COMBINED APPLICATION OF FBG AND PZT SENSORS FOR PLANTAR
PRESSURE MONITORING AT LOW AND HIGH SPEED WALKING
R. Suresha, C. Singhb, N. Kaurc, J. Haod, S. Bhallac*
a
Centre for Biomedical Engineering, IIT Delhi, Hauz Khas, New Delhi 110016, India
(currently at Bhaba Atomic Research Centre Visakhapatnam, India)
b
Bechtel India Pvt. Ltd., 244-245, Udyog Vihar, Phase IV, Gurgaon 122015, India
c
Department of Civil Engineering, IIT Delhi, Hauz Khas, New Delhi 110016, India
d
Institute for Infocomm Research, 1 Fusionopolis Way, #21-01 Connexis (South Tower) Singapore 138632
*Corresponding author, Email: [email protected], Phone: (91)-11-2659-1040 Fax: (91)-11-2658-1117
ABSTRACT
This paper presents the combined application of the Fibre Bragg Grating (FBG) and the lead
zirconate titanate (PZT) piezoceramic sensors for plantar pressure monitoring during walk at
low and high speeds. For fabrication of the pressure sensors, the FBGs are embedded within
layers of carbon composite material and stacked in an arc shape. From this embedding
technique, average pressure sensitivity of 1.3 pm/kPa and resolution of nearly 0.8 kPa is
obtained. These sensors are found to be suitable for measuring the static and the low-speed
walk generated foot pressure. Simultaneously, PZT patches of size 10×10×0.3 mm are used
as sensors, utilizing the d33 (thickness) coupling mode. A sensitivity of 7.06 mV/kPa and a
pressure resolution of 0.14 kPa is obtained from these sensors, which are found to be suitable
for foot pressure measurement during high speed walking and running. Both types of sensors
are attached to the underside of the sole of commercially available shoes. In the experiment, a
healthy male subject walks over the treadmill wearing the fabricated shoes at various speeds
and the peak pressure is measured using both the sensors. Commercially available low-cost
hardware is used for interrogation of the two sensor types. The test results clearly show the
feasibility of the FBG and the PZT sensors for measurement of plantar pressure. The PZT
sensors are more accurate for measurement of pressure during walking at high speeds. The
FBG sensors, on the other hand, are more accurate for static and quasi-dynamic (slow
walking) conditions. When present in combination, the two sensors can enable measurements
1
ranging from static to high speed conditions. Both the sensor types are rugged, small sized
and can be easily embedded in commercial shoes and enable plantar pressure measurement in
a cost-effective manner.
Keywords: PZT, FBG, Sensors, plantar pressure, diabetic foot, pressure, d33 coupling
2
INTRODUCTION
Plantar pressure monitoring is required in various pathological conditions such as diabetes
and gonarthrosis [1,2]. People with these pathological conditions are very likely to experience
foot problems associated with higher plantar pressure because of the loss of sensation caused
by nerve damage and circulatory problems associated with the disease. Skin ulceration is a
major complication for diabetic patients [3]. Diabetic foot problems are of significant clinical
and economic importance since they not only carry the risk of the amputation of the limb but
may even be life threatening [4]. It has been estimated that lower leg amputations in diabetic
patients account for more than 50% of all amputations worldwide and that two thirds of these
arise as a result of traumatic ulceration of the foot [5]. Elevated plantar foot pressure has been
identified as a major risk factor for foot ulceration in the diabetic persons. Hence, an early
detection of any abnormal foot pressure in a diabetic patient is essential to prevent the
development of foot ulceration [6]. The high cost of treating neuropathic ulcers, their impact
on the quality of life of the patient and the increased risk of lower limb amputation emphasize
the importance of preventive measures through timely diagnosis. Therefore, prevention of the
irreversible complications of diabetic neuropathy is a major priority in patient care. Plantar
pressure monitoring can be an important tool for understanding of the biomechanics of the
foot in various sports and locomotive activities such as running and walking and also in
footwear design [7, 8]. Mapping the pressure distribution on the plantar surface during
walking can timely indicate the adverse influence of the footwear, if any, on the foot. Various
aspects of plantar pressure monitoring, such as the effect of body mass index (BMI), gender,
foot size and arch have been reported in literature [9-16]. Another emerging area of
application of the plantar pressure monitoring is in biometric and human identification [17,
18].
3
Two main configurations have so far been reported in the literature for plantar pressure
monitoring: (i) plate type, and (ii) in-sole type. In the plate type system, pressure sensors are
embedded on a flat surface. This system can be used for both static and dynamic plantar
pressure monitoring. However, its application has so far been restricted to laboratory only.
Another constraint of this system is that several familiarization trials may be required by the
patients to ensure proper contact of the foot at the sensing area, to guarantee a normal gait
[19, 20]. As such, the results may not be dependable on first timers. In the in-sole system
[21], on the other hand, the pressure sensors are embedded within the footwear, enabling
pressure monitoring directly beneath the foot [22]. A feasible system capable of measuring
the foot pressure accurately needs to be tested under the same conditions encountered in an
actual foot, and under both quasi static and dynamic conditions, with the shoe worn in the
foot. The in-sole system, therefore, is particularly useful for designing appropriate footwear
as per the individual patient’s requirement.
