Development of an in-shoe pressure-sensitive device for gait analysis.
ABSTRACT In this work, we present the development of an in-shoe device to monitor plantar pressure distribution for gait analysis. The device consists in a matrix of 64 sensitive elements, integrated with in-shoe electronics and battery which provide an high-frequency data acquisition, wireless transmission and an average autonomy of 7 hours in continuous working mode. The device is presented along with its experimental characterization and a preliminary validation on a healthy subject.
- SourceAvailable from: Janez Podobnik[Show abstract] [Hide abstract]
ABSTRACT: This paper presents algorithms for detection of gait initiation and termination using wearable inertial measurement units and pressure-sensitive insoles. Body joint angles, joint angular velocities, ground reaction force and center of plantar pressure of each foot are obtained from these sensors and input into supervised machine learning algorithms. The proposed initiation detection method recognizes two events: gait onset (an anticipatory movement preceding foot lifting) and toe-off. The termination detection algorithm segments gait into steps, measures the signals over a buffer at the beginning of each step, and determines whether this measurement belongs to the final step. The approach is validated with 10 subjects at two gait speeds, using within-subject and subject-independent cross-validation. Results show that gait initiation can be detected timely and accurately, with few errors in the case of within-subject cross-validation and overall good performance in subject-independent cross-validation. Gait termination can be predicted in over 80% of trials well before the subject comes to a complete stop. Results also show that the two sensor types are equivalent in predicting gait initiation while inertial measurement units are generally superior in predicting gait termination. Potential use of the algorithms is foreseen primarily with assistive devices such as prostheses and exoskeletons.Medical Engineering & Physics 08/2013; · 1.78 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: This paper introduces the design and development of a novel pressure-sensitive foot insole for real-time monitoring of plantar pressure distribution during walking. The device consists of a flexible insole with 64 pressure-sensitive elements and an integrated electronic board for high-frequency data acquisition, pre-filtering, and wireless transmission to a remote data computing/storing unit. The pressure-sensitive technology is based on an optoelectronic technology developed at Scuola Superiore Sant'Anna. The insole is a low-cost and low-power battery-powered device. The design and development of the device is presented along with its experimental characterization and validation with healthy subjects performing a task of walking at different speeds, and benchmarked against an instrumented force platform.Sensors 01/2014; 14(1):1073-93. · 1.95 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: This paper presents a gait phase detection algorithm for providing feedback in walking with a robotic prosthesis. The algorithm utilizes the output signals of a wearable wireless sensory system incorporating sensorized shoe insoles and inertial measurement units attached to body segments. The principle of detecting transitions between gait phases is based on heuristic threshold rules, dividing a steady-state walking stride into four phases. For the evaluation of the algorithm, experiments with three amputees, walking with the robotic prosthesis and wearable sensors, were performed. Results show a high rate of successful detection for all four phases (the average success rate across all subjects >90%). A comparison of the proposed method to an off-line trained algorithm using hidden Markov models reveals a similar performance achieved without the need for learning dataset acquisition and previous model training.Sensors 01/2014; 14(2):2776-94. · 1.95 Impact Factor
33rd Annual International Conference of the IEEE EMBS
Boston, Massachusetts USA, August 30 - September 3, 2011
Abstract—In this work, we present the development of an in-
shoe device to monitor plantar pressure distribution for gait
analysis. The device consists in a matrix of 64 sensitive
elements, integrated with in-shoe electronics and battery which
provide an high-frequency
transmission and an average autonomy of 7 hours in
continuous working mode. The device is presented along with
its experimental characterization and a preliminary validation
on a healthy subject.
data acquisition, wireless
easurement of plantar pressure distribution is widely
recognized as a key tool in clinical gait analysis (e.g.
