jaja

1. Energy Harvesting from Passive
Human Power
Faruk Yildiz
Sam Houston State University
E-mail : fxy001@shsu.edu
ABSTRACT dios, became available in the market. The commer-
Sustaining the power resource for autonomous wire- cially available Freeplay’s [3] wind-up radios require
less and portable electronic devices is an important 60 turns in one minute of cranking, which allows for
issue. Ambient power sources, such as a replace- the storage of 500 Joules of energy in a spring. The
ment for batteries, can minimize the maintenance spring system drives a magnetic generator and ef-
and the cost of operation by harvesting different ficiently produces enough power for about an hour
forms of energy from the potential energy sources. of radio play.
Researchers continue to build high-energy-density Recently researchers have performed several
batteries, but the amount of energy available in studies in alternative energy sources that could pro-
the batteries is not only finite but also low, limit- vide small amounts of electricity to low-power elec-
ing the lifetime of the system. Extended lifetime of tronic devices. These studies focused on investing
electronic devices is very important and also has and obtaining power from different energy sources
more advantages in systems with limited accessibil- such as vibration, light, sound, airflow, heat, waste
ity. This research studies one form of ambient en- mechanical energy, and temperature variations.
ergy sources: passive human power generated from The piezoelectric energy-harvesting method
a shoe/sneaker insole when a person is walking or converts mechanical energy into electrical energy by
running and its conversion and storage into usable straining a piezoelectric material [4] which causes
electrical energy. Based on source characteristics, charge separation across the device, producing an
electrical-energy-harvesting, conversion, and stor- electric field and consequent voltage drop propor-
age circuits were designed, built, and tested for low- tional to the stress applied. The oscillating system
power electronic applications. is typically a cantilever beam structure with a mass
at the unattached end of the lever, since it provides
INDEX TERMS higher strain for a given input force [5]. The volt-
Circuit Analysis, Circuit Design, Energy Conserva- age produced varies with time and strain, effectively
tion, Energy Conversion, Energy Storage, Piezoelec- producing an irregular AC signal. The piezoelectric
tric Devices, Piezoelectric Materials energy conversion produces relatively higher volt-
age and power-density levels than an electromag-
I. INTRODUCTION netic system. Moreover, piezoelectricity has the
Energy harvesting is the conversion of ambient en- ability of elements, such as crystals, and some types
ergy into usable electrical energy. When compared of ceramics to generate an electric potential from a
to energy stored in common storage elements, such mechanical stress [6]. If the piezoelectric material
as batteries and capacitors, the environment repre- is not short circuited, the applied mechanical stress
sents a relatively infinite source of available energy. induces a voltage across the material.
Researchers have been working on many projects The problem of how to get energy from a
to generate electricity from human power, such as person’s foot to other places on the body has not
exploiting, cranking, shaking, squeezing, spinning, been suitably solved. For a radio frequency identi-
pushing, pumping, and pulling [1]. Several types of fication (RFID) tag or other wireless device worn on
flashlights were powered with wind-up generators the shoe, the piezoelectric shoe insert offers a good
in the early 20th century [2]. Later versions of these solution. However, the application space for such
devices, such as wind-up cell phone chargers and ra- devices is extremely limited, and, as mentioned ear-
5

2. lier, not very applicable to some of the low-powered ment, PFCB, was available to test with a small con-
devices such as wireless sensor networks. Active stant shaker. The shaker functioned as an ambient
human power, which requires the user to perform a vibration source (passive human power) and was
specific power- generating motion, is common and used to vibrate the PFCB to produce electricity for
may be referred to separately as active human-pow- the energy-harvesting circuit. A photograph of the
ered systems [7]. PFCB with the inter-digitized electrodes to align the
An example of energy harvesting using uni- field (energy-harvesting circuit) with fibers is shown
morph piezoelectric structures was conducted by in Figure 1.
Thomas, Clark, and Clark [8]. This research was fo-
cused on a unimorph-piezoelectricity circular plate, Device size (mm): 130 x 10 x 1
which is a piezoelectric (PZT) layer assembled on Active elements: AFCB
an aluminum substrate. The vibrations were driven Mode: d33
from a variable-ambient- pressure source, such as a Full scale voltage range (V): ±400
scuba tank or blood pressure monitor. The research- Full scale power range (mW): ±120
ers showed that by creating the proper electrode pat-
Figure 1. Basic specifications of the PFCB
tern on the piezoelectric element, (thermal regroup-
ing), the electrode was able to produce an increase in
The PFCB energy source is modeled as a
available electrical energy. In the system, as the can-
steady AC power source for the circuit components of
tilever beam vibrates, it experiences variable stresses
the energy-harvesting circuit. Since all the electronic
along its length. Regrouping the electrodes target-
components for the energy-harvesting and battery-
ing specific vibration modes resulted in maximum
charging circuit require DC voltage to operate, this AC
charge collection. This type of design may offer the
source is then converted to a DC voltage source.
possibility for miniaturization and practicality to
piezoelectric energy-harvesting technology.
