170 IEEE/ASME TRANSACTIONS ON MECHATRONICS, VOL. 15, NO. 2, APRIL 2010
Design, Fabrication, and Testing of a Capsule With
Hybrid Locomotion for Gastrointestinal
Tract Exploration
Massimiliano Simi, Pietro Valdastri, Member, IEEE, Claudio Quaglia, Member, IEEE,
Arianna Menciassi, Member, IEEE, and Paolo Dario, Fellow, IEEE
Abstract—This paper describes a novel solution for the active lo-
comotion of a miniaturized endoscopic capsule in the gastrointesti-
nal (GI) tract. The authors present the design, development, and
testing of a wireless endocapsule with hybrid locomotion, where
hybrid locomotion is defined as the combination between internal
actuation mechanisms and external magnetic dragging. The cap-
sule incorporates an internal actuating legged mechanism, which
modifies the capsule profile, and small permanent magnets, which
interact with an external magnetic field, thus imparting a dragging
motion to the device. The legged mechanism is actuated whenever
the capsule gets lodged in collapsed areas of the GI tract. This
allows modification of the capsule profile and enables magnetic
dragging to become feasible and effective once again. A key com-
ponent of the endoscopic pill is the internal mechanism, endowed
with a miniaturized brushless motor and featuring compact design,
and adequate mechanical performance. The internal mechanism is
able to generate a substantial force, which allows the legs to open
against the intestinal tissue that has collapsed around the capsule
body. An accurate simulation of the performance of the minia-
turized motor under magnetic fields was carried out in order to
define the best configuration of the internal permanent magnets
(which are located very close to the motor) and the best tradeoff
operating distance for the external magnet, which is responsible
for magnetically dragging the capsule. Finally, a hybrid capsule
was developed generating 3.8 N at the tip of the legged mechanism
and a magnetic link force up to 135 mN. The hybrid capsule and its
wireless control were extensively tested in vitro, ex vivo, and in vivo,
thus confirming fulfilment of the design specifications and demon-
strating a good ability to manage collapsed areas of the intestinal
tract.
Index Terms—Capsule endoscopy, endoscopic capsule, magnetic
locomotion, robotic surgery.
I. INTRODUCTION
CURRENT technologies in biomedical engineering focuson reducing pain during diagnostic procedures, thus en-
abling extensive screening among individuals who are more
likely to develop cancer. This has yet to be achieved for the
Manuscript received June 25, 2009; revised November 30, 2009. First
published February 25, 2010; current version published March 31, 2010.
Recommended by Guest Editor J. P. Desai. This work was supported in part by
the Intelligent Microsystem Center, Korean Institute of Science and Technol-
ogy, South Korea, and in part by the European Commission in the framework
of VECTOR FP6 under European Project EU/IST-2006-033970.
The authors are with the Center for Research in Microengineering Labora-
tory, Scuola Superiore Sant’Anna, 56127 Pisa, Italy (e-mail: m.simi@sssup.it;
pietro@sssup.it; c.quaglia@crim.sssup.it; arianna@sssup.it; dario@sssup.it).
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TMECH.2010.2041244
inspection of the lower gastrointestinal (LGI) tract, where long
and semiflexible video probes are introduced into the oral or
rectal cavities. This approach provides a detailed and reliable
diagnosis and currently represents the best available technique
for managing pathological diseases of the digestive tract. The
major drawback is the procedure’s invasiveness, making it fre-
quently ill-tolerated by patients.
In order to reduce invasiveness, the use of wireless capsule
endoscopy for screening procedures [1]–[3] has increased over
recent years. This approach has proved to be mostly effective
for the small bowel, with different endoscopic capsules already
available in the market from several suppliers worldwide, such
as Given Imaging [4], Olympus [5], and Intromedic [6]. These
devices are mainly composed of a 640 × 480 pixel resolution
camera, a LED-based illumination system, a data transmitter,
and a suitable battery, which guarantees sufficient operative life-
time. All these components are incorporated in a pill with size
ranging from the Intromedic MiRo (a diameter (φ) of 10.8 mm
and length (L) of 24 mm), to the Given Imaging PillCam Colon
(φ = 11 mm, L = 31 mm). Despite capsule endoscopes are
well-tolerated by patients, the diagnostic outcomes are not fully
reliable. This is mainly due to the lack of active and teleoper-
ated locomotion, because the pills move by simply exploiting
GI peristalsis. Consequently they cannot stop, turn, or change
direction to review and verify suspicious lesions, as in the case
of traditional endoscopes. Furthermore, without proper tissue
distension, obtained through air insufflation (which is standard
practice today), interesting spots behind intestinal folds could
be missed.
The ideal capsule for GI tract access should combine small
size (small enough to swallow and to move by natural peri-
stalsis) and active locomotion (ability to stop, and possibly, to
move against peristalsis), and not require inflation of the colon,
which is the source of much of the pain experienced during the
traditional procedure [7]. If LGI endoscopy were to become a
screening procedure due to such a system, this would signif-
icantly enhance public health and improve cancer screening.
With regard to this, it is important to point out that yearly death
rates from cancer dropped by 27% from 1975 to 2005, largely
due to prevention and early diagnosis [8].
The development of a wireless capsule that is capable of
moving actively inside the GI tract while teleoperated by the
endoscopist is a challenging task, given the features peculiar
to the intestinal environment: unstructured, highly deformable,
1083-4435/$26.00 © 2010 IEEE
Authorized licensed use limited to: UNIVERSITA PISA S ANNA. Downloaded on April 27,2010 at 16:59:04 UTC from IEEE Xplore.  Restrictions apply. 
SIMI et al.: DESIGN, FABRICATION, AND TESTING OF A CAPSULE WITH HYBRID LOCOMOTION FOR GI TRACT EXPLORATION 171
slippery (i.e., covered by a layer of lubricating mucus), and
partially collapsed [9], [10]. Recent works addressing this issue,
propose solutions where the source of active locomotion can be
either external or internal to the patient’s body.
External active locomotion can be obtained by using magnets
inside the capsule coupled with an external magnetic field, gen-
erated either by coils [5], [11] or permanent magnets [12]–[15].
Controlled variations of the external magnetic field drag and ori-
ent the capsule in the desired direction. This approach does not
require supply from the on-board battery, thus saving available
energy for other different functions. The most critical limitation
of this approach is that magnetic locomotion is successful only
if the tissue is distended, instead, in the collapsed regions of the
LGI tract, friction prevents effective control. Moreover, this ap-
proach does not allow per se tissue distension, thus preventing
a complete view of the lumen walls.
The alternative approach consists in placing one or more
actuators inside the video pill in order to implement an inter-
nal locomotion system. By correctly designing the locomotion
mechanism, an effective relative motion between the capsule
and the surrounding tissue can be achieved. Several devices
with embedded actuators and mechanisms have been proposed,
providing satisfactory locomotion performance inside hollow
cavities [16]–[20]. In particular, studies on legged capsules for
LGI endoscopy have demonstrated that two independent leg sets
promote effective motion in the gut [21], and that the maximiza-
tion of the number of legs could enhance tissue distension in
collapsed cavities [22]. The main drawback of this approach is
related to power supply and space constraints. The continuous
use of a couple of actuators designed for moving two indepen-
dent leg sets requires an energy source that may be hardly shrunk
down to a swallowable size.
