. MAIN DESIGN SPECIFICATIONS
ec o s. Howe e , i only one o he wo magne s is ixed,
and an ex e nal magne ic ield wi h opposi e pola i y is
applied, s ong enough o o e come he local magne ic
coupling, he cylinde ee o o a e will make a 180° u n.
Doing so, i will each an uns able equilib ium poin , and will
s o e po en ial elas ic ene gy. Once he ex e nal magne ic
ield is emo ed, he ee magne will go back o i s ini ial
posi ion, eleasing he s o ed ene gy. The main pa ame e ha
cha ac e izes he magne ic o sional sp ing is he in e nal
o que be ween he wo magne s, when he angle be ween he
co esponding magne iza ion ec o s is 90°. This is he
highes le el o o que, named peak o que, ha needs o be
o e come by he ex e nal magne ic agen o o a e one o he
wo cylinde s.
B. Analy ical desc ip ion
In ou design, an ex ended e sion o he magne ic
o sional sp ing is used o d i e he sys em. Mo e in de ails,
e e ing o he nomencla u e o Fig. 3, IPMs 1 and 5 wo k as
he ixed magne s o h ee nes ed magne ic o sional sp ings
(MTS). Acco dingly, IPM3 is he ee- o- o a e elemen o he
i s MTS, allowing he opening o he ga es. Whe eas IPM2
and IPM4 a e he o a ional magne s o he o he wo MTSs.
In o de o co ec ly dimension he sys em o
gua an eeing he desi ed beha io , and c ea e a scalable
amewo k, a s a ic model has been de eloped. Assuming
ha he en i onmen does no con ain any addi ional
e omagne ic ma e ial, i is possible o apply he p inciple o
supe posi ion o he compu a ion o o ques and o ces
induced by mul iple magne s.
Fig. 3.a shows he o ien a ion o he magne s in he es
posi ion (i.e., ga es closed and capsule ee o na iga e in he
bowel). In his con igu a ion, he in e nal o que exe ed by
he magne s ixed o he capsule (i.e., IPM1,5) on hose ee o
o a e (i.e., IPM2,3,4), has o gua an ee he s able closu e o he
capsule:
 +    +  ()
 +    +  ()
Whe e, is he o que on IPM2 gi en by IPM1, is
he o que on IPM2 gi en by IPM5, e c. Since he sys em is
axially symme ic, he o que expe ienced by IPM2 is equal o
he one o IPM4, and he same applies o o ques on IPMs 1
and 5. Conside ing Fig. 3.a, as he i s s ep owa ds he
sampling p ocess, he EPM placed a a dis ance d1 om he
capsule, need o gene a e an axial magne ic ield g adien
such ha :
F (d) = F1 - 2F2 - F3 – mg > 0 ()
F1 = F5, F2 = F4. ()
Whe e F is he o al e ical o ce exe ed on he capsule,
F1, F2 and F3 e e espec i ely o he magne ic e ical o ce
on IPM1,2,3 and mg is he capsule weigh o ce o o e come
o ensu e he adhesion o he sys em o he in es inal mucosa.
Secondly, placing he EPM close o he capsule (Fig.
3.b), a a dis ance d2 < d1, he o que induced by he EPM on
IPM3 (i.e., 36), mus o e come he in e nal o ques be ween
IPM3 and he IPMs 1 and 5, hus allowing o open he ga e:
 (d)   −  ()
A his s age, he b ushes do no o a e ye , so he ex e nal
magne ic ield does no o e come he in e nal o ques on
IPMs 2 and 4:
 (d)   +  −  +  ()
Finally, mo ing he EPM a a dis ance d3 om he
capsule (such ha d3 < d2), he ex e nal o que on IPMs 2 and
4 o e come he in e nal couplings, hus allowing he o a ion
o hese wo magne s (Fig. 3.c), while keeping he ga es
opened:
 (d)   +  −  +  ()
IV. MAGNETIC MODELLING
A. Design wo k low
The dimensioning o he ac ua ion sys em has in ol ed
mul iple s eps in o de o mee he design equi emen s.