Both these types of the plantar pressure monitoring systems require pressure sensors with
good sensitivity, linear response, repeatability, low hysteresis, low cost and ease of handling.
Both types of configurations have been reported in literature based on capacitive, piezoresistive elements, micro-electromechanical systems (MEMS) and opto-electronic sensors
[23-26]. However, most of the applications have drawbacks such as limited pressure
monitoring range [26] and susceptibility to electromagnetic interference (EMI), which may
not only reduce the signal to noise ratio (SNR), demand additional post measurement data
processing or may lead to inaccurate pressure measurement. Although modern MEMS based
sensors offer compact wireless monitoring, however, they are prone to EMI and the power
consumption level is much/ higher than the available industry standard. At the same time,
most of the reported sensors are somewhat too exorbitant for real-life implementation.
4
Besides the above lacurae, in general, most of the currently deployable sensors for plantar
pressure measurement show up to 33% margin of error, which is beyond the acceptable range
[12]. This has made very difficult for the medical experts to arrive upon the exact threshold
pressures beyond which ulceration is major risk. From the above literature review, it is
evident that the development of a simple and low-cost pressure monitoring system for
accurate measurement of plantar pressure is a crucial multidisciplinary research area.
During the recent years, smart materials have arrived at the forefront of sensing technologies.
Amongst smart materials, Fibre Bragg grating (FBG) sensors and piezoelectric ceramic lead
zirconate titanate (PZT) transducers have been the most sought after materials. Optical fiber
sensors, especially FBG sensors, are well known for their immunity to EMI, multiplexing
capability and providing wavelength encoded strain information with good level of accuracy
[27-30]. The PZT transducers, on the other hand, are well known for cost effectiveness, high
sensitivity, fast dynamic response and capability for high frequency dynamic strain
measurement [31, 32], though they are not immune from EMI. Both the sensors have been
extensively utilized in structural health monitoring (SHM) of civil, mechanical,
biomechanical and aerospace structures [33-37].
This paper presents the combined application of the FBG and the PZT sensors for plantar
pressure monitoring. Easy measurement in field environment, compactness, linearity and
negligible hysteresis, minimal hardware requirement and minimal signal processing are the
main advantages offered by the FBG sensors. The FBG sensors have been shown to possess
greater life expectancy. However, the cost of interrogation system for dynamic measurement
is still high especially where a sampling rate above few Hz is warranted, and may be very
exorbitant if it exceeds 1 kHz. Similarly, the PZT patches offer high dynamic strain
5
sensitivity (even for frequencies in 1-10 kHz range) and at the same time cost- effectiveness
with regard to sensors as well as the instrumentation. Compared to FBG sensors, the data
processing requirements are minimal for the PZT sensors. However, they are not suitable for
measuring static pressure response. Both types of sensors can thus complement one another,
the FBG for static/quasi static pressure and the PZT sensors for dynamic pressure
measurement. Due to their small size and negligible weight, both the FBG and the PZT
sensors are not likely to affect the gait parameters during plantar pressure monitoring. Hence,
they can enable a direct and easy measurement of the pressure along the plantar surface.
In the proposed measurement scheme, the PZT and the FBG pressure sensors are attached
directly to the underside of the sole of a ready to wear shoe at the appropriate locations of
forefoot and hind foot (heel: medial heel). The pressure values have been measured at these
locations during treadmill walking by a healthy male subject at various speeds. In the
following sections, first, a brief review of the PZT/FBG sensor technologies is first presented,
followed by specimen preparation, conduct of the experimental study on the treadmill, results
and inferences.
FABRICATION OF FBG BASED PLANTAR PRESSURE SENSOR
The working of a FBG sensor is based on the principle that a periodic variation of refractive
index occurs in the core of a single-mode fibre, which renders it a wavelength selective
mirror. When light from a wide band source is launched into the fibre, a particular
wavelength, which is known as the “Bragg wavelength”, is reflected back, given by [38]
λb = 2 neff Λ
(1)
where λb is the reflected Bragg wavelength, neff the effective index of the fibre core and Λ the
grating periodicity. Any external perturbation that results in an axial strain in the fibre can
6
alter the grating periodicity and/or the effective refractive index due to strain- optic effect,
thus shifting the Bragg wavelength, which provides a measure of the applied perturbation.
The shift in the Bragg wavelength with strain and temperature can be expressed as [39]
b/b= [1-0.5 neff2 {12- (11 + 12)}] + T
(2)
where  is the axial strain in the fibre,  the Poison’s ratio of the fibre material, ρij the strainoptic coefficients,  the thermo-optic coefficient, T the temperature change and neff the
refractive index of the core. As given by Eq. (2), the reflected Bragg wavelength changes with
the variation of either the strain and/or the temperature. Thus, the wavelength shift from a
single FBG cannot explicitly measure the strain in the presence of temperature variation. To
overcome this problem, temperature compensation using two fibres is generally resorted to.