By combining external magnetic fields—which drag and ori-
ent the capsule—with a simplified internal active system—
which creates space—a hybrid solution for controlled capsule
endoscopy can be achieved. This technique would benefit from
the specific advantages of both approaches and enable effective
active locomotion in a swallowable size. The capsule described
in this paper is a step toward this goal: it features a set of per-
manent magnets for external control in uncollapsed areas of the
GI tract and a leg-based mechanism to be used as soon as the
capsule gets stuck along its journey. This single leg set can ad-
ditionally be exploited to lightly distend surrounding tissue and
so obtain a reliable diagnostic outcome. Hybrid locomotion in
capsule endoscopy is therefore firstly defined in this paper as
the combination and synergy of magnetic and legged strategies
for active control and locomotion in the LGI tract.
An overall description of the novel device is reported in
Section II-A, followed by a detailed overview of the dimension-
ing of the internal (legged) and external (magnetic) locomotion
mechanisms in Sections II-B and II-C, respectively. Control unit
and power supply are discussed in Section II-D, while packag-
ing issues are outlined in Section II-E. Section III describes the
experimental assessment of the capsule, conducted to confirm
technical requirements and performance of hybrid locomotion.
Finally, conclusions and future improvements of the system are
illustrated in Section IV.
II. HYBRID CAPSULE DESIGN
A. System Overview
The patient undergoing the devised procedure is requested
to enter a purposely equipped ambulatory environment once
the capsule has reached the LGI tract. Then, the endoscopist
employs magnetic locomotion as much as possible [12] and
guides the capsule using real-time vision feedback. Since natural
gases are frequently present in the LGI tract, most of the capsule
journey may occur by simple magnetic dragging, relying on the
low dynamical friction coefficient of intestinal tissue, which
is lubricated by a thick mucus layer [23]. If the capsule gets
stuck in collapsed areas of the intestine, the on-board legged
mechanism is activated and magnetic locomotion can take over
again. Specifically, leg activation pushes the capsule outside a
tissue fold if the capsule is entrapped. Alternatively, by keeping
the legs open, the capsule can be driven by an oscillatory motion
due to the magnetic link. This provides a lever-like action, which
expands the tissue and allows the capsule to overcome sharp
bends. Finally, the internal mechanism may also help dilate
tissue and obtain a better view of intestinal walls. Fig. 1 shows
an example of hybrid locomotion.
Starting from this operative scenario, it is possible to outline
the desirable features that the hybrid capsule should have as
follows.
1) Small size: The capsule should be small enough to swal-
low. However, since the hybrid locomotion concept still
needs to be extensively validated, scalability of the pro-
totype down to a swallowable size was considered ac-
ceptable in this paper. Thus, a diameter of 14 mm was
assumed for this device, with a target of 11 mm for future
developments.
2) Controllability by external magnetic fields in uncollapsed
regions: Considering that the external source of the mag-
netic field is at an average operative distance of 10 cm
from the capsule, the attraction force should be adequate
to lift the capsule against its own weight. If this is satisfied,
we can reasonably assume that the external magnetic field
would also be able to drag the pill inside uncollapsed or
lightly collapsed intestine.
3) Ability to expand surrounding tissues in collapsed regions:
Previous studies have demonstrated that about 3 N of total
force at leg tips are required to lift and stretch out collapsed
tissue by a legged mechanism [10], [22].
4) Real-time vision system: Since the devised procedure is
based on a real-time visual feedback from the capsule, a
wireless vision module with adequate frame rate needed to
be integrated. Despite research in this field [24], no minia-
turized real-time vision system was available at the time
of the work, therefore an estimated volume (450 mm3 as
the ongoing efforts [25] toward a wireless vision module)
for this module was left available in the hybrid capsule for
future integration.
5) User-friendly control interface: Since capsule steering is
obtained by applying two synergetic approaches, the en-
doscopist must be able to switch between the two easily
and intuitively.
Authorized licensed use limited to: UNIVERSITA PISA S ANNA. Downloaded on April 27,2010 at 16:59:04 UTC from IEEE Xplore.  Restrictions apply. 
172 IEEE/ASME TRANSACTIONS ON MECHATRONICS, VOL. 15, NO. 2, APRIL 2010
Fig. 1. (a) Legs are closed inside the body and the external magnet moves
the capsule in lightly collapsed region. Collapsed tissue, corresponding to the
dashed line, stops the magnetically driven motion. (b) External magnet only
allows capsule orientation, and does not provide net locomotion. (c) Activation
of the legged mechanism lifts the tissue and moves the capsule slightly forward.
Magnetic locomotion starts again after removing the tissue around the capsule.
The oscillating motion of the external magnet in combination with translation
can help to overcome the sharp bends of the intestine.
B. Internal Mechanism
1) Legged System Design: The internal legged mechanism
must be activated whenever the capsule is lodged within col-
lapsed tissue or more lumen distension is required. To achieve
this goal, a viable actuator was first selected, then the mecha-
nism was designed and its parts were dimensioned considering
the aforementioned requirements.
Due to previous extensive analysis regarding actuators for
legged locomotion in capsule endoscopy [22], an electromag-
netic brushless dc motor (SBL04-0829, Namiki Precision Jewel
Company Ltd., Tokyo, Japan) was selected as the best commer-
cially available tradeoff between output torque and compact-
ness. The selected actuator had an external diameter of 4 mm, a
total length of 17.4 mm (not including the shaft), and a nominal
Fig. 2. (a) Schematic view of the hybrid capsule internal mechanism. Gears
1 and 2 do not have teeth in the picture for the sake of clarity. (b) Cross section
view of the locomotion unit.
output stall torque of 5.7 mNm [26]. Being an electromagnetic
motor, permanent magnets placed close to the rotor or exter-
nally generated intense magnetic fields reduce effective output
torque. As reported in the following sections, this issue was
deeply analyzed for a proper dimensioning of the magnetic sys-
tem. Alternative actuators, such as piezoelectric motors were
also considered. Despite their immunity to magnetic fields, those
commercially available do not guarantee a high torque in a small
size, as desirable for this application.
As regards the mechanism, the use of three legs placed ra-
dially at 120◦, was considered adequate for positioning both
mechanism components and magnets, and to obtain an effec-
tive tissue engagement. The pivoting point of the three legs was
placed in the middle of the locomotion unit, thus obtaining a
symmetrical and reversible 180◦ movement. Each leg snapped
into a rectangular groove obtained on the lateral surface of a
brass shaft, which was connected to a bronze helical gear [part
4 in Fig. 2(a)] by means of a custom-shaped groove.
A steel worm gear [part 3 in Fig. 2(a)] placed on the central
axis of the locomotion unit matched the three helical gears,
whereas a brass main gear [part 2 in Fig. 2(a)] transmitted the
motion to the worm gear due to a long (5.6 mm) axial shaft.