Conside ing he maximum olume o he capsule and lea ing
space o he wo ese oi s o RNAla e , he in e nal
magne s sizing is a ade-o be ween s abili y and ex e nal
magne ic ac i a ion. As a ma e o ac , he bigge he
in e nal o que on he IPM3, he mo e obus is he closu e o
he ga es. Howe e , i implies also ha ing a bigge ex e nal
magne o emo e ac i a ion. Hence, i s he IPMs we e
dimensioned and hei ela i e dis ances compu ed. Secondly,
he magne ic ield able o o e come he o ques on he IPMs
du ing he opening and b ushing phase was compu ed.
Finally, he EPM able o p o ide he compu ed ield was
selec ed, and he ope a ional dis ances om he capsule, o
each wo king phase, we e assessed. The op imal olume,
shape and ype o magne iza ion o he EPM we e
in es iga ed o p o ide an almos uni o m magne ic ield a
he capsule si e. To his end, wo models we e used o
sol ing he sys ems o equa ions: one analy ical and one
nume ical. The analy ical modelling is an implemen a ion o
he cha ge and cu en model o app oxima ing he magne ic
ield lux densi y, in e ac ion o ces and o que de i ing om
a magne ic sou ce [11]–[13]. The amewo k was de eloped
using Ma lab (Ma hwo ks). Whe eas he nume ical model
in ol es he solu ion o he magne os a ic equa ions using he
Fini e Elemen Me hod (FEM) wi h he so wa e COMSOL
Mul iphysics® (S ockholm, Sweden). Las ly, o a inal
alida ion o he modelling amewo k, bench op
expe imen s we e conduc ed wi h es magne s (Sec ion V).
Figu e 3. Diag am o he h ee ope a ional phases and he pola iza ion o he
IPMs: a) capsule con igu a ion in es posi ion, he EPM gene a e an
a ac ion o ce on he IPMs; b) he EPM mo es o wa d igge ing he
IPM3 o a ion ( i s uns able equilib ium poin ); c) he EPM mo es close ,
enabling he o a ion o IPM2 and IPM4 (second uns able equilib ium poin )
and he consequen b ushing.
.
B. In e nal o ques be ween IPMs
The dimensions o he in e nal magne s we e maximized,
o inc ease he o que gene a ed by he EPM on he IPMs. As
a ma e o ac , bigge magne s expe ience highe o ques
and o ces unde a ixed magne ic ield. The inal alues a e
epo ed in Table 1. All he magne s a e NdFeB wi h N52
magne iza ion g ade, he highes a ailable (i.e., B =1.45T).
As a second s ep, a se o simula ions was conduc ed o
in es iga e he op imal solu ion o loca ing he magne s
inside he capsule. To his end, he in e nal o ques, o he
h ee wo king phases we e compu ed a ying he leng h o
he capsule om 26.5 o 30.5 mm, wi h s eps o 1 mm. This
speci ic ange alls be ween he leng h o he Pillcam™ SB3
(i.e., 26.2 mm) and he Pillcam™ Colon (i.e., 32.3 mm),
lea ing ew mm om he maximum leng h o accoun o any
addi ion ex e nal co e s o shea hs ha migh inc ease he
inal size. Meanwhile, he su ace-su ace dis ance be ween
IPM1 and IPM2 was a ied om 0.5 mm o 4 mm wi h s eps
o 0.5 mm. The esul s a e shown in Fig. 4. As he bes ade-
o be ween s abili y (low in e nal o ques cause he capsule
o open oo easily, e en when no desi ed) and emo e
ac i a ion, he capsule dimensions we e es ablished, in he
i s place, o 28.5 mm, and maximum dis ance be ween
IPMs 1 and 2 (i.e., d =3 mm). Acco dingly, he o que needed
o o a e he IPM3 du ing he opening phase is 3.7 mNm, and
o o a e IPMs 2 and 4 du ing he b ushing is 6.05 mNm
(solu ion compu ed analy ically). These alues a e p o ided
by an ex e nal uni o m magne ic luid densi y o a leas
B=30 mT (opening) and B=75 mT (b ushing), as esul ing
om a second se o simula ions.