in case of gait abnormalities , diabetes mellitus,
peripheral neuropathies and musculoskeletal disorders ),
as well as in footwear evaluation , and in sport training
Two types of devices are commonly used for plantar
pressure monitoring: force platforms and pressure-sensitive
foot insoles. The former, e.g. the EMED®-SF (Novel USA,
Inc., Minneapolis, MN, USA), are high-resolution, high-
frequency floor-mounted matrices of pressure sensors, which
can capture three-axial pressure data of the barefoot. While
these platforms are very valuable for clinical studies, they
are not suitable when measurements of pressure at the shoe-
foot interface are required , or when a high-portability is
desired. To tackle these needs, pressure-sensitive foot
insoles have been developed through several sensing
technologies, ranging from force-sensing resistors, as in the
F-Scan® system, (Tekscan Inc., Boston, MA, USA), to
capacitive sensors, as in the Pedar® system (Novel GmbH,
Germany) to piezo-resistive sensors, as in the paromed®
(Vertriebs GmbH & Co, Neubeuern, Germany). However,
most of these systems require electrical wires to connect to
wireless communication modules that are strapped around
the waist or the ankle of the user, causing discomfort and
making long-time monitoring and recordings during daily
life or sport activities difficult to obtain. For these
applications, devices which fit entirely inside the shoe, with
a long time autonomy, and with wireless communication are
desirable. Some devices of this kind have been developed in
This work was partly supported by the EU within the EVRYON
Collaborative Project STREP (Evolving Morphologies for Human-Robot
Symbiotic Interaction, Project FP7-ICT-2007-3-231451).
All authors are with The BioRobotics Institute, Scuola Superiore
Sant’Anna, viale Rinaldo Piaggio 34, 56025, Pontedera (Pi), Italy.
S.M.M. De Rossi is corresponding author to provide phone: +39 050
883472; fax: +39 050 883 497; e-mail: firstname.lastname@example.org.
the last years , tough with a reduced temporal and
spatial resolution (e.g. six sensors are used in , four in
), and including a small-scale integrated acquisition
In this work, we present a new in-shoe device to monitor
plantar pressure, consisting in a matrix of 64 sensitive
elements, integrated with in-shoe conditioning electronics
and battery. The device provides high-frequency data
acquisition, on-board computation of center of pressure and
ground reaction force, and wireless communication, with an
autonomy of about 7 hours in continuous working mode.
The device can replace the insole of commercially available
shoes, without interfering with the normal gait. The system
is presented along with an experimental characterization and
a preliminary validation on a healthy subject.
II. DEVELOPMENT OF THE INSOLE HARDWARE AND
Our in-shoe device comprises two hardware parts: an array
of pressure sensors, and the on-board electronics for signal
conditioning and wireless data transmission. The sensor
array is made of 64 silicone-covered opto-electronic pressure
sensors, is deputed for pressure transduction, and is
connected to the on-board electronics through two flat cables
carrying unamplified analog voltage signals. The on-board
electronics performs high-frequency data acquisition,
filtering and de-sampling, and also computes the estimated
Development of an in-shoe pressure-sensitive device for gait analysis
S.M.M. De Rossi, Student Member, IEEE, T. Lenzi, Student Member, IEEE, N. Vitiello, Student
Member, IEEE, M. Donati, A. Persichetti, F. Giovacchini, F. Vecchi, Member, IEEE, M.C. Carrozza,
Figure 1. Base sensitive element of the pressure-sensitive insole. (a)
Silicone cover of the sensitive element. (b) Representation of the
transduction principle. When a load is applied (on the right), the
consequent deformation of the structure occludes the light path, and
results in a diminished output from the photodiode.
978-1-4244-4122-8/11/$26.00 ©2011 IEEE5637
center of pressure and ground reaction force. All data is
transmitted wireless at a 100 Hz frequency. The autonomy
of the device in continuous working conditions is about 7
A. Sensor Array
1) Sensitive Element
The pressure sensitive element presented in this work is
based on a tactile sensing technology relying on an opto-
electronic transduction principle, conceptually analogous to
the one presented by the authors in   .
This sensitive element is made of a LED light emitter facing
a photodiode as light receiver. These optical components are
covered by a shell made of opaque silicone, which deforms
under the effect of external force.