The approach involves instrumenting two sensors, which have different strain (K1, K2) and
temperature sensitivities (KT1, KT2). On the monitored component one sensor is exposed to
both the strain and temperature fields and the other to temperature variation only, by keeping
it unbonded to the component in question. Then, a matrix equation can be obtained as [38]
𝐾
∆𝜆
( 1 ) = ( 𝜀1
𝐾𝜀2
∆λ2
𝐾𝑇1 𝜖
)( )
𝐾𝑇2 𝑇
(3)
where, the subscripts 1 and 2 refer to FBG 1 and 2, FBG 1 bonded to structure (exposed to
strain and temperature) and 2 not bonded to the structure (hence exposed to temperature
only). Solving this equation, strain and temperature can be obtained from the wavelength
shifts of the two FBG sensors. At constant temperature (or after thermal compensation), the
relation between the wavelength shift and the pressure (P) can be simplified as [40]
 1  2  neff 2

1  2 212  11 P
P  B 

E
2E


(4)
where E is the Young’s modulus of the martial of the fibre. Since, the wavelength is an
absolute parameter (does not depend on the intensity), any fluctuation in the source intensity
7
or loss in the sensor network (such as coupling loss) will not cause any error in the recorded
wavelength. Also, as each FBG works in a narrow wavelength range, several FBGs can be
multiplexed in a single fibre, thereby reducing the required hardware rendering the sensor
network compact and easy to handle.
In the present study, a standard telecommunication single mode optical fibre with a 250
micron acrylate coating was used for FBG sensor fabrication. The fibre was first hydrogen
loaded at room temperature for about two weeks to enhance photosensitivity. After stripping
off the acrylate coating of a short section of about 10-20 mm, an FBG of 5 mm length was
imprinted in the fibre core using the standard phase mask technique. The FBG wavelength
was stabilized by annealing the fibre for 24 hours at 100oC, after which, each FBG was
embedded into five layers of carbon composite material (CCM) which were stacked to form
an arc shaped sensor module, as shown in Fig. 1 for higher sensitivity [41]. After embedding,
the sensor module was cured in vacuum oven.
CCM was chosen as the embedding material due to its high strength to weight ratio, excellent
corrosion resistance and elasticity and ease to be moulded into complex shapes for sensor
fabrication. After curing, the resulting sensor module was 0.625 mm thick and 5 mm wide,
with the effective length (a) of 40 mm and of the arc height (Dmax) of 2.2 mm (see Fig. 1).
The arc shape embedding ensures a higher sensitivity of about 1.3 pm/kPa as compared to a
sensitivity of 0.5 pm/kPa obtained with linear embedding [42].
FOOT PRESSURE MEASUREMENT USING PZT PATCHES
The phenomenon of piezoelectricity occurs in certain classes of noncentro-symmetric
crystals, such as quartz, in which electric dipoles (and hence surface charges) are generated
8
due to mechanical deformations. The same crystals also exhibit the converse effect; that is,
they undergo mechanical deformations when subjected to electric fields. The constitutive
relations for piezoelectric materials for 1D interaction, through d33 coupling, as illustrated in
Fig. 2, can be expressed as [36]
T
D3   33
E3  d 33T3
S3 
T3
 d 33 E3
YE
(5)
(6)
where, S3 is the strain in the principal direction ‘3’ (thickness), T3 the corresponding stress,
D3 the electric displacement over the PZT patch and d33 the piezoelectric strain coefficient
providing coupling between the mechanical strain (along axis ‘3’) and electric field (also
along axis ‘3’). Y E  Y E (1  j ) is the complex Young’s modulus of elasticity of the PZT
T
T
  33
(1  j ) the complex electric permittivity in
patch (at constant electric field) and  33
direction ‘3’ (at constant stress), with j   1 , and,  and  respectively denoting the
mechanical loss and the dielectric loss factors of the PZT material. Equation (5) is used in
sensing applications and Equation (6) in actuation applications of the piezo-electric materials.
If a PZT patch surface bonded on a structure is desired to be used as a sensor only (with no
external electric field across its terminals, i.e. E3 = 0), its governing sensing equation Eq. (5)
will be reduced to
D3  d 33T3
(7)
From the theory of parallel plate capacitors, the charge density can also be expressed as
D3 
T
 33
V
9
h
(8)
where V is the potential difference across the terminals of the PZT patch of thickness h.
Combining Equations (7) and (8), the voltage generated across the terminals of the PZT patch
can be expressed in terms of the normal stress (or pressure) T3 acting on the patch as [32]
d h
 1 
T
V   33 T3  
K  3
 T 
 p
 33 
(9)
where kp, which depends only on the PZT parameters, is the voltage to pressure conversion
constant of the PZT patch. The output voltage can be easily measured by an oscilloscope
(with or without conditioning circuit) or directly using the modern digital multimeters, such
as Agilent 34411A [43]. A pressure resolution of 0.89 kPa can be obtained using this d33
coupling mode for a PZT patch conforming to grade PIC 151 [44].