C. Ex e nal o ques induced by an EPM on he IPMs
The i s s ep owa ds he selec ion o he EPM was
choosing i s op imal shape and he ype o magne iza ion, a a
ixed olume, inducing he highes o ques on IPM2 and
IPM3. Following, he olume can be scaled o allow he
p ope igge ing o he opening and b ushing phase, a
dis ances compa ible wi h he ana omy o he pa ien . As a
esul o an ini ial pilo expe imen a ion, an EPM olume o
a ound 1000 cm3 is needed o ac i a e IPMs 2 and 4 a a
dis ance o 100 mm, which is he usual ange used o
endoscopic magne ic locomo ion [10]. Hence, ixing he
olume o he EPM o 1000 cm3, he o que on IPM2 induced
by se e al cylinde s wi h axial and diame ical
magne iza ion, and cubes, wi h di e en sizes a io we e
compu ed (EPMs B =1.45T), a a iable dis ances. The
esul s a e shown in Fig. 5, and highligh ha he highes
o que is p o ided by he axially-magne ized cylinde wi h
dimensions: D=15 cm L=5,6 cm.
Wi h hese dimensions, he EPM o e come he IPM2 peak
o que a a dis ance o 100 mm om he capsule (su ace- o-
su ace) and he IPM3 peak o que a 160 mm. Hence,
e e ing o Fig. 3, d2 could be se o 160 mm (opening), and
d3 o 100 mm (b ushing). These alues all in he ange
usually conside ed as he a e age dis ance be ween an
endoscopic capsule and he skin, o people wi h a no mal
body mass index (BMI) [10]. Howe e , gi ing he scalabili y
o he amewo k and he lack o any es ic ion on he size o
he ex e nal ac ua o , he EPM olume migh be inc eased in
he u u e o be compa ible wi h pa ien s ha ing highe BMI.
V. EXPERIMENTAL VALIDATION
To assess he magne ic models and e i y he p edic ed
o ques, a se ies o ou bench op expe imen s we e
conduc ed (named A, B, C, D). Since capsule’s selec ed
magne s equi e a cus om p oduc ion, high cos and long
deli e y imes, easy- o-acqui e and comme cially-a ailable
magne ic cylinde s ha e been used o es ing he model.
Each es equi ed he use o a obo ic a m (RV-3SB,
Mi subishi Co p., Tokyo, Japan) o high p ecision egula ion
o he dis ances. A iaxial o ce/ o que senso , (Nano 17-E,
ATI Indus ial Au oma ion, U.S.) was used in es s A and B
o measu ing he o ques. Each ime, he i e eadings om
he senso we e acqui ed, and i s a e age alue was used as a
e e ence. Fo all he ou expe imen s, he suppo s we e
ei he 3D p in ed in ABS, o manu ac u ed wi h Te lon® and
aluminum o a oid magne ic in e e ences. The sc ews used
o ix he suppo s and he sha we e non- e omagne ic (i.e.,
Figu e 5. To ques induced on IPM2 by di e en EPM wi h ixed olume o
1000 cm3, a a iable dis ances IPM-EPM ( esul s o he analy ical model).
The EPMs in es iga ed a e axially magne ized cylinde s, diame ically
magne ize cylinde s, and cubes. The ed line shows he peak o que o
IPM2, which when eached, allow o o a e he magne (b ushing).
Figu e 4. Peak o que cu es o IPM2 and IPM3 a ying he leng h o he
capsule (L [mm]) and he dis ance be ween IPM1 and IPM2 ( esul s o he
analy ical model). These alues a e hose ha need o be o e come by he
EPM o open he ga e (3) and b ush he mucosa (2).
aluminum, Te lon®). Each es was epea ed i e imes, and
he esul s ob ained we e compa ed wi h he p edic ed alues
o bo h he analy ical and nume ical models.