II. SYSTEM DESIGN
The piezoelectric active-fiber composites
For the purpose of energy harvesting and storage,
(AFCs) are made by Advanced Cerametrics Incor-
shoe or sneaker insoles are good sources of me-
porated (ACI) [9] from a uniquely-flexible ceramic
chanical stress, deformation, and vibration when a
fiber that is able to capture wasted ambient energy
person is walking or moving his/her feet. With this
from mechanical vibration sources and convert it
method, waste-ambient mechanical energy was con-
into electric energy. The piezoelectric composites’
verted to electrical voltage through a unique ener-
fiber-spinning lines are capable of generating elec-
gy-harvesting circuit. An overall energy-harvesting
tricity when exposed to an electric field. In piezo-
model is shown in Figure 2 to explain implemen-
electric fiber composite bimorph (PFCB) architec-
tation steps and potential applications. In order to
ture, the fibers are suspended in an epoxy matrix
have the best efficiency and output power, the cir-
and connected using inter-digitized electrodes to
cuit was designed and developed according to the
create an AFC. Advance Cerametrics Incorporated
ambient-source, characteristics, PZT ceramic- fiber
has demonstrated through testing that thin fibers
composite and load constraints. The energy- har-
with a dominant dimension, a length, and a very
vesting system is capable of capturing even minute
small cross-sectional area are capable of optimizing
amounts of stress and vibrations, then converting
both the piezo and the reverse-piezo effects.
them to electric power sufficient to run low-power
An investigation into the improvement of
electronic systems.
an energy-harvesting-system performance and ef-
ficiency using a PFCB is considered in this research.
The PFCB characteristics and properties were inten-
sively studied in order to build an efficient energy
harvesting circuit for further study. The efficiency
of the PFCB was measured by building an operation-
al-difference, amplifier instrumentation test circuit
and following an energy-harvesting and battery-
charging circuit. Only one type of piezoelectric ele-
6 j o u r n a l o f a p p l i e d s c i e n c e & e n g i n e e r i n g t e c h n o l o g y 2011

3. age level, frequency, and corresponding time. AC
signals and voltage outputs were similar to each
other; however, the time required for the vibration
to decay was observed to be longer when more mass
was attached on the PFCB. All signal outputs, that
occurred until the vibration decayed completely,
could not be shown on the oscilloscope screen.
However, it was observed that the time until vibra-
tion stops is longer with more mass attached on the
PFCB. The resonant frequency does not depend upon
the pencil flicks. It depends upon the mass (including
distributed mass and lumped-tip mass) and distrib-
uted spring constant of the cantilevered beam. For a
sinusoidal excitation, the most energy is transferred
when excited at the resonant frequency. The decay
depends upon the input resistance of the measur-
ing device (electrical damping) and the mechanical
damping from the material and from the air.
Figure 2. Overall energy-harvesting model
Power Characteristics
A piezoelectric energy source is most often mod-
eled as an AC voltage source because of its AC pow- Figure 3.The voltage decay of the open-circuit PFCB
er characteristics and features. Piezoelectric fiber
composite can be connected in series with the ca- The PFCB was carefully clamped on the
pacitors and resistors to reduce or smooth a high- shaker with plastic bumpers to avoid damaging
voltage input produced by PFCB. For test purposes, the part during its vibrations. The wiring between
the tip of a mechanical pencil was used to flick the all modules was done carefully to allow for read-
tip of the PFCB product in order to provide the ini- ing the voltage outputs from the oscilloscope and
tial disturbance. The test equipment used included multi-meter displays. The PFCB layer and material
a multi-meter, shaker, and oscilloscope, connected properties were not known to predict the frequency
to each other properly to obtain voltage readings rate accurately, so the value had to be determined
from the PFCB. The first test for voltage output de- experimentally on the test fixture.
pended on time variation and was conducted with- In order to allow the calculation of the cur-
out any mass placed on the tip of the PFCB. This rent output, wires from the PFCB electrodes were
was followed by a second test with variable masses connected to the oscilloscope probes through a 1
that were placed on the tip of the PFCB to observe kΩ resistor. The current outputs could not be mea-
the output voltage levels. As the more mass that is sured by a multi-meter and were observed from the
added on the tip of the PFCB, the more time passes oscilloscope screen when the PFCB was vibrated by
until vibration of the PFCB stops. The voltage from the shaker at 60Hz.
the PFCB increases depending on the mass and force It was not possible to plot the current and
applied to the tip of the PFCB. power outputs due to a lack of the proper data-
The plot in Figure 3 is the peak-to-peak volt- acquisition system during the research. However,
Energy Harvesting From Passive Human Power 7

4. very-low-current outputs were still produced that Technology based DC-DC buck-boost converter and
might be harvested with a proper energy harvesting battery-charging circuit components [15].