The sensitive element is shown in Figure 1(a). It can be seen
that the sensor cover has the shape of a square pyramidal
frustum, with the base having a side of 12 mm, and the top
face of 10mm. This contact surface provides a 1cm2
resolution to the estimation of pressure distribution and
center of pressure. The cover is made of a black-dyed
Figure 1(b) represents the transduction principle of the
sensitive element: A load applied to the top face of the cover
causes a deformation, and lowers a silicone ‘curtain’ which
obstructs the light path from the LED (on the right) to the
photodiode (on the left). The sensor thus works as an inverse
force vs. voltage transducer having high vertical sensitivity
and low tangential sensitivity. Finally, the specific design of
the cover ensure a maximum loading range of 1MPa before
damaging the device.
Figure 2(a) and (b) show the characterization curve of the
sensor and its structural behavior, respectively. Figure 2(a)
shows the typical relationship between the applied force and
the output voltage, were data was acquired in quasi-static
conditions. A non-amplified dynamic output range of about
1.1V, corresponding to a 50N load on the sensor (500kPa),
was observed. Notably, the maximum measurable load
generates a vertical deformation of about 1.8mm. The
resulting stiffness (about 28N/mm) adds compliance to the
insole, increasing its comfort and wearability.
The curve in Figure 2(a) can be approximated with a third-
order polynomial interpolant (see also next Section) which
was found to be the best compromise in terms of complexity
and goodness of fit. Further trials show that due to the
rigidity of the silicone, no significant dynamic effects can be
observed on the output (see ). Figure 2(b) portraits the
structural behavior of the sensor, which shows no significant
static hysteresis, and an average stiffness of 28.7N/mm.
2) Sensor Board
The sensor board comprises 64 sensitive, placed as in Figure
3(a). The sensitive area covers about 80% of the foot-insole
contact region, leaving the Medial Arch area free from
sensors, and available for the integration of the electronics.
The Medial Arch area is known to be the least stressed area
of the foot plant, in terms of peak pressure, mean peak
pressure and average pressure during a gait cycle .
The covered area contributes to about 99% of the total
pressure-time integral  during a gait cycle and to 100%
of the pressure peaks during the gait in healthy adult subjects
. Compared to devices with fewer sensors
(), this insole, covering the whole interaction area,
allows a direct evaluation of the center of pressure position
and of the total loading force. Moreover, it permits to
estimate pressure peaks with a much higher accuracy.
The sensor board is based on a thin PCB, housing, for each
sensor, a power and ground wire, and a signal wire. The 64
signal channels, and the two common power and ground
channels are routed through two flat cable connectors,
placed in the Lateral Arch area. The cable connectors are not
in contact with the foot, being positioned at a height lower
than the sensors. Figure 3(b) shows the final appearance of
the sensor board, complete with silicone cover for the 64
B. In-Shoe Electronics
As shown in Figure 3(a), we used the space under the
Medial Arch area to house an in-shoe electronic board for
signal acquisition, processing and wireless communication.
Based on anatomical measurements, we determined the
maximum encumbrance of the electronics board so as to not
impact the comfort of the wearer.
The board, shown in Figure 4(a), comprises analog-digital
converters (ADCs), and a microcontroller to perform signal
Figure 2. Characterization of the sensitive element. (a) Shows the force
vs. Voltage output characterization, while (b) shows the structural
characterization of the sensor.
Figure 3. Sensor board. (a) Disposition of sensors, on-board electronics
and connectors. (b) Picture of the sensor array complete with silicone
cover and connectors.
processing. The board is connected to a Bluetooth
receiver/transmitter (RoboTech s.r.l., Pisa, Italy) on a UART
socket. Power to the board and to the sensors is supplied by
a Li-poly 700mAh battery operating at 3.6V (25x25x10mm).
The board acquires the 64 signals at a 1.8kHz frequency
through four 16-channels 14-bit ADCs. Signals are low-pass
filtered and de-sampled to 100Hz. Each digitalized voltage is
used to determine the pressure on the corresponding
sensitive element through a third-order polynomial function
(see Section III). The 64 pressure values are used to
determine the ground reaction force on the insole and the
position of the center of pressure. The pressure map (64
values), ground reaction force and center of pressure are sent
at 100Hz frequency through a Bluetooth transmitter.