A. Tes A
A i s s ep in assessing he model was o alida e he
es ima ed o que be ween wo concen ic cylind ical magne s,
a ying hei ela i e dis ance and angle. This con igu a ion
is he one used o es ima ing he in e nal o ques be ween he
IPMs. To his end, wo iden ical hollow cylinde s wi h
diame ical magne iza ion (NdFeB 35H, B =1.17 T, L=6 mm,
D=8 mm and d=2 mm) we e used. As shown in Fig. 6.a, one
magne was ixed o he bench op, while he o he one was
a ached o he load cell h ough a connec o made in
aluminum, long enough o a oid any in e e ences o he
senso wi h he magne ic ield. The load cell was hen
a ached o he obo ic a m, o con ol he magne posi ion. In
a i s es (Tes A.1), he wo cylinde s we e placed one in
on o he o he , wi h a 90° misma ch be ween hei
pola iza ion ec o s, and a su ace- o-su ace dis ance
anging om 5 o 57 mm, conside ing 1 mm s eps.
In a second es (Tes A.2), he dis ance be ween he
cylinde s was ixed o 10 mm, and he magne a ached o he
obo was o a ed om 0 o 360°, wi h s eps o 30° each. In
bo h cases, o each s ep, he axial o que was eco ded and
no ed.
B. Tes B
The second expe imen al se -up was designed o alida e
he o que es ima ed be ween one IPM and he EPM. As in
he p e ious expe imen , one diame ically magne ized
cylinde (NdFeB 35H, B =1.17 T, L=6 mm, D=8 mm and
d=2 mm) was a ached o a load cell, connec ed o he obo ic
a m. Howe e , in his case, a bigge cylind ical magne
(NdFeB N35, B =1.35 T, D=50 mm and L=25 mm), wi h
axial magne iza ion, was ixed o he bench op. In he i s
es (Tes B.1), he magne a ached o he obo ic a m was
p og essi ely mo ed owa ds he ixed magne om a
dis ance o 34 o 12 mm, wi h 1 mm s eps. Meanwhile, he
angle be ween he wo magne iza ion ec o s was ixed o
90° as shown in Fig. 6.b. Whe eas, in he second es (Tes
B.2), he dis ance su ace- o-su ace be ween he wo
componen s was ixed o 50 mm, and he IPM was o a ed
a ound i s main axis om an angle o 0° o 180°, wi h s eps
o 30°. In bo h cases, a each s ep, he o que was eco ded by
he load cell and no ed down.
C. Tes C
The hi d se o expe imen s aimed a alida ing he
magne ic o sional sp ing model, and he es ima ed
ope a ional dis ance o he EPM. In his case, he wo
cylind ical magne s used in he i s es (i.e., B =1.17 T, L= 6
mm, D=8 mm and d=2 mm) we e moun ed on a sha , which
was ixed h ough a suppo o he bench op (Fig. 6.c). One
magne was ixed, while he o he ( ee o o a e), was placed
a a su ace- o-su ace dis ance o 6.5 mm om he i s one.
A disk made o a low ic ion ma e ial (i.e., Te lon®) was
used as a space o he magne s. The p e iously in oduced
axially-magne ized cylinde , was a ached o he obo ic a m
using a non- e omagne ic connec o (B =1.35 T, D=50 mm
and L=25 mm). S a ing om a dis ance o 100 mm om he
IPMs, he EPM was p og essi ely mo ed o wa d. The
dis ance IPM-EPM a which he magne ee o o a e made a
u n o 180° was no ed.