circuit. From the vibration test results, it was de- Initially, a full-wave bridge rectifier was
termined that at the variable frequency, the power placed between the PFCB and the operational- am-
generated from the PFCB is sufficient for low-pow- plifier-instrumentation circuit that converts AC
er electronic devices. The obtained values would signals to DC signals. A full-wave bridge rectifier is
be enough to build an energy harvesting circuit to very efficient, converting positive and negative cy-
charge a small-scale storage device such as a bat- cles from the PFCB and supplying DC voltage to the
tery, capacitor, or super capacitor, albeit slowly. battery through the battery- charging part of the en-
ergy-harvesting circuit. Since the current produced
III. ENERGY HARVESTING CIRCUIT DESIGN from PFCB was low, an intermediate-operational-
After testing the power output and the working amplifier (op-amp) circuit was used to test various
characteristics of the PFCB at different vibration current levels that was generated by PFCB in differ-
levels and attached masses, the researcher built the ent decaying times. [16]. This instrumentation cir-
energy-harvesting circuit used to charge the bat- cuit consisted of operational amplifiers, resistors,
teries under low-current levels. The mechanical-to- and intermediate/storage capacitors to implement
electrical energy conversion is usually managed by the circuit at ±15V.
the energy-harvesting circuits including convention- A buck-boost converter and battery- charg-
al buck-boost converters, bridge rectifiers, and bat- ing circuit is the last part of the simulation inter-
tery-charging circuits [10,11,12,13]. The energy har- face before the storage unit. The op-amp part of the
vesting circuit was designed, developed, and built energy-harvesting circuit consisted of three single
according to the ambient source and piezoelectric operational amplifiers that are configured as differ-
fiber composites’ low-current constraints in order ence amplifiers.
to produce efficient power output. This implies that the voltage differential be-
The energy-harvesting and battery charg- tween the two branches is the output of the circuit.
ing circuit design was built with typical compo- This op-amp-instrumentation-circuit design (Figure
nents that could decrease high-input voltages and 5) helped to observe voltage outputs from the PFCB
increase low-input currents from the PFCB to pro- and the capacity changes of the capacitors when the
vide sufficient charge currents to the batteries. The PFCB was being vibrated.
circuit was designed to start charging when the bat- Capacitors C1 and C11 were charged de-
tery voltage drops beyond a nominal value, and it pending on the voltage generated by the PFCB,
stops charging when voltage is reached at the bat- which was observed by the oscilloscope through the
tery nominal voltage. operational-amplifier-instrumentation circuit.
The Linear Technology SPICE simulator
(LTSPICE) simulation interface that shows the over-
all circuit is depicted in Figure 4; it represents the
system circuit modules that are simulated together
to test the output power level of the circuit [14].
All the necessary simulations were conducted us-
ing SwitcherCADTM Spice III, because of the Linear
Figure 4. Energy-harvesting circuit simulation interface
8 j o u r n a l o f a p p l i e d s c i e n c e & e n g i n e e r i n g t e c h n o l o g y 2011

5. voltage drops across the capacitor when measuring
the voltage with a digital multi-meter. However, the
capacity readings across the capacitor C11 would be
accurate, since the operational amplifier keeps the
initial voltage-level constant.
DC buck-boost converter and battery charging
circuit
DC-DC converters efficiently step-up (boost), step-
down (buck), or invert DC voltages without the neces-
sity of transformers. In these structures, switching
capacitors are usually utilized to reduce or increase
physical size requirements. DC-DC converters assist
Figure 5. Operational amplifier instrumentation in product-size reduction for portable electronic
circuit devices where increased efficiency and regulation
of input power are necessary for optional require-
The general operational amplifier in Figure ments. Taking the above features of the buck-boost
5 was used to observe the charging phase of the converters into consideration, a linear-technology
C11 intermediate-storage capacitor. The 5V IC (ini- based LT1512 integrated circuit (DC-DC buck-boost
tial voltage) was supplied across capacitors in both SEPIC constant current/voltage battery charger) was
circuits. In circuit A, the voltage across the capacitor used to regulate the high-output voltage that was
with the 10MEG impedance, representing a flux digi- produced from the PFCB to charge small-scale bat-
tal multi-meter, was measured. The voltage across teries for test purposes [17].
capacitor C1 dropped almost 2V when the circuit A LT1512 battery-charging circuit was
was simulated at the same input voltage. added to the energy-harvesting circuit, considering
However, the voltage across capacitor C11 in the its characteristics, based upon an application data
operational-amplifier circuit stayed at constant volt- sheet. Since buck-boost converters are very sensitive,
age other than negligible voltage drops. The voltage proper design, in conjunction with supporting com-
level across both capacitors, C1 and C11, was simu- ponents and physical layout, is necessary to avoid
lated and is plotted in Figure 6 in order to compare electrical noise generation and instability. The con-
voltage drops across the capacitors. siderations, including LTSPICE modeling, converter
When the PFCB was placed on the constant selection, circuitry building, debugging, and power-
shaker, it started generating voltages by charging output improvements, were followed step-by-step to
the capacitors. The operational- amplifier circuit obtain adequate energy-harvesting circuitry.