Finally, a brass pinion gear [part 1 in Fig. 2(a)] connected the
mechanism to the dc brushless motor. As visible in the cross
section view of Fig. 2(b), the mechanism design divides the
capsule in three longitudinal sections. One section hosts the dc
brushless motor, while the other two are left available for the
on-board magnets.
2) Legged System Dimensioning: In order to define the de-
sign parameters for the different gears of the mechanism, several
issues were considered, such as dc brushless motor features, cap-
sule size, and ease of fabrication. A high reduction ratio is a key
point for this application, because it allows the force available
at the leg tip to be increased.
The relation between output (Mout) and input (Mmot) torques
as function of total mechanism efficiency (ηtot) and total
Authorized licensed use limited to: UNIVERSITA PISA S ANNA. Downloaded on April 27,2010 at 16:59:04 UTC from IEEE Xplore.  Restrictions apply. 
SIMI et al.: DESIGN, FABRICATION, AND TESTING OF A CAPSULE WITH HYBRID LOCOMOTION FOR GI TRACT EXPLORATION 173
transmission ratio (τtot) was as follows:
Mout =
Mmotηtot
τtot
. (1)
Considering the hybrid mechanism model, Mout was the
torque of both legs and respective helical gears, Mmot was
the torque transmitted by the motor to the mechanism, τtot
was the total transmission ratio of the hybrid mechanism gears,
whereas ηtot was the total efficiency.
Considering the kinematic chain, τtot and ηtot can be written
as follows:
τtot = τ12τ23τ34 (2)
ηtot = η12η23η34 (3)
where subscripts 1, 2, 3, 4 indicate the gears as in Fig. 2(a). The
output torque can also be written as follows:
Mout = Flegsb (4)
where b was the lever arm defined by the leg length and Flegs
was the total force at the leg tips.
The motor torque can be defined as follows:
Mmot = 0.5Cmotγmot (5)
where Cmot was the stall torque for the selected motor (Cmot =
6.95 mN·m, when powered with 3.7 V). In order to introduce
a safety margin for the mechanism design, we adopted the half
of this value as the operative motor torque. As previously men-
tioned, permanent magnets placed close to an electromagnetic
motor reduce performance of the actuator. To consider this, we
introduced an experimental coefficient (γmot) that related the
loss of torque available for moving the mechanism to the mag-
netic field surrounding the motor.
From (1) and considering (2)–(5), the hybrid three-legged
mechanism can be described by
Flegsb =
(0.5Cmotγmot)(η12η23η34)
τ12τ23τ34
⇒
Flegs =
(0.5Cmotγmot)(η12η23η34)
(τ12τ23τ34)b
≥ 3 N (6)
for an optimal management of collapsed tissue.
On the basis of (6) and the 4-leg capsule described in [16],
the mechanism gears were designed to obtain the best tradeoff
between maximum reduction ratio, high efficiency, small size,
reliability, fabrication easiness, and costs, according to guide-
lines reported in [27] and [28]. The main features of gears and
mechanism are summarized in Tables I and II, respectively.
The mechanical transmission between worm and helical gears
allowed a very high reduction ratio (τ34 = 0.056), a very com-
pact size, and high reliability to be achieved. On the other hand,
the high friction between the helical teeth hampered efficiency,
resulting equal to η34 = 0.415 as in [16].
Transmission between worm and main gears was designed
with a fixed joint (also the connection between pinion gear
and motor shaft) and defined unitary transmission ratio and
efficiency (τ23 = 1 and η23 = 1).
TABLE I
GEAR FEATURES
TABLE II
MECHANISM FEATURES
A low modulus (equal to 0.15) was selected for both the
pinion and the main gears, thus obtaining homogeneous rotation.
Considering the relation in [28], these two gears also featured
a transmission ratio of τ12 = 0.364 and an efficiency of η12 =
0.9.
From the gear features, it is possible to calculate total mech-
anism transmission ratio and total efficiency (τtot = 0.02 and
ηtot = 0.3735).
Both pinion and main gears were fabricated by using a
five-axis micro-CNC machining center (HSPC, KERN GmbH,
Germany). The steel worm gear and the bronze helical gears
were custom designed by the authors and fabricated by an
external workshop. The gears were modified by sink and mi-
crowire electrodischarge machining (EDM) (Micro Sink, Sarix,
Switzerland, and AP 200L, Sodick, Japan, respectively) to pro-
vide proper couplings between connecting parts. In particular,
the worm gear was cut at one end in order to obtain a rectangular
groove, while a T-shaped hole was made in each helical gear to
enable proper connection with the brass shaft. The three brass
shafts were fabricated by the CNC machining center. The motor
shaft was machined by microwire EDM and fitted into the hole
previously obtained in the center of the pinion gear. Finally, the
gears were assembled on custom made bushes in synthetic ruby
with buffing surface minimizing friction force and overall di-
mensions. The fabricated components of the legged mechanism
are showed in Fig. 3.
After dimensioning and fabricating the gears, the legs were
designed in terms of shape and length, on the basis of (6) and
size requirements. Since the legs needed to lift tissues rather
than to propel, the leg movement was conceived as reversible
Authorized licensed use limited to: UNIVERSITA PISA S ANNA. Downloaded on April 27,2010 at 16:59:04 UTC from IEEE Xplore.  Restrictions apply. 
174 IEEE/ASME TRANSACTIONS ON MECHATRONICS, VOL. 15, NO. 2, APRIL 2010
Fig. 3. Internal components of the legged mechanism. (a) Pinion gear.
(b) Main gear. (c) Worm gear. (d) Helical gear. (e) Brass shaft. (f) Bushes.
(g) NiTinol legs. All components are depicted on millimetric sheets.
(moving forward and backward, and backward and forward).
Therefore, a simple symmetrical leg geometry was selected. A
round shape was adopted for the leg tip in order to minimize
local tissue stresses. The legs were fabricated from a 0.5 mm
thickness superelastic alloy (NiTinol) plate by microwire EDM
[see Fig. 3(g)]. Superelastic alloy was chosen because it provides
high elasticity, particularly useful at the leg-shaft junction [29].
The total length of the locomotion module was 21 mm, while
the legs were 10.5 mm long and could be symmetrically closed
on both sides. Finally, considering (6) and assuming the force re-
quirement Flegs as 3 N, it was possible to calculate the minimum
value for γmot , which still met our specifications as 0.5.
The following paragraph describes the dimensioning of mag-
netic system, which led to the determination of 35% perfor-
mance loss (γmot = 0.65), this corresponds to Flegs = 3.8 N,
thus providing a safety margin on the desired specifications. If
extensive tests were to demonstrate that a lower motor perfor-
mance is sufficient, the force requirements could be reduced, as
also the size of the entire locomotion module.
C. Magnetic System
The magnetic system provides external active locomotion in
partially collapsed or uncollapsed regions of the LGI tract by
coupling of external and internal magnetic fields.
To help endoscopists easily handle the external source of
magnetic field, a cylindrical NdFeB N35 permanent magnet
placed on a passive hydraulic arm, was adopted (see Fig. 4).