D. Tes D
Finally, he las expe imen was designed o es he
whole sys em. In his case, i e di e en diame ically
magne ized hollow cylinde s, esembling hose sized o he
capsule, we e moun ed on a suppo . The mos ex e nal
magne s (in Fig. 6.d IPMs 1 and 5) we e ixed (NdFeB 35H,
B =1.17 T, L=6 mm, D=8 mm and d=2 mm), as in he
endoscopic capsule. Two addi ional hollow cylinde s
(Sm2Co17 YXG30, B =1.08 T, L=2 mm, D=8 mm, d=2
mm), ep oducing IPMs 2 and 4, we e spaced (su ace- o-
su ace) 2 mm apa , om he closes ixed magne , and le
ee o o a e. Finally, a hin cylinde (Sm2Co17 YXG30,
B =1.08 T, L=4 mm, D=12 mm and d=8.5 mm) was moun ed
on a cylind ical suppo , ixed o he sha . This magne ,
mimicking IPM3, was le ee o o a e, and was spaced 2 mm
espec i ely om magne s 2 and 4 (see Fig. 6.d). All he
space and he suppo we e buil in Te lon®, o minimize he
con ac ic ion. The p e iously men ioned axially-
magne ized cylinde (B =1.35 T, D=50 mm and L=25 mm),
was a ached o he obo ic a m, wi h i s su ace cen e ing he
magne s moun ed on he suppo . The cylinde was mo ed
om a dis ance o 130 mm owa ds he IPMs. The poin s in
which magne s 2 and 4 and magne 3 u ned was eco ded.
E. Resul s
The aim o he expe imen s was o alida e he modelled
o ques, o ces and EPM ope a ional dis ances. The esul s
a e summa ized in Table 2. To his end, Tes A and B
Figu e 6. a) Expe imen al se up o alida e he in e nal o que be ween wo IPMs a ying hei ela i e dis ance d and he angle θ; b) expe imen al se up
o alida e he o que induced by he EPM on a IPM a ying hei ela i e dis ance d and he angle θ; c) expe imen al se up o alida e he dis ance a
which he EPM igge s he magne ic o sional sp ing, and allows he o a ion o an IPM; d) expe imen al se up o alida e he dis ances a which he
EPM igge s he o a ion o IPM3 (opening) and IPM2,4 (b ushing).

p o ided good esul s, ha ing he expe imen al o que cu es
o e lapping wi h he es ima ed ones (see Fig. 7); o Tes
A.2, FEM ou pe o med he analy ical solu ion. Rega ding
Tes C a sligh misma ch be ween simula ed and eal esul s
was epo ed. This is p obably due o he non-pe ec
alignmen be ween he EPM and IPM cen e , which induces
addi ional o ques ha con ibu e o he o a ion o he IPM
(happening a a sligh u he dis ance IPM-EPM han he
simula ed one). Finally, in Tes D, bo h IPM3 and IPM2,4
o a ed when he EPM was close o he es ima ed ope a ional
dis ance. The eason o his obse ed beha io comes om
an addi ional esis i e ac o which con ibu e o he o e all
axial o que o he IPM. As a ma e o ac , while o a ing
unde he in luence o he EPM, he IPMs expe ience in e nal
a ac i e/ esis i e magne ic o ces and sligh o ques on he
o he wo axes. These, oge he wi h he con ac ic ion,
inc ease he o e all mechanical o que on IPM3 o a 30%.
Whe eas, o IPM2,4, i doubles he esis i e peak o que.
VI. DISCUSSION AND CONCLUSION
In his pape , we p esen a no el magne ic ac ua ion
sys em o an inges ible capsule designed o enabling he
collec ion o mic obio a samples h oughou he whole GI
ac . The sys em allows in si u sampling o he mic obio a,
and emo e ac i a ion wi h an ex e nal pe manen magne .