kept the initial voltage level constant to allow for This circuit would maximize the power flow
an accurate reading of the voltage levels of the in- from the piezoelectric device and was implemented
termediate-storage capacitor. In circuit A, the volt- in coordination with a full-wave bridge rectifier, in-
age readings would not be accurate because of the termediate-storage capacitor, and voltage-sensitive
switching circuit. It was observed that, when using
the energy-harvesting circuit, over twice the amount
of energy was transferred to the battery compared
to direct charging alone. However, if the power-
harvesting medium produced less than 2.7V, power
flow into the battery was reduced due to losses in
the additional circuit components and the threshold
characteristics of the LT1512.
For the purpose of storing energy in the
intermediate storage unit, a capacitor was placed
before the voltage-sensitive circuit and buck-boost
Figure 6. Voltage levels across intermediate storage converter. The voltage-sensitive circuit consists of di-
capacitors
Energy Harvesting From Passive Human Power 9

6. Figure 7. Intermediate-voltage-sensitive switch with hysteresis
odes, MOSFET switches, and resistors to transfer the intermediate capacitor and transfers it to the battery
energy from the intermediate capacitor to the battery through a DC-DC buck-boost converter and battery-
through the DC-DC buck-boost converter [18]. charging circuit. The fourth module is the model of
The MOSFET switches and zener diodes on a buck-boost converter and battery-charging circuit
the voltage-sensitive circuit sense the voltage in the representing the exact characteristics of the LT1512
intermediate capacitor and transfer the energy when SEPIC-constant-current/voltage integrated circuit.
the capacitor reaches specific voltage levels. The volt- The voltage level after the intermediate ca-
age level in the intermediate capacitor is controlled pacitor and voltage-sensitive switch simulation is
by the zener diodes until the capacitor is discharged plotted in Figure 8. VIN (the capacitor voltage level)
by transferring its energy to the battery. Depending reaches only 15V, and starts discharging by trans-
on the zener diode values, the stored energy in the ferring voltage to the battery. When (VIN) starts
capacitor is transferred to the storage unit through decreasing, (VOUT) increases until 15V with the ca-
DC-DC buck-boost converter and battery-charging pacitor voltage at 15V. Both (VIN) and (VOUT) start de-
circuit (The switch should be between 5V-15V). creasing by transferring energy to the battery. (VOUT)
Due to known high-discharge rates of the ca- reaches zero voltage while transferring its energy
pacitors, the zener-diode-voltage values of 12V and to the battery; simultaneously, the (VIN) value de-
6.2V were chosen (which are small values for the pur- creases in order to reach 15V again. The charge and
pose of energy harvesting from PFCB) to avoid loos- discharge steps are repeated while PFCB produces
ing stored energy in the intermediate capacitor. One electricity from vibrations.
of the biggest benefits of the intermediate capacitor
and voltage-sensitive switching circuit is to increase
the amount of transferred energy from the PFCB. Cir-
cuit loss is reduced throughout the energy-harvesting
circuit caused by the electronic components.
The circuit shown in Figure 7 was designed
has four phases to represent the overall energy- har-
vesting circuit modules. Figure 8. Voltage input and output simulation of
The first module is a mechanical-to- elec- voltage-sensitive switch
trical energy conversion module and functions the
same as PFCB producing AC power. The second The DC-DC converter and battery-charging
module has rectification (the conversion of AC volt- circuit design, a part of the energy harvesting circuit
age to DC voltage) and an energy-storage unit (inter- simulation interface (Figure 9) is employed to han-
mediate capacitor). The third module is a voltage- dle the decrease or increase of voltage levels and
switching circuit that senses the voltage level of the adjust it according to the battery specifications. The
10 j o u r n a l o f a p p l i e d s c i e n c e & e n g i n e e r i n g t e c h n o l o g y 2011

7. voltage output of the circuit can be easily modified battery at the nominal voltage level, which is 3.6V at
by using different resistance values if a different 60mAh for the test battery.
battery is integrated to the system. The battery-charging current (IROUT) and
battery-charging voltage (VOUT) simulation plots are
depicted to indicate battery-charging values. Both
voltage and current levels were supplied at a steady
state for proper battery charging, IROUT =5mA,
which is a standard charging current for the bat-
tery, and VOUT =3.6V nominal charging voltage.