In general, for most SHM applications involving strain measurements [32] so far reported,
the PZT transducers have utilized the d31 (along the axis) mode. However, the foot pressure
sensor proposed in this paper utilizes the d33 coupling (thickness) mode. Neville et al. [45]
reported a pressure sensor utilizing d33 coupling effect in a piezoelectric thin film based
transducer and measured current for pressure calculation. However, use of piezoelectric thin
film increased the cost and reduced the ruggedness of the sensor system. Measurement of
current in place of voltage may also proved to be practically troublesome. The d33 coupling
mode has not been explored very well with low cost commonly available thin sheet type lead
zirconate titanate (PZT) patches. Voltage across the surface can be accurately measured using
digital multimeter or oscilloscopes, and then converted into pressure using Equation (9).
thereby making the system low cost and commercially viable. This is the main consideration
of using the d33 mode in the present paper.
10
INSTRUMENTATION DETAILS
In this study, standard commercial sports shoes of size nine (UK) were used for
instrumentation with the proposed FBG and PZT sensors. The FBG and the PZT pressure
sensors were attached to the sole of the shoes; at the locations of the forefoot and the heel
(medial heel), as shown in Fig. 3. Two grooves were first made in the sole at the desired
locations. For the left footwear, fabricated FBG sensors were placed in these grooves.
Araldite epoxy adhesive was applied at the ends to secure the sensors. PZT patches were also
bonded to the sole adjacent to the FBG sensors using the epoxy adhesive, as shown overall in
Fig. 3(a) and illustrated in detail Fig. 3(b).
For sensors attached on the left footwear, the FBGs had the Bragg wavelengths of 1548.858
nm and 1552.729 nm respectively. One additional FBG sensor (not shown in the figure) was
placed in contact with the shoe (but not bonded to it) to facilitate thermal compensation at the
time of the experiment. All the FBG sensors were calibrated before attachment to the shoe
sole. Figs. 4 (a) and (b) show the variation of the thermally compensated wavelength shift of
the individual sensors with the applied pressure for the left footwear [34]. Due to the linear
variation of the wavelength shift with pressure, the curves can be easily employed for
pressure measurement.
On both the shoes (left/ right), two PZT patches were instrumented, one on the heel and the
other on the forefoot area as shown in Fig. 3. The PZT patches were of size 10x10x0.3 mm
and confirmed to grade PIC 151 [44]. The patches were bonded using a thin layer of the two
part Araldite epoxy adhesive, as shown in Fig. 3. After 24 hours of room temperature curing,
wires were soldered on the two patches, followed by the application of an additional outer
layer of epoxy adhesive for protection purpose. The PZT sensor patches were thus in adjacent
11
locations to the FBG sensors. However, as can be seen from Figs. 3(a) and (b), whereas the
FBG sensors would directly touch the ground as the subject walks, the PZT patches would
directly not be in contact but rather under a cover of about 5 mm thick epoxy layer.
EXPERIMENTAL DETAILS OF FOOT PRESSURE MEASUREMENT STUDY
To verify the feasibility of the FBG and the PZT sensors in measuring the plantar pressure,
experiments were performed on human subject using a treadmill. Healthy male subject of
BMI 20.5 volunteered for this experiment. Informed consent was taken from the subject
before experimentation. The apparatus was set up with the subject wearing the shoes as
shown in the Fig. 5. The FBG sensors were monitored using optical sensor interrogator
SM125 interrogator from Micron Optics [46]. The two FBGs were bonded on the left
footwear and the one for thermal compensation were connected to channels 1, 2 and 3 of the
interrogator, respectively. A pressure resolution of 0.89 kPa and a sampling rate of 2Hz was
available during the experiment. The four PZT patches were connected to the four channels
of the TDS 2004B oscilloscope [47], where a pressure resolution of 0.14 kPa and a sampling
rate of 1 kHz was available.
The experiments were conducted at different speeds to account for: normal walking (1-3
kmph), brisk walking (5kmph) and running (7kmph). Treadmill speed controller was used to
set the desired speed and the subject walked over the treadmill for 5 minutes at each walking
speed. The oscilloscope and SM125 data were recorded continuously for a particular speed.
The experimental results are discussed in the next section.
12
EXPERIMENTAL RESULTS AND DISCUSSION
The wavelength shifts of the FBG sensors embedded in the forefoot of the left shoe at the
speeds of 1 kmph and 3 kmph are shown in the Fig. 6. It is observed that as the speed
increases, the number of peaks in given time interval also increases. Each peak corresponds
to the instance when the foot touches the ground. Similar curves were obtained for both the
FBG sensors at other walking speeds. The FBG wavelengths for left foot were recorded
continuously and converted to pressure values using the calibration curves shown in Fig. 4.