Once the hybrid locomotion has been assessed, a robotic arm
may be used to replace the passive support, as proposed in [15].
The main features of the external permanent magnet (EPM)
were: a cylindrical shape (φ = 60 mm, L = 70 mm) with an
axial hole (φ = 10 mm) to allow fixation, a 1.5 kg weight, a
magnetic remanence of 1.21 T due to the selected material, and
a magnetization vector perpendicular to the main axis of the
cylinder.
First finite-elements method (FEM) analysis was performed
with COMSOL Multiphysics 3.4 (COMSOL Inc., Sweden) with
the aim to obtain the relationship between magnetic field and
distance from the surface of the EPM. All the EPM properties,
such as shape, dimension, magnetization direction, remanence,
and magnetic permeability (µ0 = 1.05) were set into the simu-
lation. The selected mesh consisted in about 450.000 elements
Fig. 4. EPM mounted on the hydraulic arm.
with a maximum element size set at 1/50 of the maximum geo-
metric feature in the scenario. The incremental ratio of the mesh
elements was set to 1.3, while the curvature factor and the mesh
curvature cutoff were 0.2 and 0.001, respectively.
In order to quantify γmot , we assumed that the effect of the
magnetic field acts as a torque working against the motor torque
itself. To verify this assumption, we performed the following
experimental procedure: firstly, given a fixed voltage supply
(3.7 V), motor rotational speed, and the related current absorp-
tion were measured as a function of distance between the motor
and the EPM surface. Then, considering the motor character-
istic curves, relating shaft rotational speed, current absorption,
and motor torque [26], we confirmed that rotational speed and
current absorption behave as if an increasing load was applied
to the shaft when the EPM-motor distance decreased. This al-
lowed us to assume the effect of the magnetic field surrounding
the motor as an external antagonistic torque (Cmag ) exerted on
the motor shaft. This perturbation reduced the torque (Mmot)
available for moving the mechanism.
The FEM analysis and the experimental results were used
to obtain a plot describing the experimental trend of the motor
rotational speed and the current absorption, both as a func-
tion of magnetic flux density. The rotational speed of the shaft
decreased with magnetic field, whereas the current absorption
increased. The torque Cmag was derived form the motor char-
acteristic curves and the measures of speed and current. Since
rising with the magnetic field, the coefficient γmot described by
the (7) decreased, as shown in Fig. 5
γmot =
0.5Cmot − Cmag
0.5Cmot
. (7)
The internal magnetic field was generated by a set of small
permanent magnets [internal permanent magnets (IPM)] em-
bedded into the capsule. The optimal IPM configuration would
maximize the volume of magnets into the space left available
inside the locomotion unit (roughly 500 mm3). It would also
minimize motor interferences (γmot = 0.65) and maximize mag-
netic attraction force with the selected EPM, according to the
requirements presented in Section II-A.
Off-the-shelf cylindrical NdFeB N50 permanent magnets,
having a 1.45 T magnetic remanence were selected as IPM
with the purpose of maximizing the magnetic flux density for a
Authorized licensed use limited to: UNIVERSITA PISA S ANNA. Downloaded on April 27,2010 at 16:59:04 UTC from IEEE Xplore.  Restrictions apply. 
SIMI et al.: DESIGN, FABRICATION, AND TESTING OF A CAPSULE WITH HYBRID LOCOMOTION FOR GI TRACT EXPLORATION 175
Fig. 5. Experimental relationship between γm ot and magnetic flux density.
Fig. 6. (a) FEM simulation of the selected IPM configuration: plot of magnetic
flux density. (b) Experimental test bench to quantify γm ot .
given volume. Given the shape and the volume (i.e., 500 mm3)
of the space available inside the capsule, as also our previous
experience in magnetic steering of endoscopic capsule [15], the
ten most suitable IPM configurations, in terms of number and
arrangement of the magnets composing the IPM, were simu-
lated. The main goal was to obtain a rapid solution for each
configuration in terms of magnetic field engaging the motor and
magnetic attraction force with the EPM. A complete analytical
solution for the forces between permanent magnets was based
on Kelvin’s formula and on Biot-Savart’s law as reported in [30].
The selected IPM configuration [see Fig. 6(a)] was composed
by six cylindrical NdFeB N50 magnets (φ = 3.175 mm, L =
6.35 mm) and 12 cylindrical NdFeB N50 magnets (φ =
1.58 mm, L = 3.175 mm) with axial magnetization and to-
tal volume of 475 mm3 (KJ Magnetics, Jamison). FEM analysis
of this configuration showed that the motor was engaged in
an average magnetic field of 25 mT. Thus, the parameter γmot
was estimated as 0.65 (see Fig. 5), meeting the requirements.
The simulated magnetic attraction between IPM and EPM [see
Fig. 7(a)] was calculated as 135 mN at 10 cm and 550 mN at
5 cm, enabling the capsule (approximately 13.5 g) to be lifted
within this distance range. Another outcome of this simulation
Fig. 7. (a) FEM simulation to evaluate magnetic attraction force between EPM
and IPM. (b) Experimental test bench with the IPM connected to a load cell.
Fig. 8. Experimental plot of magnetic attraction force versus EPM–IPM dis-
tance in the selected configuration.
was the negligible effect of the EPM magnetic field on the motor
performance at the operative distance of 10 cm. The force re-
quirement of 3 N defined in Section II-A, was still fulfilled at an
EPM-capsule distance of 9 cm (γmot = 0.5), while at shorter dis-
tances the interference increased until the mechanism stopped at
5 cm. In this case, only magnetic locomotion could be used. It is
worth noting that in the case of obese patients, a 10 cm distance
between EPM and IPM may not be a feasible option. In this
case, a different EPM needs to be used, either by increasing its
volume or selecting a material with higher remanence.
In order to verify FEM results, experimental tests were per-
formed with the selected configuration. Firstly, a γmot of 0.65
was experimentally assessed with a dedicated test bench re-
ported in Fig. 6(b). Then, a monoaxial load cell (FMI-210, Al-
luris, Germany) was used to verify the attraction force between
EPM and IPM at different distances [see Fig. 7(b)]. The force
measured with the load cell plotted in Fig. 8, confirmed the FEM
analysis results.
Authorized licensed use limited to: UNIVERSITA PISA S ANNA. Downloaded on April 27,2010 at 16:59:04 UTC from IEEE Xplore.  Restrictions apply. 
176 IEEE/ASME TRANSACTIONS ON MECHATRONICS, VOL. 15, NO. 2, APRIL 2010
Fig. 9. Communication architecture.