The charging current that was generated by the
PFCB was less than 1mA but was increased to 5mA
by the intermediate capacitors. The intermediate
capacitors were charged to the minimum charging
threshold of the battery and then released to the
battery terminals by discharging themselves, allow-
Figure 9. Energy-harvesting and charging circuit ing the capacitors to accept charge voltages from
the PFCB again. However, the current level could not
Circuit Simulation increase to charge the battery at the quick-charge
The simulation of the important circuit components phase because of the low current produced by the
through the LT1512-SEPIC-battery- charging circuit PFCB. The specific voltage and current levels that
(including input voltage, battery- charging voltage, are specified in the simulation plot can charge at
and current) are depicted in Figure 10. All three im- 3.6V at 60mAh for a fully discharged battery in ap-
portant aforementioned parameters of the energy- proximately 27 hours with constant vibrations.
harvesting and battery- charging circuit were sim-
ulated together to examine the consistency of the Building the Circuit
voltage/current levels on the circuit-design-simula- Considering the variable output voltages from the
tion interface. PFCB, an energy-harvesting circuit was designed to
charge small-scale NICD/NIMH batteries at the con-
stant-charging phase. The printed circuit board for
the energy-harvesting circuit was designed and built
as small as possible to fit even the smallest places
for power generation, including the battery soldered
Figure 10. Battery-charging values simulation on the circuit. However, for test purposes, the in-
strumentation circuit on the prototyping board and
The input voltage (VIN) simulation plot was the energy-harvesting circuit were placed near the
generated by the vibrations through the PFCB while measuring equipment with the PFCB assembled to
being shaken. This voltage level was measured after
rectification of the AC voltage signal that came from
the PFCB unit as a DC voltage and served as the in-
put for the buck-boost converter and the voltage-
regulator circuit. Since the maximum input voltage
of an LT1512- integrated circuit is 30VMAX, a zener
diode was placed between VIN and the ground of
the LT1512 in order to avoid damaging the internal
chip components of the LT1512. The input voltage
(VIN) and regulated voltage (VOUT) are compared in
order to check the input and output voltage differ-
ences after regulation. The input voltage levels that
were larger or less than (3.6V) were regulated by the
Figure 11. Energy-harvesting, conversion, and charging
LT1512 buck-boost converter in order to charge the
circuit
Energy Harvesting From Passive Human Power 11

8. allow for the reading of output values on the oscillo- (as reported by tests). The graph in Figure 12 com-
scope display. The energy-harvesting circuit, which pares voltage and current levels and average power
is soldered on the printed circuit board, is shown in output in 13 seconds.
Figure 11. A protective box should be designed and When a resistive load is relatively large, the
built to protect the circuit components and the bat- power output from the PFCB does not produce sig-
tery from bending and experiencing deformations nificantly more power. The results of using a larg-
from the vibration sources. er capacitor to smooth the voltage output suggest
that the size of the smoothing capacitor affects the
Storage Unit Tests amount of power that can be delivered to a resis-
One problem often encountered when using pow- tive load (battery). This result is attributed to the
er-harvesting systems is that the power produced non-ideal behavior of the capacitor, which leads to
by the piezoelectric material is often not sufficient internal losses. Following construction of the ener-
to power most electronics. Therefore, methods are gy-harvesting circuit, NICD- and NIMH-type batter-
needed to accumulate energy in an intermediate ies were charged to determine the battery charging
storage device so that it may be used as a power time that could be effectively observed for each
source. A capacitor is typically used to accumulate battery, with constant frequency. After testing the
the energy. However, capacitors have characteristics voltage levels of the PFCB using the capacitors, the
that are not ideal for many practical applications PFCB was then tested with the batteries to observe
such as limited capacity and high leakage rates. For battery-charging efficiency. For this purpose, a per-
the purpose of intermediate storage units, typical manent magnet shaker was used to induce vibra-
capacitors were used in the energy-harvesting cir- tions; three rechargeable batteries were used for
cuit without causing any critical issues. A group of the experiment (Fig. 13). A PFCB consisting of two
capacitors were connected in parallel with the re- active-fiber composites (bimorph) was clamped to
sistors in order to smooth the delivered voltage, a thin piece of metal on the constant shaker for the
making the output voltage easily read by the multi- energy-harvesting experiment. The constant vibra-
meter. According to the approximate displacement tions from the shaker were applied to the PFCB at
and frequency levels, stored energy in the capaci- 60Hz. The voltage measurements from the batteries
tors was calculated. were taken every hour, and it appeared that the in-
The current levels for battery-charging pur- crease was minor. The reason for the slow charging
poses were calculated according to the input voltage was the very low current produced from the PFCB
levels. If the input voltage is increased, the output and the losses across the energy-harvesting and bat-
current would automatically increase by decreas- tery-charging circuits. Because of the low-charging
ing the battery-charging time. All calculations were current, the test battery was not able to be charged
done according to the energy stored in 13 seconds at the specified standard charging time. However,
this charging experiment was conducted only with
a single PFCB, which is not recommended for charg-
ing batteries. In some applications, more than three
PFCBs are connected in parallel to increase the cur-
rent levels and efficiency of the energy harvesting
system. The number of PFCBs would be increased
to charge the batteries at a specified time frame to
avoid voltage drops across the batteries while pow-
ering the electronic application. The batteries used
in the experiment are listed in Table I with the basic
specifications that are needed as charging param-
eters. In order to charge batteries, the PFCB inter-
digitized electrodes were connected to the battery
terminals through the energy harvesting circuit.