The power consumption of the device (sensors and
electronics) is about 100 mA at 3.6-3.7V. The on-board
battery can power the unit for up to 7 hours in continuous
As shown in Figure 4(b) the in-shoe electronics is protected
by a rigid plastic cover, providing protection from impacts,
sweating and humidity.
C. Remote Electronics and Communication
The data from the pressure-sensitive insole can be received
by any remote device (PC, tablet or smarphone) equipped
with a Bluetooth receiver. A 921.6Kbit/s connection is
required to sustain a 100Hz communication rate. Each data
packet sent from the insole includes a timestamp to verify
transmission reliability on the remote host.
The remote host can command the device to initiate (or
stop) a data acquisition and communication sequence.
Commands can be sent to require force de-offsetting, battery
level as well as for debugging information.
D. Software Modules
We developed a graphical user interface (GUI) to allow
for data monitoring and logging, as well as to remotely
command the device. The interface was written using
National InstrumentsTM Labview® 2010, and is shown in
Figure 5. The GUI gives a real-time monitoring of the foot
pressure map, of the ground reaction force trend, and of the
position of the center of pressure. Taking into account that
no computation is required from the remote host side (all
calculations are done on-board), similar interfaces could be
developed for less powerful architectures like tablets or
III. CHARACTERIZATION OF THE PRESSURE-SENSITIVE
While all the sensors of the array have identical electronic
components and silicone covering, the imprecision of the
silicone molding and polymerization process, along with the
variability in the characteristics of LEDs and photodiodes
require a separated characterization for each pressure-
Characterization was performed using a three-axial
robotic platform able to provide controlled loads or
deformations to any given position. Each sensitive element
was compressed to a maximum reaction force value of 50N,
and then unloaded at a constant speed of 5mm·min-1 to give
a quasi-static condition. Three loading-unloading cycles
were performed for each sensitive element.
The force vs. voltage behavior was characterized for each
element, with a result similar to that depicted in Figure 1(a).
The curve was fitted with a third-order polynomial, unique
for each sensitive element. Figure 6(a) and (b) show the
results of the characterization of each element in terms of
goodness of fit (R2 and RMSE). It can be seen that a third
order polynomial approximant introduces errors between 2
Figure 4. In-shoe electronics. (a) without and (b) with its cover that is
housed inside the shoe.
Figure 5. Graphical User Interface for the remote host.
Figure 6. Characterization results. On the top, the RMSE and on the
bottom, the R2 value for each polynomial fitting.
and 5% of the full-scale range, and that goodness of fit is
high (minimum R2 95%).
The polynomial coefficients are loaded on the firmware of
the processing electronics, and allow for force/pressure
estimation on each sensitive element to be performed on-
The final appearance of the insole is shown in Figure 7(a)
and (b). As a preliminary validation of the device, we
performed a walking experiment on a healthy subject
wearing the insole inside a sport shoe. Data relative to 5
steps were recorded. Figure 8 shows the center of pressure
position profile during the stance phases of the 5 steps.
We developed a new in-shoe device to monitor plantar
pressure consisting in a matrix of 64 pressure-sensitive
elements, integrated with in-shoe conditioning electronics
and battery. The device can replace the insole of
commercially available shoes, without interfering with the
normal gait, and allowing the user to wear his/her own shoe.
The device allows to monitor shoe-foot interaction with a
high temporal (100Hz) and spatial (1cm2) resolution,
compared to state-of-the-art in-shoe devices. The on-board
electronics acquires and digitalizes the signals, computes the
pressure map in real-time using proper calibration curves,
,evaluates the center of pressure, and the ground reaction
force and finally transmit all these information through
Bluetooth. This allows also low-computational-power
remote devices (like smartphones) to provide visualization
and logging of data from the insole.
These features, along with its long-time autonomy, make
the proposed device very useful to monitor the gait and for
the assessment of quality of the walk in healthy subjects.
Most importantly, this device can used for long-term
monitoring of gait (all-day monitoring), thanks to its
simplicity of use and to its versatility.