D. On-Board Electronics
The hybrid capsule integrated a four-layer circular electronic
board (φ = 10.3 mm, thickness = 2.5 mm) with a wireless 8051
microcontroller (CC2430, Texas Instruments) and a driver for
brushless dc motor activation [31]. The microcontroller embed-
ded an IEEE 802.15.4 compliant wireless transceiver working
at 2.4 GHz. This proved to be an effective solution for LGI tract
telemetry, as reported in [32]. The capsule was interfaced with
a standard PC through a custom designed universal serial bus
(USB) dongle, featuring a second CC2430 and a serial USB
converter (UM232, FDTI Chip, U.K.). A purposely developed
human–machine interface (HMI) ran on the PC, enabling the
user to control the capsule and receive a feedback on its status
(see Fig. 9). The HMI was implemented with labview (National
Instruments) and it was also compatible with voice commands.
The endoscopist, therefore, by pronouncing the proper com-
mands (including direction and angle) could open the legs while
operating the magnet.
Deep sleep modality was also implemented, thus allowing
the capsule to use just 0.3 µA when in idle mode. Two external
contacts if short-circuited, enabled the system to wake up and
start communication with the HMI.
As regards the energy source, a single rechargeable lithium
ion polymer battery (LP50, Plantraco, Canada) was used to
maximize the ratio between capacity and volume. Its main fea-
tures were: peak current of 500 mA, 3.7 V voltage, compact
size (9.5 × 11 × 3.3 mm3), and a 50 mAh nominal capacity.
The battery only needs to be charged during the development
phase, since multiple tests may be conduced without the need
to replace the energy source.
In order to estimate the battery lifetime, current absorption
in different working modalities was measured with a digital
oscilloscope. Average current drain during periodic activation
of wireless transmission was 2.75 mA, whereas motor activation
required an average current of 120.6 mA. This resulted in a
range going from 18 h (only telemetry activities) to 25 min
(continuative motor activation) depending on the use of the
legged mechanism. Considering that the legs requires 16.7 s
Fig. 10. (a) Chassis. (b) Body. (c) Tank. (d) Frontal plug. (e) Rear plug.
(f) Cap.
Fig. 11. (a) Assembly of worm gear and helical gears in the chassis.
(b) Assembly of the motor.
to complete a 180◦ span, 90 cycles can be performed by the
capsule before the battery expires. From a different perspective,
each 180◦ leg movement consumes about 12 min of device
autonomy.
E. Packaging and Assembly
After defining all the main features of the capsule, a com-
pact packaging was designed, which focused on small overall
dimensions, robustness, and ease of assembly. Biocompatibil-
ity of materials was not addressed at this stage of development
and will be pursued, once the hybrid capsule will be completely
assessed.
The legged mechanism was included in two complementary
parts, named chassis and body, as in Fig. 10 that can be joined
together by five screws. This solution ensured stability of the
mechanical components and correct protection of the electri-
cal parts. Chassis and body were fabricated in Ergal (i.e., an
aluminum alloy) by CNC machining. Assembly of the legged
mechanism components in the chassis is illustrated in Fig. 11.
A cylindrical tank [see Fig. 10(c)] was fabricated in pol-
yarylether ether ketone (PEEK), a thermoplastic semicrystalline
polymer, which hosted both electronics and battery. Additional
parts (frontal and rear plugs, [Fig. 10(d)–(e)]) were fabricated
either in PEEK or in acrylic photopolymer to match the single
parts of the capsule and to make the device waterproof.
Three holes in the rear plug contained three magnetic con-
tacts that wake up the capsule from deep sleep and recharge
the battery. A hemispherical cap [see Fig. 10(f)], simulating the
presence of a vision system, was fabricated with a stereolitho-
graphic technique (3-D Printer Invision Si2) and fixed frontally.
Authorized licensed use limited to: UNIVERSITA PISA S ANNA. Downloaded on April 27,2010 at 16:59:04 UTC from IEEE Xplore.  Restrictions apply. 
SIMI et al.: DESIGN, FABRICATION, AND TESTING OF A CAPSULE WITH HYBRID LOCOMOTION FOR GI TRACT EXPLORATION 177
Fig. 12. First hybrid capsule prototype.
Fig. 13. Leg force measurements. (a) Cable is fixed at the tip of each leg. The
load cell measures the tension of the cable. (b) Simple model for calculating
effective leg force.
Fig. 12 shows the wireless hybrid capsule prototype. It is
composed roughly of 60 different parts, and once assembled,
measures 14 mm in diameter, 44 mm in length, and weights
13.5 g.
III. EXPERIMENTAL VALIDATION
A. Bench Testing
The first experiments were performed to assess if the me-
chanical features of the prototype met the design specifications.
In order to measure the force generated by the three legs, three
nonextensible cables of the same length were fixed on the tip
of each leg. The cables were then tied together and fixed to a
monoaxial load cell, as in Fig. 13(a). When the motor was ac-
tivated by a wireless command, the legs pulled the cables and
an average force of 3.58 N was acquired by the load cell. By
measuring the cable and leg angles (θ1 = 8◦ and θ2 = 28◦), as
defined in Fig. 13(b), we derived the Flegs .
By using a simple vectorial sum
Flegs =
Fm
cos(θ1)cos(θ2 − θ1) (8)
it was possible to calculate the effective force generated by the
three legs as 3.85 N. This value met the mechanical requirements
listed in Section II-A.
Fig. 14. (a) In vitro bench test. (b) Capsule is stuck in the cavity. (c)–(e)
Activation of the legged mechanism to get rid of collapsed lumen (see white
arrow).
Fig. 15. (a) Ex vivo bench test. (b)–(d) Activation of the legged mechanism to
get rid of collapsed tissue. (e) Magnetic locomotion starts again after removing
the tissue around the capsule.
B. Locomotion Performance Assessment
Setting up a proper test bench and a reliable protocol to as-
sess hybrid capsule performance is no trivial matter, since un-
collapsed regions in the human LGI are not easily predictable in
terms of number and extension. We implemented preliminary in
vitro, ex vivo, and in vivo trials to estimate average transit time
of the capsule provided with active locomotion. However, an
extensive statistical analysis of a relevant number of trials must
be performed in order to obtain a definitive evaluation of the
proposed device. Throughout the experiments, the capsule was
observed by a gastroscope (Karl Storz Endoscopy, Tu¨ttlingen,
Germany) in order to verify capsule motion and to assess the
lumen dilatation generated by the three legs.
Ten preliminary in vitro tests were conducted in a 20 cm tract
of latex colon simulator (Kyoto Kagaku, Kyoto, Japan). Then,
ten ex vivo trials with a 20 cm fresh porcine colon segment
were performed. Considering the weight of the abdominal wall
on the colon [33] and the force required for the dilatation of a
collapsed tract [10], [22], the intestine was covered with water
filled plastic bags (about 350 g each) [see Figs. 14(a) and 15(a)]
in order to simulate the surrounding tissues.
The EPM was placed on the hydraulic arm and maintained at
an operative distance of 10 cm from the capsule during all tests.
The capsule was dragged by the EPM as much as possible and
the legged mechanism was activated only when the capsule was
stuck in tissue folds. Based on the time required to complete a
traditional colonoscopic procedure [34], the trial was considered
failed if the capsule was not able to cross the entire segment after
10 min.