Static pressure for standing on both feet as well as standing only on left foot was also
measured using the FBG sensors. Table 1 summarizes the pressure values for the FBG and
PZT sensors on the left foot at various walking speeds, including the standing (purely static)
condition.
For the PZT sensors, at each walking speed, the oscilloscope displayed a plot of the voltage
versus time. Fig. 7 shows a typical oscilloscope plot obtained for the four PZT sensors
attached at the four locations at a walking speed of 3 kmph. Similar plots were obtained at
other speeds. Each peak in the plot corresponds to the instant the concerned portion of the
foot touches the ground, i.e. the stance phase. Due to somewhat weak signals from the heel
(right), the data was not considered. The voltage values obtained were converted to the
pressure values using Eq. (9) with the constant kp equal to 141.6 KPa/V. Figs. 8 and 9 shows
the converted pressure variation with time for the left and the right foot and heel at walking
speed of 5 kmph and 7 kmph, respectively. The results are summarized in Table 1 alongside
the FBG sensors. The measured pressures can be observed to be lower than those of the FBG
sensors, especially at lower speeds. The possible cause is the dispersion of the load due to a
thick layer of epoxy covering the patch. It is observed that both the forefoot and the heel
pressures increase with the speed of walking. As observed in Figs. 8 and 9, the trace of the
13
pressure against time has a well documented double hump form [48]. The first foot makes
contact with the treadmill at the heel when a very short duration transient occurs. The foot
progressively rolls into contact with the floor and the pressure builds up to its first sustained
peak, which is higher than the average pressure. After a fall of this weight, the force reaches a
second peak of similar magnitude of the first, before rolling off at the end of the step. This
observation is much more distinct and realistic than the FBG sensors (see Fig. 6 for
comparison), due to the fact that the sampling rate is 1 kHz for PZT sensors than 2 Hz in the
case of the FBG sensors.
Both the FBGs as well as the PZT patches at forefoot and the heel were able to measure
pressure values at various walking speeds. The peak pressure at both the locations changes
with the change in speed of walking. The plot of peak pressure variation with walking speed
is shown in Fig. 10 for both the sensors. From this figure, looking at the PZT based trend, it
can be observed that the pressure at forefoot as well as heel increases with the increase in the
walking speed. This is expected, due to higher dynamic effects associated with high speed
walking or running. However, with regard to the FBG sensors, somewhat opposite nature of
observation holds, where the foot pressure seems to decrease with speed. The possible reason
is the low sampling rate (2Hz), which led to the peaks being missed out. From this
consideration, the FBG based measurements may not be reliable beyond 2kph when the
frequency of stepping is typically higher than 1Hz.
Looking at the PZT results, somewhat greater increase in the pressure is observed at the heel
area as compared to the forefoot. This could be due to the specific posture adopted during
walking, resulting in higher pressure applied in the heel area. The pressures measured by the
two sensor types on the left heel are comparable. However, the pressure values obtained at
14
the left forefoot by PZT sensor are lower as compared to those obtained by the FBG sensor.
This could be due to the specific placement of the PZT sensor (not exactly same at same
location as the FBG) on the sole of the left footwear, and also due to a cover of over 5mm of
epoxy adhesive, which led to the phenomenon of dispersion. Further, as the PZT sensors can
only measure dynamic pressure, no pressure values are obtained for standing position using
PZT sensors.
CONCLUSIONS
This paper has demonstrated the proof-of-concept experimentation to measure foot pressure
distribution using PZT patches and FBG sensors. The PZT sensors carry out the measurement
using d33 coupling. A pressure resolution up to 0.89 kPa has been obtained with PZT sensors.
The FBG sensors employ a special arch type configuration for higher sensitivity. The
pressure values measured by the two sensors are comparable in nature. The pressures at both
the locations (forefoot/heel) are found to increase with increase in the walking speed for PZT
sensors. The PZT patches, which can measure the pressure with a high sampling rate
(typically few kHz per second), can provide near real time pressure measurement and are
most suitable for fast walking speeds, typically higher than 3 kmph. The FBG sensors are
suitable for both static as well as low frequency dynamic measurements typically less than 3
Kmph. However, the cost of high frequency measurement system for FBG is still very high
as compared to cost incurred during dynamic monitoring by PZT patches.
In the present study, the shoes samples were prepared with only two sensors at forefoot and
heel. However, typically, there are about 15 pressure points along the plantar surface [21].
Shoes samples with more number of sensors are required before the test shoes can be utilized
for plantar pressure monitoring for clinical application. For practical applications, certain
15
aspects, such as minimum/ maximum sensors size, spatial resolution and contact area
between the sensor and the foot need to be well studied and understood. Further experiments
are in progress entailing more detailed foot pressure distribution with larger number of
sensors, and the results will be published subsequently.
ACKNOWLEDGEMENTS
Dr. Rupali Suresh gratefully acknowledges the financial support through Fast Track Young
Scientist Scheme from Department of Science and Technology, Government of India. Thanks
are also due to Ms Mansi Dhiman, B. Tech. final year student, for assistance in carrying out
preliminary experiments on treadmill.