Localiza ion o he sys em is possible, pe o ming
iangula ion o he magne ic lux densi y measu ed wi h
ex e nal Hall e ec senso s (e.g., asking he pa ien o wea a
bel wi h hall e ec senso s) [14]. Howe e , i s de elopmen
is ou o he scope o his pape . The passi e na u e o he
magne s, eplacing s anda d elec ical componen s,
in insically makes he sys em sa e and easie o ansla e in
he clinical p ac ice. The ac ua ion sys em, based on mul iple
magne ic o sional sp ings, was dimensioned h ough a
scalable amewo k, based on nume ical and analy ical
simula ions. The alidi y o he compu a ional model and he
unc ions o he ac ua ion sys em we e e alua ed
expe imen ally wi h a se ies o bench op es s. Acco dingly,
a scaled se -up assessed he alidi y o he compu ed o ques
and e i ied he o e all ope a ions o he sys em. To his
end, he expe imen al pla o m included a se o pe manen
magne s wi h sligh bigge size o hose designed o he
inal p o o ype. The es s allowed o es ima e a esis i e
mechanical ac o which con ibu es o he o que on he
IPMs o a second- ound design and inal implemen a ion.
As a inal s ep, conside ing he disco e y o he
co ec i e pa ame e h ough he expe imen s, he p o o ype
design is modi ied. Acco dingly, he inal capsule leng h is
se o 30.5 mm. Doing so, he pu e magne ic peak o que on
IMP2,3,4 is dec eased. Following, he size o he EPM is
inc eased o L=73 mm and D=195 mm: o allow he
ac i a ion despi e he addi ional esis i e con ibu ion on he
IPMs o ques. In his way he new EPM ope a ional
dis ances a e d2=200 mm (i.e., opening) and d3=100 mm
(i.e., b ushing), s ill compa ible wi h ope a ing dis ances. As
a ma e o ac , he inal dimensioning is made in o de o
keep he minimum dis ance capsule (IPMs)-EPM (i.e., d3) o
100 mm, which is a easonable es ima ion o he a e age
dis ance be ween he skin and he bowel [10]. In his way,
he EPM dimensions we e modi ied main aining he shape,
ype o magne iza ion, and he dimensional a io be ween he
diame e and he heigh , se in Sec ion IV and conside ed
op imal. Wi h a inal olume o 2179 cm3, and a weigh o
16.3 kg, he EPM migh be ixed o a suppo i e a m (e.g.,
Ma in’s A m), and manipula ed by hands om he clinician.
The design pa ame e s, de i ed om he alida ed
compu a ional models and he bench op expe imen s, will
enable he au ho s o build he inal capsule p o o ype.
Hence, all he componen s will be moun ed inside a bio-
compa ible case and a se o u u e expe imen s will aim a
assessing he abili y o he designed de ice o collec and
s o e mic obio a samples, in ex- i o and in- i o scena ios.
Fo eseen challenges include ensu ing he sealing o he
capsule, o a oid any con amina ion o he samples, bu
keeping he ic ion o ces low. Possible solu ions migh
in ol e a combina ion o di e en echnologies, e.g., silicon-
based plugs, expanding hyd ogels, eccen ic o a ional
mechanism and ex e nal en e ic co e ings.
TABLE 2. SUMMARY OF RESULTS OF TEST A AND TEST B
Tes
Es ima ed o que e o (Mean ± SD)
Analy ical s
expe imen al da a
Nume ical s
expe imen al da a
A.1
0.02±0.02 mNm
0.05±0.04 mNm
A.2
0.1 ± 0.04 mNm
0.08±0.02 mNm
B.1
0.13±0.05 mNm
0.21±0.12 mNm
B.2
0.02±0.02 mNm
0.03±0.01 mNm
Es ima ed dis ance
Expe imen al
dis ance
(Mean ± SD)
Analy ical model
Nume ical model
C
77 mm (Δ=5 mm)
76 mm (Δ=6 mm)
82±1.4 mm
D
Open: 105 mm
B ush: 32 mm
Open: 110 mm
B ush: 32 mm
Open: 95 ±4.5 mm
B ush: 18 ± 3 mm
Figu e 7. To que cu es ob ained om es s A and B; compa ison o expe imen al esul s, wi h he analy ical and nume ical models.