Both in vitro and ex vivo bench tests confirmed that the mag-
netic field coupling can move and steer the capsule in slightly
Authorized licensed use limited to: UNIVERSITA PISA S ANNA. Downloaded on April 27,2010 at 16:59:04 UTC from IEEE Xplore.  Restrictions apply. 
178 IEEE/ASME TRANSACTIONS ON MECHATRONICS, VOL. 15, NO. 2, APRIL 2010
collapsed regions, whereas activation of the legged mechanism
(especially when moving forward, and then, backward) was able
to lift and distend collapsed tissues or to overcome tissue folds.
In all in vitro tests the capsule reached the end of the simulator
within the 10 min time frame, whereas in the ex vivo trials it
was able to travel along the intestine tract in only seven out
of ten cases. The latex colon simulator had an averagely larger
lumen (about φ = 35 mm) than the ex vivo specimen and it
also tended to hold the original shape of the cavity despite
the weight of the bags, thus facilitating magnetic locomotion.
Although the ex vivo colon specimen was very slippery and was
positioned on the bench without bends, the three failed trials
were mainly due to the absence of muscular tone (as in the in
vivo condition), which resulted in completely collapsed tissue
around the capsule. It is worth mentioning that failure does not
mean capsule malfunctioning, but inability to cover the entire
20 cm of the specimen in the targeted time frame.
The capsule needed an average time of 2 min and two legged
mechanism activations to go through the latex colon simulator
about 4 min and five mechanism activations were required in the
porcine lumen (excluding the three failed trials). The average
speed was estimated as 10 and 5 cm/min for in vitro and ex vivo
tests, respectively.
Ten in vivo trials were performed on a total of four female
pigs with a weight ranging from 25 to 35 kg. The experiments
were conducted in an authorized laboratory with the assistance
and collaboration of a specially trained medical team, in compli-
ance with all ethical requirements and regulatory issues related
to animal experiments. After intravenous sedation of the animal
and preparation of the bowel with water enemas, an endoscope
was used to introduce the capsule and position it 40 cm from
the anus. The legs of the capsule were closed into the body in
order to facilitate sliding. Then, while the gastroscope observed
the capsule without insufflation, the hybrid device was moved
and steered by means of magnetic forces in order to simulate
a diagnostic procedure. Magnetic locomotion always allowed
precise orientation of the device, but the motion was only pos-
sible in uncollapsed or slightly collapsed regions. During the
in vivo tests, the capsule was often lodged because either the
tissues surrounded it completely or the cavity in front of it was
highly collapsed. Activation of the legged mechanism enabled
the layout of the capsule to be modified and the lumen to be en-
larged. The motion of the three legs, moving from front to back,
lifted the tissue easily and propelled the capsule forward. Exper-
imental evidence showed that partial opening (around 45◦) of
the legs on the rear side was the best capsule layout for magnet-
ically assisted locomotion, since the tissue was kept away from
the device surface. With this configuration, a leverage effect
of the legs on the collapsed tissue was achieved by oscillating
the EPM, and sharp bends were effectively handled. Finally,
legs motion from back to front allowed the capsule to move
backward. This was beneficial whenever the capsule was stuck
in a tissue pocket, such as a diverticulum. Naturally, the legs
could be closed either toward the front or the back, however,
they could not be closed a priori in order to impart the most
favorable motion to the capsule when approaching an obstacle.
Nevertheless, by correctly manoeuvering the EPM, a 135◦ mo-
Fig. 16. Capsule opens its legs while immersed in physiological fluids.
tion can be obtained without tissue entrapment. At this point,
the legs could be ready to move in the most favorable direction.
Ten locomotion trials were performed during the in vivo ex-
perimental sessions. In six cases, the capsule covered the entire
40 cm tract toward the anus in an average time of 5 min and an
average speed of 8 cm/min. An average number of five legged
mechanism activations was required. However, in four trials, the
capsule was not able to complete the journey within the 20 min
time frame. This was mainly due to the geometry of the porcine
colon, which is much more complex than in humans. Under-
standing the correct leg activation sequence, also contributed
to determining the high number of failures (three out of four
failures occurred during the first half of the trials).
During the in vivo tests, the capsule was also moved against
natural peristalsis, showing satisfactory performance. It is also
worth mentioning that the hybrid capsule engaged in physio-
logical fluids, as represented in Fig. 16, without any significant
loss in performance, thus demonstrating its reliability and ro-
bustness.
IV. CONCLUSION
The hybrid capsule developed, fabricated, and tested by the
authors represents an effective solution for the active locomo-
tion of endocapsules in the GI tract. The capsule incorporates
several innovative features in terms of combination of different
actuation principles. In fact, two different solutions for active
locomotion (i.e., external magnetic actuation and internal actu-
ation structures) were merged into the same device to overcome
the most significant limitations of the two separate approaches.
Combining permanent magnets and electromagnetic actuators
is not a trivial matter due to interference that reduces the motor
torque needed to move the mechanism. The paper contributes to
the design of miniaturized devices combining magnetic and dc
motor actuation, and provides quantitative guidelines for merg-
ing the two principles.
The fabricated hybrid locomotion capsule is a wireless sys-
tem conceived with a “totally on-board” philosophy: it embeds
an internal motor mechanism, several permanent magnets, a
wireless microcontroller, and a rechargeable battery. Once as-
sembled, it measures 14 mm diameter, 44 mm in length, and
weights 13.5 g. Experimental bench tests reveled that the in-
ternal mechanism produces 3.85 N force at the leg tip and that
magnetic field coupling between internal and external magnets
Authorized licensed use limited to: UNIVERSITA PISA S ANNA. Downloaded on April 27,2010 at 16:59:04 UTC from IEEE Xplore.  Restrictions apply. 
SIMI et al.: DESIGN, FABRICATION, AND TESTING OF A CAPSULE WITH HYBRID LOCOMOTION FOR GI TRACT EXPLORATION 179
provides an attraction force capable of performing controlled
locomotion at an operative distance of 10 cm.
Comparison with other locomotion modules designed for en-
doscopic capsules presented in literature can be conducted on
the basis of several parameters. The most significant in our
opinion, is power consumption. From this standpoint, the hy-
brid capsule combines the zero power required by magnetic
dragging (which becomes ineffective in collapsed regions of
the LGI tract), with the specific advantages that a leg set can
provide (e.g., tissue dilatation and freeing of the capsule from
tissue folds). Our most advanced legged capsule prototype [22]
requires a 10 mm diameter and 30 mm long battery to complete
colon examination.
Another significant parameter considered when comparing
locomotion modules is speed. With special reference to our
previous legged capsule prototypes, the 4-leg capsule [16] speed
was 3 cm/min, the 8-leg capsule [21] speed was 4 cm/min, and
the 12-leg capsule [22] speed was 5 cm/min, while the hybrid
capsule achieved a speed of 8 cm/min in the in vivo trials.
The results obtained with this first prototype support the fea-
sibility of the hybrid locomotion concept. The next generation
of hybrid capsules could be reduced to a swallowable size by
further optimizing each component. Considering that the 3.85 N
force produced by the internal structure seems to be grater than
the force required to lift collapsed tissue, a miniaturized mech-
anism may be designed as next step. Smaller integrated motors
could reduce the size of the capsule, whereas many other fea-
tures could be modified to improve functionality. New leg shapes
(e.g., with a smoother profile) could be studied to minimize local
tissue stresses or to facilitate colon distension.