This experimental test showed that batter-
ies were being charged with constant current/volt-
Figure 12. Energy stored in 13 seconds
12 j o u r n a l o f a p p l i e d s c i e n c e & e n g i n e e r i n g t e c h n o l o g y 2011

9. Quick
Charge Charge
Nom. Ampacity Charge
Type time
V mAh I Tim I T
h
mA h mA h
NIMH 1.2 80 8 15 47
NICD 3.6 60 6 14 20 7 36
NIMH 3.6 60 6 14 20 7 41
Table 1. Rechargeable Battery Specifications
age in longer time frames than the specific time son starts walking. Several pennies were attached
frame in the battery datasheets. The last column in to the tip of the PFCB to increase the vibration time
Table I shows the time required to charge the batter- while the PFCB was being shaken by the foot-steps.
ies with one PFCB. If more PFCBs are used for the en- The power output of the PFCB was directly propor-
ergy harvesting, charging time would be decreased tional to the force and mass applied on the PFCB
considerably. to induce vibrations. However, the output power of
the energy-harvesting circuit was a constant voltage
IV. SNEAKER SOLE EXPERIMENT and current to avoid damaging the battery (because
A sneaker-insole was considered as a possible am- of the unregulated voltages produced by the PFCB).
bient energy source to generate electricity through Therefore, the amount of energy generated by hu-
a PFCB that could be used to charge low-scale re- man power through the PFCB was determined by
chargeable batteries. The batteries are expected to human power that was available to shake the piezo-
power low-power electronic applications such as a electric material. The stronger the force and mass
radio, MP3 player, mobile phone, and GPS unit. A typ- applied, the more electrical power was generated
ical MP3 player was chosen as an electronic device to from the PFCB while the person walked or ran. As
make the power estimation between generated- and a result of vibrations, the fibers in the frame of the
consumed- power levels. Table 1 demonstrates that shoe sole generated electricity with high potential
the PFCB was able to produce enough voltage with and low current. The electricity was conducted and
low current as an input power to charge a small- scale stored in a battery or capacitor in the circuit placed
rechargeable battery, depending on the time a person into a special compartment of the shoe insole. The
walks or runs. The analytical estimation of the bat- photograph of the redesigned sneaker insole with
tery-charging time and walking-time relationship was the assembled PFCB is shown in Figure 13.
calculated for the fully discharged batteries. Howev-
er, in the case of powering the electronic application,
the batteries were placed in the system fully charged.
It is essential to determine if the gained and stored
power compensates for the consumption of the elec-
tronic device while it is operating. If the produced
power compensates for the daily consumptions and
the leakages of the electronic device, a sneaker insole
as an ambient energy source is a feasible source for
the electronic application.
For experiment purposes, the sneaker sole
was cut carefully to place the PFCB in the correct po-
Figure 13. Redesigned sneaker insole with PFCB
sition for maximum efficiency when a person walks
or runs. The base where the PFCB was placed was su-
The wires coming from the PFCB were ex-
per glued with a piece of thin wood in order to place
tended by wires to the energy-harvesting-circuit in-
the PFCB properly on a smooth surface. Following
put. In order to avoid voltage drops due to long wires,
the proper placement of the PFCB, the insole of the
the PFCB and energy-harvesting circuit were placed
sneaker was again covered with the piece taken from
in close proximity. The efficient point of this experi-
the sneaker sole to avoid any damages when a per-
Energy Harvesting From Passive Human Power 13

10. ment demonstrated that if two sneakers (each with also necessary to determine the power consumption
one PFCB attached) are worn, the output from the vi- of the MP3 player (PMP3) separately per usage, before
brations of both PFCBs never stops, even if a person a calculation of the overall energy consumption. It
walks slowly. Once the first PFCB starts vibrating, it is assumed that a person walks/runs about 1 hour
takes at least one second until the vibrations stop. In in 24 hours. The power characteristics of a small,
this time period, the other PFCB starts shaking when low-power MP3 player were provided by one of the
the person takes another step. In this manner, battery- MP3 manufacturer’s data sheets. The specifications
charging is decreased to almost half of the standard of the MP3 player were used to calculate the power
charging time as specified in the previous sections. consumption for one hour of use (PMP3). According
Since the current level produced through to the specifications, the power usage of the MP3
the energy-harvesting circuit for battery charging player in one hour is calculated as 72mW.
has been calculated, the next comparison estimated
whether the system could produce enough power IMP3 = 60mA;
for a typical MP3 player. The walking time also was
VMP3 = 1.2V; and PMP3 = I * V = 60mA * 1.2V =
calculated to compute how much walking is needed
to compensate the energy consumption of the elec- 72mWh.
tronic application (Figure 14).