REFERENCES
[1] E. N. Jano , “The mic obiome and human disease
pa hogenesis: how do you do wha you do o me
…?,” T ansla ional Resea ch, ol. 179. Mosby Inc.,
pp. 1–6, Jan. 01, 2017, doi:
10.1016/j. sl.2016.10.007.
[2] L. K. U sell, J. L. Me cal , L. W. Pa ey, and R.
Knigh , “De ining he human mic obiome,” Nu .
Re ., ol. 70, no. SUPPL. 1, p. S38, Aug. 2012, doi:
10.1111/j.1753-4887.2012.00493.x.
[3] E. Thu sby and N. Juge, “In oduc ion o he human
gu mic obio a,” Biochemical Jou nal, ol. 474, no.
11. Po land P ess L d, pp. 1823–1836, Jun. 01,
2017, doi: 10.1042/BCJ20160510.
[4] G. Ciu i e al., “F on ie s o obo ic endoscopic
capsules: a e iew,” Jou nal o Mic o-Bio Robo ics,
ol. 11, no. 1–4. Sp inge Ve lag, Jun. 01, 2016, doi:
10.1007/s12213-016-0087-x.
[5] M. Ben Salem, G. Aiche, L. Rubbe , P. Renaud, and
Y. Haddab, “Design o a Mic obio a Sampling
Capsule using 3D-P in ed Bis able Mechanism,” in
P oceedings o he Annual In e na ional Con e ence
o he IEEE Enginee ing in Medicine and Biology
Socie y, EMBS, Oc . 2018, ol. 2018-July, pp. 4868–
4871, doi: 10.1109/EMBC.2018.8513141.
[6] H. Rezaei Nejad e al., “Inges ible Osmo ic Pill o
In Vi o Sampling o Gu Mic obiomes,” Ad . In ell.
Sys ., ol. 1, no. 5, p. 1900053, Sep. 2019, doi:
10.1002/aisy.201900053.
[7] “PillCamTM SB 3 Sys em | Med onic.”
h ps://www.med onic.com/co idien/en-
us/p oduc s/capsule-endoscopy/pillcam-sb-3-
sys em.h ml (accessed Oc . 18, 2020).
[8] G. Li e al., “Di e si y o duodenal and ec al
mic obio a in biopsy issues and luminal con en s in
heal hy olun ee s,” J. Mic obiol. Bio echnol., ol.
25, no. 7, pp. 1136–1145, 2015, doi:
10.4014/jmb.1412.12047.
[9] M. J. Claesson, A. G. Clooney, and P. W. O’Toole,
“A clinician’s guide o mic obiome analysis,”
Na u e Re iews Gas oen e ology and Hepa ology,
ol. 14, no. 10. Na u e Publishing G oup, pp. 585–
595, Oc . 01, 2017, doi: 10.1038/n gas o.2017.97.
[10] M. Simi, G. Ge boni, A. Menciassi, and P. Valdas i,
“Magne ic o sion sp ing mechanism o a wi eless
biopsy capsule,” J. Med. De ices, T ans. ASME, ol.
7, no. 4, Sep. 2013, doi: 10.1115/1.4025185.
[11] Pe manen Magne and Elec omechanical De ices.
Else ie , 2001.
[12] E. P. Fu lani, S. Reznik, and W. Janson, “A Th ee-
Dimensional Field Solu ion o Bipola Cylinde s,”
IEEE T ans. Magn., ol. 30, no. 5, pp. 2916–2919,
1994, doi: 10.1109/20.312547.
[13] J. Li e al., “Magne ically-d i en medical obo s: An
analy ical magne ic model o endoscopic capsules
design,” J. Magn. Magn. Ma e ., ol. 452, pp. 278–
287, Ap . 2018, doi: 10.1016/j.jmmm.2017.12.085.
[14] F. Bianchi e al., “Expe Re iew o Medical
De ices Localiza ion s a egies o obo ic
endoscopic capsules: a e iew Localiza ion
s a egies o obo ic endoscopic capsules: a
e iew,” 2019, doi:
10.1080/17434440.2019.1608182.