In order to enable actively controlled endoscopic procedures,
a real-time wireless vision module with a high frame rate must
be integrated on-board the capsule. Although a system with
these features is not commercially available, on-going efforts
toward this goal are provided in literature [24]. In particular, our
group is developing a 20 fps real-time wireless camera system
for capsule endoscopy, weighing 3.5 g and measuring 450 mm3
in volume [25]. The system requires approximately 100 mW
to operate, inclusive of four white LED illumination system.
Once this module will be available for integration, the whole
design process of the magnetic link will be reiterated, following
the procedure described in this paper and taking into account the
modified weight of the capsule. As regards power supply, the
additional burden introduced by the vision subsystem can be
powered by replacing the battery with a wireless power supply
solution, as reported in [35].
While real-time image stream should be the main feedback
for the operator, additional sensing strategies may be imple-
mented in order to attain fully autonomous hybrid locomotion.
In such a scenario, the EPM could be mounted as end-effector
on a robotic arm, as in [15], and real-time localization of the
capsule (based on magnetic fields [36]) would allow the sys-
tem to autonomously run the legged mechanism whenever the
capsule does not follow the EPM movement. Additionally, con-
tact or pressure sensors on the surface of the capsule could
be exploited in order to autonomously trigger activation of the
legs.
ACKNOWLEDGMENT
The authors would like to thank N. Funaro and the Staff of
the Center for Research in Microengineering Laboratory me-
chanical workshop for manufacturing the prototypes. They are
also grateful to Dr. Trivella, Dr. Burchielli, and the Staff of the
Experimental Surgery Center of San Piero a Grado (Pisa) for
their help during the in vivo testing phase of the device. Finally,
they would like to thank Dr. Piccigallo for his technical support
on brushless motor analysis.
REFERENCES
[1] G. Iddan, G. Meron, A. Glukhovsky, and P. Swain, “Wireless capsule
endoscopy,” Nature, vol. 405, no. 6785, pp. 417–418, May 2000.
[2] M. Waterman and R. Eliakim, “Capsule enteroscopy of the small intes-
tine,” Abdom. Imaging, vol. 34, pp. 452–458, 2008.
[3] A. Moglia, A. Menciassi, P. Dario, and A. Cuschieri, “Clinical update:
Endoscopy for small-bowel tumours,” Lancet, vol. 14, no. 370, pp. 114–
116, 2007.
[4] Y. C. Metzger, S. N. Adler, A. B. Shitrit, B. Koslowsky, and I. Bjarnason,
“Comparison of a new PillCam SB2 video capsule versus the standard
PillCam SB for detection of small bowel disease,” Rep. Med. Imaging,
vol. 2, pp. 7–11, 2009.
[5] Olympus Endocapsule. (2010, Feb.). [Online]. Available: www.olympus-
europa.com/endoscopy/
[6] S. Bang, J. Park, S. Jeong, Y. H. Kim, H. B. Shim, T. S. Kim, D. H. Lee, and
S. Y. Song, “First clinical trial of the miro capsule endoscope by using a
novel transmission technology: Electric-field propagation,” Gastrointest.
Endosc., vol. 69, no. 2, pp. 253–259, 2009.
[7] P. Swain, “The future of wireless capsule endoscopy,” World J. Gastroen-
terol., vol. 14, no. 26, pp. 4142–4145, 2008.
[8] American Cancer Society. (2010, Feb.). [Online]. Available: http://www.
cancer.org/
[9] P. Dario, P. Ciarletta, A. Menciassi, and B. Kim, “Modelling and exper-
imental validation of the locomotion of endoscopic robots in the colon,”
Int. J. Robot. Res., vol. 23, no. 4–5, pp. 549–559, 2004.
[10] C. Stefanini, A. Menciassi, and P. Dario, “Modeling and experiments
on a legged microrobot locomoting in a tubular, compliant and slippery
environment,” Int. J. Robot. Res., vol. 25, no. 5–6, pp. 551–560, 2006.
[11] X. Wang and M. Meng, “A magnetic stereo actuation mechanism for
active capsule endoscope,” in Proc. 29th Annu. Int. Conf. IEEE EMBS,
2007, pp. 2811–2814.
[12] F. Carpi, S. Galbiati, and A. Carpi, “Controlled navigation of endoscopic
capsules: Concept and preliminary experimental investigations,” IEEE
Trans. Biomed. Eng., vol. 54, no. 11, pp. 2028–2036, Nov. 2007.
[13] F. Carpi and C. Pappone, “Magnetic manouvering of endoscopic capsules
by means of a robotic navigation system,” IEEE Trans. Biomed. Eng.,
vol. 56, no. 5, pp. 1482–1490, May 2009.
[14] F. Volke, A. Schneider, J. Gerber, M. Reimann-Zawadzki, E. Rebinovitz,
C. Mosse, and P. Swain, “In vivo remote manipulation of modified capsule
endoscopes using an external magnetic field,” Gastrointest. Endosc.,
vol. A, pp. 121–122, 2008.
[15] G. Ciuti, P. Valdastri, A. Menciassi, and P. Dario, “Robotic magnetic
steering and locomotion of capsule endoscope for diagnostic and surgical
endoluminal procedures,” Robotica, vol. 28, pp. 199–207, 2010.
[16] M. Quirini, A. Menciassi, S. Scapellato, C. Stefanini, and P. Dario, “Design
and fabrication of a motor legged capsule for the active exploration of the
gastrointestinal tract,” IEEE/ASME Trans. Mechatronics, vol. 13, no. 2,
pp. 169–179, Apr. 2008.
[17] H. Park, S. Park, E. Yoon, B. Kim, J. Park, and S. Park, “Paddling based
microrobot for capsule endoscopes,” in Proc. IEEE Int. Conf. Robot.
Autom., 10–14 Apr. 2007, pp. 3377–3382.
[18] P. Glass, M. Sitti, A. Pennathur, and R. Appasamy, “A swallowable teth-
ered capsule endoscope for diagnosing barrett’s esophagus,” Gastrointest.
Endosc., vol. 69, no. 5, p. AB106, Apr. 2009.
[19] W. Li, W. Guo, M. Li, and Y. Zhu, “A novel locomotion principle for
endoscopic robot,” in Proc. IEEE Int. Conf. Mechatronics Autom., Jun.,
2006, pp. 1658–1662.
[20] B. Kim, S. Park, and J.-o. Park, “Microrobots for a capsule endoscope,” in
Proc. IEEE/ASME Int. Conf. Adv. Intell. Mechatronics, 14–17 Jul. 2009,
pp. 729–734.
Authorized licensed use limited to: UNIVERSITA PISA S ANNA. Downloaded on April 27,2010 at 16:59:04 UTC from IEEE Xplore.  Restrictions apply. 