16
REFERENCES
[1] Formosa C, Gatt A and Chockalingam N. Diabetic foot complications in Malta:
Prevalence of risk factors. The Foot. 2012; 22: 294.
[2] Motooka T, Tanaka H, Ide S, Mawatari M and Hotokebuchi T. Foot pressure distribution
in patients with gonarthrosis. The Foot. 2012; 22: 70.
[3] Levin ME, O’Neil LW, Bowker JH. The diabetic foot, 5th edition, St Louis, Mosby
yearbook. 1993; 31.
[4] Periyasamy R, Anand S, and Ammini AC. Prevalence of standing plantar pressure
distribution variation in north Asian Indian patients with diabetes mellitus: A study to
understand ulcer formation. Proceedings of the Institution of Mechanical Engineers, Part H:
Journal of Engineering in Medicine. 2013; 227 (2): 181.
[5] Larsson J, Apelquist J. Towards lessamputations in diabetic patients- Incidence cause,
cost, treatment and prevention- a review. Acta Orthopaedica Scandinavica. 1995; 66: 181.
[6] Boulton AJM. The importance of abnormal foot pressure and gait in the causation of foot
ulcers. In: Conner H, Boulton AJM, Ward JD editors, Chichester. 1987.
[7] Mueller M. Application of plantar pressure assessment in footwear and insert design.
Journal of Orthopaedic and. Sports Physical Therapy. 1999; 29: 747.
[8] Queen R, Haynes B, Hardaker W and Garrett W. Forefoot loading during 3 athletic tasks.
The American Journal of Sports Medicine. 2007; 35: 630.
[9] Hills AP, Henni EM, Mcdonal M and Bar-Or O. Plantar pressure differences between
obese and nonobese adults: a biomechanical analysis. International Journal of Obesity. 2001;
25: 1674.
[10] Ho IJ, Houl YY, Ynag CH, Wu W L, Chen SK and Guo LY. Comparison of plantar
pressure distribution between different speed and incline during treadmill jogging. Journal of
Sports Science and Medicine. 2010; 9: 154.
17
[11] Kellis E. Plantar pressure distribution during barefoot standing, walking and landing in
preschool boys. Gait and Posture. 2001; 14: 92.
[12] Periyasamy R, Mishra A, Anand S and Ammini AC. Foot pressure distribution variation
in pre-obese and non-obese adult subject while standing. The Foot. 2012; 22: 276.
[13] Periyasamy R and Anand S. The effect of foot arch on plantar pressure distribution
during standing. Journal of Medical Engineering and Technology. 2013; 37 (5): 1.
[14] Lee YC, Lin G and Wang MJ. Evaluating in-sole design with joint motion, plantar
pressure and rating of perceived exertion measures. Work: A Journal of Prevention,
Assessment and Rehabilitation. 2012; 41: 1114.
[15] Gu YD, Sun D, Li JS, Graham MR and Ren XS. Plantar Pressure variation during
jogging with different heel height. Applied Bionics and Biomechanics. 2014 (In press).
[16] Haw JT, Lee JH, Lee EJ, Lim CH and Kim WB. Comparison of plantar pressure
between flat and normal feet when crossing an obstacle at different heights. Journal of Back
and Musculoskeletal Rehabilitation. Dec 2013; 1-5.
[17] Yamakawa T, Taniguchi K, Asari K, Kobashi S and Hata Y. Biometric Personal
Identification Based on Gait Pattern using Both Feet Pressure Change. Proceeding of 2010
World Automation Congress (WAC), Kobe, Japan, 19–23 September. 2010: 1–6.
[18] Yong F, Yunjian G and Quanjun S. A Human Identification Method Based on Dynamic
Plantar Pressure Distribution. Proceeding of 2011 IEEE International Conference on
Information and Automation (ICIA), Shenzhen, China, 6–8 June. 2011: 329–332.
[19] MacWilliams BA and Armstrong PF. Clinical Applications of Plantar Pressure
Measurement in Pediatric Orthopedics. Proceeding of Pediatric Gait, 2000. A New
Millennium in Clinical Care and Motion Analysis Technology, Chicago, IL, USA, 22 July.
2000: 143–150.
18
[20] Novel Quality in Measurement. Available online: http://www.novel.de/ (accessed on
November 15, 2013).
[21] Shu L, Hua T, Wang Y, Li Q, Feng D and Tao, X. (2009). Inshoe plantar pressure
measurement and analysis system based on fabric pressure sensing array. IEEE Transaction
on Information Technology and Biomedical. 2000; 14: 767.
[22] Bamberg S, Benbasat AY, Scarborough DM, Krebs DE and Paradiso JA. Gait analysis
using a shoe-integrated wireless sensor system. IEEE Transactions on Information
Technology in Biomedicine. 2008; 2: 413.