There are certain components in the sys-
tem that is always at stand-by to sense the wake-up
signals. These components drain some quiescent
currents from batteries while they are on standby
(including the MP3 player, the energy-harvesting cir-
cuit, and the battery that keeps the device up and
running). The total leakages and quiescent current
for the system components during the playing of
the MP3 player were calculated using Equation 1.
Power consumption of the MP3 player (PMP3) was
found separately and included in overall current
consumption in 24 hours as calculated below.
I1 (LOSS/24HRS) = [(IBATTERY_LEAKAGE) + (IHARVEST_ LEAKAGE) + (IMP3_
LEAKAGE
) + (IMP3)]; so
Figure 14. Block diagram of comparison estimation for
energy harvesting I1 (LOSS/24HRS) = [(270μA)
+ (984μA) + (13mA)] +
[(60mA)] =74mAh.
Overall energy used/produced relationship
estimations were conducted considering (I1), overall
The value for I1 is converted to the power unit in or-
current consumption, and (I2), current gained from the
der to compare the power gain and the power loss.
PFCB in the sneaker insole, through the human power.
Then
I1 (LOSS/24HRS) = [(IBATTERY_LEAKAGE) +
P1 (LOSS/24HRS) = 0.074A * 1.2V = 0.088Wh (88mW).
(IHARVEST_ LEAKAGE) + (IMP3_ LEAKAGE) + (IMP3)] (1)
The total energy drained from the batter-
where I1 (LOSS/24HRS) is the current loss per 24 hours,
ies was estimated at 0.088W for one day. This value
and IMP3 is the current consumption of the MP3.
is not exact energy consumption; it may change
according to how often a person changes the pa-
Equation 1 calculates the overall current
rameters of the MP3 player while using the device.
consumption of the MP3 player including the to-
Since the total energy loss (P1) was estimated, the
tal leakages on stand-by mode, energy- harvesting
next step was to calculate the energy gain from the
circuit, and batteries. According to Equation 1, it is
energy-harvesting system. This equation enables
14 j o u r n a l o f a p p l i e d s c i e n c e & e n g i n e e r i n g t e c h n o l o g y 2011

11. calculation of the total gained current from human
power through the PFCB that is assembled in the
sneaker insole during one day (24 hours), the bat-
tery charging time assumed.
Thus:
I2 (GAIN/24HRS)
= IG * T; (2)
where I2 (GAIN/24HRS) is the total current recovered/
stored per 24 hours, IG is the current gathered while
a person walks for one hour, and T is the time per-
son walked/run (in hour).
For this application, the gained current from
the PFCB (I2) should be more than or equal to the Figure15. Energy gain/loss depends on the time a
overall current loss (I1) in 24 hours (I1≤I2). Otherwise, person walks/runs
the MP3 player would be operating inconsistently
due to insufficient current. The total energy gained Figure 15 shows that the gained power is
from the system would depend on the time a person too low to play an MP3 player for one hour with one
walks/runs during a 24-hour period: hour of walking. To improve the energy gain, more
than one PFCB should be placed into the sneaker in-
I2 = (0.005A * 2hrs) = 0.01A. sole to increase the generated power to balance the
(GAIN/24HRS)
power consumption of the MP3 player. Another so-
The value for I2 is converted to the power unit in lution to make the harvesting circuit efficient would
order to make comparison between the power gain be improving or redesigning the circuit to increase
and the power loss. Then the current flow to the battery to decrease the bat-
tery charging time. Also, if the time a person walks
P2 (GAIN/24HRS) = 0.01A * 1.2V = 0.012Wh (12mW). increases (more than one hour), the battery charging
time would be decreased depending on how long a
P1 and P2 were calculated and converted to the en- person walks in a day.
ergy value in order to make a comparison ratio if
energy gain is greater than energy loss in order to V. CONCLUSION
balance the system power. The advances made from this research build the
framework for further experimentation with the
(3) tools necessary to use the PFCB effectively in nu-
merous applications. The sensing capabilities of
where EG is the overall energy gain, EOUTPUT is energy the PFCB were investigated, and shown in an ener-
loss through the MP3 player, and EINPUT is the energy gy-harvesting system through an experiment and
gain from PFCB; battery-charging circuit. The energy-harvesting
circuit can be improved to increase current levels
energy loss ratio. from the PFCB while decreasing voltage levels for
battery-charging purposes. The increase of current
As estimated above, the energy gain is 7.3 during the vibration of PFCB would decrease the
times smaller than the overall energy consumption battery- charging time by supplying more energy to
of the MP3 player in a day. The harvested and stored the electronic device. The special device should be
energy is not sufficient to run an MP3 player for one in the proper position so that the maximum amount
hour with one-hour of daily walking. The energy of vibration can be created for energy-harvesting
loss/gain graph depending on time given in Figure15 purposes. As the energy yield increases and wear-
to permit a visual comparison of these variables. able electronic devices become more efficient, foot-
powered energy scavenging systems can drive more
components of wearable computers, reducing the
Energy Harvesting From Passive Human Power 15

12. need for batteries, or enabling them to be charged
while the energy is being discharged. These sys- [10] “DC-DC Converter Basics,”2008. [Online].