 M. Chen, B. Huang, Y. Xu, and A. Motivation, “Intelligent shoes for
abnormal gait detection,” IEEE ICRA, May. 2008, pp. 2019-2024.
 M.N. Orlin and T.G. McPoil, “Plantar Pressure Assessment,” Physical
Therapy, vol. 80, 2000, pp. 399 - 409.
 M.J. Mueller, “Application of plantar pressure assessment in footwear
and insert design,” The Journal of orthopaedic and sports physical
therapy, vol. 29, Dec. 1999, pp. 747-55.
 A. Gioftsidou, P. Malliou, G. Pafis, A. Beneka, G. Godolias, and C.N.
Maganaris, “The effects of soccer training and timing of balance
training on balance ability,” European journal of applied physiology,
vol. 96, Apr. 2006, pp. 659-64.
 J. Burnfield, “The influence of walking speed and footwear on plantar
pressures in older adults,” Clinical Biomechanics, vol. 19, Jan. 2004,
 R.M. Queen, B.B. Haynes, W.M. Hardaker, and W.E. Garrett,
“Forefoot loading during 3 athletic tasks.,” The American journal of
sports medicine, vol. 35, Apr. 2007, pp. 630-6.
 M. Saito, K. Nakajima, C. Takano, Y. Ohta, C. Sugimoto, R. Ezoe, K.
Sasaki, H. Hosaka, T. Ifukube, S. Ino, and K. Yamashita, “An in-shoe
device to measure plantar pressure during daily human activity,”
Medical Engineering & Physics, Feb. 2011.
 M.C. Carrozza, A. Persichetti, C. Laschi, F. Vecchi, R. Lazzarini, P.
Vacalebri, and P. Dario, “A Wearable Biomechatronic Interface for
Controlling Robots with Voluntary Foot Movements,” IEEE/ASME
Transactions on Mechatronics, vol. 12, Feb. 2007, pp. 1-11.
 L. Shu, T. Hua, Y. Wang, Q. Qiao Li, D.D. Feng, and X. Tao, “In-
shoe plantar pressure measurement and analysis system based on
fabric pressure sensing array.,” IEEE Transactions on Information
Technology in Biomedicine, vol. 14, May. 2010, pp. 767-75.
 S.M.M. De Rossi, N. Vitiello, T. Lenzi, R. Ronsse, B. Koopman, A.
Persichetti, F. Giovacchini, F. Vecchi, A.J. Ijspeert, H. van der Kooij,
and M.C. Carrozza, “Soft artificial tactile sensors for the measurement
of human-robot interaction in the rehabilitation of the lower limb,”
Proceedings of the 32nd Annual International Conference of the IEEE-
EMBS , Jan. 2010, pp. 1279-82.
 S.M.M. De Rossi, N. Vitiello, T. Lenzi, R. Ronsse, B. Koopman, A.
Persichetti, F. Vecchi, A.J. Ijspeert, H. van der Kooij, and M.C.
Carrozza, “Sensing Pres sure
Exoskeleton Physical Human-Machine Interface,” Sensors, vol. 11,
2010, p. 207–227.
 T. Lenzi, N. Vitiello, S.M.M. De Rossi, A. Persichetti, F. Giovacchini,
S. Roccella, F. Vecchi, M.C. Carrozza, “Measuring human-robot
interaction on wearable robots: A distributed approach”, Mechatronics
 A. Bryant, P. Tinley, and K. Singer, “Normal Values of Plantar
Pressure Measurements Using the EMED-SF System,” Journal of the
American Podiatric Medical Association, vol. 90, 1999, pp. 295-299.
 S. Lundeen, K. Lundquist, M.W. Cornwall, and T.G. McPoil, “Plantar
pressures during level walking compared with other ambulatory
activities,” Foot & Ankle International, vol. 15, Jun. 1994, pp. 324-8.
Distribution on a Lower-Limb
Figure 8. Trajectory of the center of pressures during 5 steps.
Figure 7. Overview of the device. (a) outside of the shoe and (b) inside