[23] Razak AHA, Zayegh A, Begg RK and Wahab Y. Foot Plantar Pressure Measurement
System: A Review. Sensors. 2012; 12: 9884.
[24] Gefen A. Pressure-sensing devices for assessment of soft tissue loading under bony
prominences: Technological concepts and clinical utilization. Wounds. 2007; 19: 350.
[25] Urry S. Plantar pressure-measurement sensors. Measurement Science and Technology.
1999; 10: R16.
[26] De Rossi SMM, Lenzi T, Vitiello N, Donati M, Persichetti A, Giovacchini F, Vecchi F
and Carrozza MC. Development of an In-Shoe Pressure-Sensitive Device for Gait Analysis.
Proceeding of 2011 Annual International Conference of the IEEE Engineering in Medicine
and Biology Society, Boston, MA, USA, 30 August–3 September. 2011: 5637.
[27] Yang SM, Han JH, Lee I, Characteristics of smart composite wing with SMA actuators
and optical fiber sensors, International Journal of Applied Electromagnetics and Mechanics,
2006; 23 (3-4): 177.
[28] El-Sherif MA. Smart structures and intelligent systems for health monitoring and
diagnostics. Applied Bionics and Biomechanics. 2005; 3-4: 161.
[29] Tjin SC, Hao J, Lam Y Z, Ho YC and Ng BK. A pressure sensor using fiber Bragg
grating. Fiber and Integrated Optics. 2001; 20 (1): 59.
19
[30] Tjin, SC, Suresh R and Ngo NQ. Fiber Bragg Grating Based Shear-Force Sensor:
Modeling and Testing. Journal of Lightwave Technology. 2004; 22 (7): 1728.
[31] Bhalla S, Soh CK. High Frequency Piezoelectric Signatures for Diagnosis of Seismic/
Blast Induced Structural Damages. Nondestructive Testing and Evaluation International.
2004; 37(1): 23.
[32] Shanker R, Bhalla S and Gupta A. Dual Use of PZT Patches as Sensors in Global
Dynamic and Local EMI Techniques for Structural Health Monitoring. Journal of Intelligent
Material Systems and Structures. 2011; 22 (16): 1841.
[33] Suresh R, Tjin SC and Ngo NQ. Application of a new fiber Bragg grating based shear
force sensor for monitoring civil structural components. Smart Materials and Structures.
2005; 14(6): 982.
[34] Suresh R, Singh C, Hao J Anand S. Fibre Bragg grating sensors for plantar pressure
monitoring at different walking speed. International Journal of Biomedical Engineering and
Technology. 2014; 14 (1): 34.
[35] Moyo P, Brownjohn JMW, Suresh R and Tjin SC. Development of fiber Bragg grating
sensors for monitoring civil infrastructure. Engineering Structures. 2005; 27: 1828.
[36] Bhalla S and Soh CK. Structural impedance based damage diagnosis by piezotransducers. Eartquake Engineering Structural Dynamics. 2003; 32: 1897.
[37] Bhalla S, Soh CK and Liu Z. Wave Propagation Approach for NDE Using Surface
Bonded Piezoceramics. Nondestructive Testing and Evaluation International. 2005; 38: 143.
[38] Kersey AD, David MA, Patrick HJ, LeBlanc M, Koo KP, Askins CG, Putnam MAJ and
Friebele E. Fiber grating sensors. Journal of Lightwave Technology. 1997; 15: 1442.
[39] Hao J. Design and Fabrication of a Fiber Optic Pressure Sensor. Ph.D. Thesis, Nanyang
Technological University, Singapore. 2000.
20
[40] Rao YJ. In-fibre Bragg grating sensors. Measurement Science and Technology. 1997; 8:
355.
[41] Hao J, Jayachandran M, Kng PL, Foo SF and Aung PWA. A Smart Multifunctional
Fiber Bragg Grating Sensor System for Healthcare Applications. International Conference on
Material for Advanced technologies, 28 June – 3 July. 2009.
[42] Hao JZ, Tan KM, Tjin SC, Liaw CY, Chaudhuri PR, Guo X and Lu C. Design of a footpressure monitoring transducer for diabetic patients based on FBG sensors. Lasers and
Electro-Optics Society, LEOS 2003, The 16th Annual Meeting of the IEEE, 27-28 October.
2003: 23.
[43] Agilent Technologies (2013) http://www.agilent.com
[44] PI Ceramic (2012) http://www.piceramic.de
[45] Nevill AJ, Peppa MG and Whiting M. In-sole foot pressure measurement system
utilizing piezoelectric film transducer. Medical and Biological Engineering and Computing.
1995; 33: 76.