tems provide for a host of applications in situations Available: http://www.powerdesigners.com/
where power resources are inaccessible, such as InfoWeb/design_center/articles/DC-DC/con-
verter.shtm [Accessed February 15, 2010]
during hiking expeditions or military missions. As
[11] “Bridge Rectifier,” 2008. [Online]. Available:
applications of energy harvesting from human pow-
http://hyperphysics.phy-astr.gsu.edu/Hbase/
er increase, so should ease of use. The piezoelectric electronic/rectbr.html [Accessed January 11,
application conditioning electronics to harvest en- 2010].
ergy from sneaker insoles could be commercialized [12] “Designing A SEPIC Converter: Introduction,
as separate components to be modularly linked to 2008. [Online]. Available: http://www.na-
electronics embedded in the heel via a weatherproof tional.com/an/AN/AN-1484.pdf [Accessed
connector. As shoe/sneaker insoles become worn, January 11, 2010].
they could be replaced by the consumer. With ap- [13] “Switching regulator, Glossary of Terms, 2008.
propriate adaptations, sneaker/shoe mounted en- [Online]. Available: http://www.elpac.com/
ergy harvesting systems are likely to power a wide resources/glossary/index. html. [Accessed
range of low power electronic devices. January 16, 2010].
[14] “LTspice/SwitcherCAD III, 2007. [Online].
Available: http://www.linear.com/design-
REFERENCES
tools/software/index.jsp#Spice [Accessed
[1] T. Starner and J.A., Paradiso, Human-generated February 14, 2010].
power for mobile electronics,”. Low-Power [15] “Linear Technology,” 2007. [Online]. Available:
Electronics Design, C. Piguet, Ed. CRC Press, http://www.linear.com [Accessed February 1,
New York, NY, chp 45, pp. 1–35, 2004. 2010].
[2] “Self-contained generating and lighting unit,” [16] “Maxim,” Operational Amplifiers, [Online].
U.S. Patent and Trademark Office, patent 2007. Available: http://www.maxim-ic.com/
1,184,056, 1916. appnotes.cfm/an_pk/1108 [Accessed Janu-
[3] FreePlay, Energy, 2009 [Online]. Available: ary 22, 2010].
http://www.freeplayenergy.com/page-view. [17] “LT1512 SEPIC constant-current/constant-
php?pagename=Products-copy&language=US voltage battery charger,” Linear Technology,
[4] H.A. Sodano, D.J. Inman, and G. Park, “A review 2007. [Online]. Available: http://cds.linear.
of power harvesting from vibration using com/docs/Datasheet/1512fa.pdf [Accessed
piezoelectric materials,” The Shock and Vi- January 1, 2010].
bration Digest, vol. 36, no 3, pp.197-205, [18] “Diodes,” 2008. [Online]. Available: http://
2004. www.kpsec.freeuk.com/components/diode.
[5] S. Roundy, and P.K. Wright, “A piezoelectric vi- htm [Accessed February 18, 2010].
bration based generator for wireless electron-
ics,” Smart Materials and Structures, vol. 13, Faruk Yildiz attended Taraz State University, Ka-
pp. 1131-1142, 2004. zakhstan, and graduated with a BS Computer Sci-
[6] D.A. Skoog, J.F. Holler, and S.R. Crouch, Prin- ence degree. He received his Master’s Degree from
ciples of Instrumental Analysis, 6th ed., Flor-
City University of New York with a Computer Sci-
ence, KY: Cengage Learning, Brooks/ Cole,
ence and completed his Doctoral degree from the
2006.
University of Northern Iowa in Industrial Technol-
[7] S. J. Roundy, Energy Scavenging for Wireless
Sensor Nodes with a Focus on Vibration to ogy. He is a faculty member in the Industrial Tech-
Electricity Conversion.” PhD dissertation, The nology Program at Sam Houston State University.
University of California, Berkeley, 2003. His research interests include energy harvesting,
[8] J.J. Thomas, W. W. Clark, “Harvesting Energy conversion, charging circuits, and ambient energy
from Piezoelectric Material,” IEEE, University sources. He is a current member of IEEE, ATMAE,
of Pittsburgh: IEEE CS, vol. 4, no.1, pp. 69-71 ASEE, EPT communities.
1536-1268, 2005.
[9] Product Overview, “Advanced Cerametrics In-
corporated,” 2007. [Online]. Available: http://
www.advancedcerametrics.com/pages/prod-
ucts [Accessed January 16, 2010]
16 j o u r n a l o f a p p l i e d s c i e n c e & e n g i n e e r i n g t e c h n o l o g y 2011