180 IEEE/ASME TRANSACTIONS ON MECHATRONICS, VOL. 15, NO. 2, APRIL 2010
[21] M. Quirini, S. Scapellato, A. Menciassi, P. Dario, F. Rieber, C. N. Ho,
S. Schostek, and M. O. Schurr, “Feasibility proof of a legged locomotion
capsule for the GI tract,” Gastrointest. Endosc., vol. 67, no. 7, pp. 1153–
1158, 2008.
[22] P. Valdastri, R. J. Webster, C. Quaglia, M. Quirini, A. Menciassi, and
P. Dario, “A new mechanism for meso-scale legged locomotion in com-
pliant tubular environments,” IEEE Trans. Robot., vol. 25, no. 5, pp. 1047–
1057, Oct. 2009.
[23] D. Accoto, C. Stefanini, L. Phee, A. Arena, G. Pernorio, A. Menciassi,
M. C. Carrozza, and P. Dario, “Measurements of tha frictional properties
of the gastrointestinal tract,” in Proc. World Tribol. Congr., 3–7 Sep. 2001,
p. 728.
[24] M. Vatteroni, D. Covi, C. Cavallotti, P. Valdastri, A. Menciassi, P. Dario,
and A. Sartori, “Smart optical CMOS sensor for endoluminal applica-
tions,” Procedia Chem., vol. 1, no. 1, pp. 1271–1274, 2009.
[25] The VECTOR project website. (2010, Feb.). [Online]. Available: http://
www.vector-project.com/
[26] Namiki Precision Jewel Co., Ltd. (2010, Feb.). [Online]. Available:
http://www.namiki.net
[27] Boston Gear Transmission Products. (2010, Feb.). [Online]. Available:
http://www.bostongear.com/products/open/theory.html
[28] RoyMech Engineering Design. (2010, Feb.). [Online]. Available: http://
www.roymech.co.uk/Useful_Tables/Drive/Gears.html
[29] E. Buselli, P. Valdastri, M. Quirini, A. Menciassi, and P. Dario, “Su-
perelastic leg design optimization for an endoscopic capsule with active
locomotion,” Smart Mater. Struct., vol. 18, no. 1, pp. 015001-1–015001-8,
2009.
[30] J. Agashe and D. Arnold, “A study of scaling and geometry effects on the
forces between cuboidal and cylindrical magnets using analytical force
solution,” J. Phys. D, Appl. Phys., vol. 41, pp. 105001-1–105009, 2008.
[31] E. Susilo, P. Valdastri, A. Menciassi, and P. Dario, “A miniaturized wire-
less control platform for robotic capsular endoscopy using pseudokernel
approach,” Sens. Actuators A, vol. 156, pp. 49–58, 2009.
[32] P. Valdastri, A. Menciassi, and P. Dario, “Transmission power require-
ments for novel zigbee implants in the gastrointestinal tract,” IEEE Trans.
Biomed. Eng., vol. 55, no. 6, pp. 1705–1710, Jun. 2008.
[33] M. Jensen, J. Kanaley, J. Reed, and P. Sheedy, “Measurement of abdom-
inal and visceral fat with computed tomography and dual-energy X-ray
absorptiomentry,” Amer. J. Clin. Nutr., vol. 61, pp. 274–278, 1995.
[34] American Society for Gastrointestinal Endoscopy. (2010, Feb.). [Online].
Available: http://www.askasge.org/
[35] R. Carta, G. Tortora, J. Thone´, B.Lenaers, P. Valdastri, A. Menciassi,
P. Dario, and R. Puers, “Wireless powering for a self-propelled and steer-
able endoscopic capsule for stomach inspection,” Biosens. Bioelectron.,
vol. 25, no. 4, pp. 845–851, 2009.
[36] M. Hocke, U. Schone, H. Richert, P. Gornert, J. Keller, P. Layer, and
A. Stallmach, “Every slow-wave impulse is associated with motor activity
of the human stomach,” Amer. J. Physiol. Gastrointest. Liver Physiol.,
vol. 296, no. 4, pp. 709–716, 2008.
Massimiliano Simi received the Master’s degree in
biomedical engineering from the University of Pisa,
Pisa, Italy, in April 2009. He is currently working
toward the Ph.D. degree in biorobotics in the Center
for Research in Microengineering (CRIM) Labora-
tory, Scuola Superiore Sant’Anna, Pisa.
He was engaged in the design, fabrication, and test-
ing of an endocapsule with hybrid locomotion. Since
June 2008, he has been with the CRIM Laboratory as
a Research Assistant. His current research interests
includes medical robotics and biomechatronics.
Pietro Valdastri (M’05) received the Master’s
(Hons.) degree in electronic engineering from the
University of Pisa, Pisa, Italy, in 2002, and the Ph.D.
degree in bioengineering from the Scuola Superiore
Sant’Anna, Pisa, in 2006.
He is currently an Assistant Professor of biomed-
ical robotics at the Scuola Superiore Sant’Anna,
where he is engaged in research on implantable
robotic systems and active capsule endoscopy. He is
also involved in several European projects regarding
the development of minimally invasive and wireless
biomedical devices.
Claudio Quaglia (M’05) received the Master’s de-
gree in mechanical engineering from the University
of Pisa, Pisa, Italy, in 2005. He is currently working
toward the Ph.D. degree in microsystem engineering
at the Scuola Superiore Sant’Anna, Pisa.
He was engaged in the employment of a triaxial
accelerometer for tool condition monitoring in auto-
matic machining. Since 2006, he has been with the
Scuola Superiore Sant’Anna. His current research in-
terests include medical robotics and micromechanical
systems.
Arianna Menciassi (M’02) received the Master’s de-
gree in physics from the University of Pisa, Pisa, Italy,
in 1995, and the Ph.D. degree in bioengineering from
the Scuola Superiore Sant’Anna, Pisa, in 1999.
She is currently an Associate Professor of biomed-
ical robotics at the Scuola Superiore Sant’Anna. She
has authored or coauthored more than 130 interna-
tional papers, about 70 in Institute for Scientific In-
formation (ISI) journals, and five book chapters on
medical devices and microtechnologies. Her research
interests include biomedical micro- and nanorobotics
for the development of innovative devices for surgery, therapy, and diagnostics.
Paolo Dario (M’99–SM’01–F’03) received the Mas-
ter’s degree in mechanical engineering from the Uni-
versity of Pisa, Pisa, Italy, in 1977.
He is currently a Professor of biomedical robotics
at the Scuola Superiore Sant’Anna, Pisa, where he su-
pervises a team of about 150 young researchers. He
has authored or coauthored more than 200 ISI jour-
nal papers, many international patents, and several
book chapters on medical robotics. His research in-
terests include biorobotics, mechatronic and robotic
systems for rehabilitation, prosthetics, surgery, and
microendoscopy.
Mr. Dario is a recipient of the Joseph Engelberger Award as a Pioneer of
Biomedical Robotics.
Authorized licensed use limited to: UNIVERSITA PISA S ANNA. Downloaded on April 27,2010 at 16:59:04 UTC from IEEE Xplore.  Restrictions apply.