[46] Micron Optics Inc. Available at http://www.micronoptics.com (accessed on 15
November 2013)
[47] Tecktronics Inc, 2013. http://www.tektronixindia.com/oscilloscope
[48]
Dananberg
H,
DPM
and
Williams
B,
Tekscan
www.breakthroughpodiatry.com/pdf/forms/Intro-Force-vs-Time-Curve.pdf)
21
Inc.,
(http://
LIST OF TABLE
Table 1 Pressure values at left foot sole as measured by PZT/ FBG sensors on left foot
LIST OF FIGURES
Fig. 1
Stacking sequence of CCM lamina and FBG attachment [45]
Fig. 2
Electro-mechanical (voltage-pressure) interaction in PZT patch Fig. 3 Fabricated shoes
samples. FBG sensors/ PZT patches are attached at forefoot and heel
Fig. 3(a) Experimental shoes withFBG sensors/ PZT patches are attached at forefoot and
heel.
Fig. 3(b) Details of attachment of PZT sensors
Fig. 4
Calibration of FBG sensors embedded at (a) left forefoot (b) left heel (c) right heel
Fig. 5
Experimental set up at the treadmill
Fig. 6
Wavelength shift of FBG sensor embedded at forefoot at the speeds 1kmph and
3kmph
Fig. 7
Typical oscilloscope screenshot at 3kmph
Fig. 8
Variation of pressure below (a) left foot and (b) right foot captured by PZT sensor at
walking speed of 5 kmph
Fig. 9
Variation of pressure below (a) left foot and (b) right foot captured by PZT sensor at
walking speed of 7 kmph
Fig. 10
Variation of pressure below the left foot under varying speed for PZT sensor and FBG at
(a) forefoot and (b) heel
22
Table 1 Pressure values at left foot sole as measured by PZT/ FBG sensors on left foot
Pressure values measured by FBG
Walking speed sensors (kPa)
(Kmph)
Forefoot
Heel
Standing on
796.49
524.81
both feet
Standing on
1518.8
1158.1
one foot (left)
1
1117.7
669.05
2
1061.6
623.86
3
964.24
582.28
5
863.48
539.67
7
740.57
500.52
23
Pressure values measured by PZT
sensors (kPa)
Forefoot
Heel
-
-
-
-
169.92
396.48
424.80
453.12
509.76
56.64
249.22
339.84
543.74
996.86
Arc
Height
Effective length
Fig. 1 Stacking sequence of CCM lamina and FBG attachment [45]
24
E3
3
2
T3
1
PZT patch
T3
Fig. 2 Electro-mechanical (voltage-pressure) interaction in PZT patch
25
FBG Sensor at
forefoot
PZT patches
PZT patches
FBG sensor
(medial heel)
(a)
Shoe
Inner adhesive layer
Outer
adhesive layer
PZT patch
(b)
Fig. 3 (a) Experimental shoes withFBG sensors/ PZT patches are attached at forefoot and heel.
(b) Details of attachment of PZT sensors
26
Wavelength Shift (pm)
500
450
400
350
300
250
200
150
100
50
0
0
100
200
300
400
Pressure (kPa)
(a)
Wavelength Shift (pm)
600
(a)
500
400
300
200
100
0
0
100
200
300
400
Pressure (kPa)
(b)
Fig. 4 Calibration of FBG sensors embedded at (a) left forefoot (b) left heel (c) right heel
27
Instrumented
Shoe Pair
Tread Mill
(to maintain speed)
Fig. 5 Experimental set up at the treadmill
28
3 kmph
Wavelength shift (pm)
1 kmph
Time
Fig. 6 Wavelength shift of FBG sensor embedded at forefoot at the speeds
1kmph and 3kmph
29
Forefeet (Left)
Heel (Left)
Forefeet
(Right)
Heel (Right)
Fig. 7 Typical oscilloscope screenshot at 3kmph
30
600
Left Forefoot
450
Pressure (kPa)
300
150
0
-150
-300
Left Heel
-450
-600
15
16
17
18
19
20
Time (s)
(a)
1200
30
800
20
400
10
0
0
-400
-10
Right Heel
-800
Heel Pressure (kPa)
Forefoot Pressure (kPa)
Right Forefoot
-20
-1200
-30
15
16
17
18
19
20
Time (s)
(b)
Fig. 8 Variation of pressure captured by PZT sensor at walking speed of 5 kmph below
(a) left foot and (b) right foot
31
450
Pressure (kPa)
200
-50
-300
Left Forefoot
-550
-800
Left Heel
-1050
10
11
12
13
14
15
Time (s)
(a)
1600
50
Right Forefoot
30
800
400
10
0
-10
-400
Heel Pressure (kPa)
Forefoot Pressure (kPa)
1200
-800
-30
Right Heel
-1200
-1600
-50
10
11
12
13
14
15
Time (s)
(b)
Fig. 9 Variation of pressure captured by PZT sensor at walking speed of 7 kmph below
(a) left foot and (b) right foot
32
1200
FBG
1000
Pressure (kPa)
800
600
400
PZT
200
0
0
2
4
Speed (kmph)
6
8
6
8
(a)
1200
Pressure (kPa)
1000
800
FBG
600
400
PZT
200
0
0
2
4
Speed (kmph)
(b)
Fig. 10 Variation of pressure below the left foot under varying speed for PZT sensor and FBG at
(a) forefoot and (b) heel
33