Original Article

Journal of Intelligent Material Systemsand Structures1–10� The Author(s) 2019Article reuse guidelines:sagepub.com/journals-permissionsDOI: 10.1177/1045389X19828512journals.sagepub.com/home/jim

Boosting the efficiency of a footsteppiezoelectric-stack energy harvesterusing the synchronized switchtechnology

Haili Liu1, Rui Hua2, Yang Lu3, Ya Wang2 , Emre Salman3 andJunrui Liang4

AbstractIn this article, the self-supported power conditioning circuits are studied for a footstep energy harvester, which consistsof a monolithic multilayer piezoelectric stack with a force amplification frame to extract electricity from human walkinglocomotion. Based on the synchronized switch harvesting on inductance (SSHI) technology, the power conditioning cir-cuits are designed to optimize the power flow from the piezoelectric stack to the energy storage device under real-timehuman walking excitation instead of a simple sine waveform input, as reported in most literatures. The unique propertiesof human walking locomotion and multilayer piezoelectric stack both impose complications for circuit design. Threecommon interface circuits, for example, standard energy harvesting circuit, series-SSHI, and parallel-SSHI, are comparedin terms of their output power to find the best candidate for the real-time-footstep energy harvester. Experimentalresults show that the use of parallel-SSHI circuit interface produces 74% more power than the standard energy harvest-ing counterpart, while the use of series-SSHI circuit demonstrates a similar performance in comparison to the standardenergy harvesting interface. The reasons for such a huge efficiency improvement using the parallel-SSHI interface aredetailed in this article.

Keywordspiezoelectric energy harvesting, synchronized switch energy harvesting on inductance, human walking locomotion, self-powered, piezoelectric stack

1. Introduction

Power demands are increasing everyday with increasingpurchase and usage of mobile electronics. Although thegreat progress in large-scale integrated circuit hasdecreased the power requirement of the mobile electro-nics, the stagnancy of modern battery technologies cre-ates the need for the alternative methods to meet thepower demands from the large number of the mobileelectronics (Paradiso and Starner, 2005). Human foot-steps are a great source of power, and previous studieshave revealed the forces generated from human foot-steps (Giakas and Baltzopoulos, 1997) along with theirpotential to generate power (Starner and Paradiso,2004). A healthy person is expected to walk 10,000steps a day and thus can output considerable power ifharvested efficiently. This power can be stored andused to charge a mobile phone afterward or directlypower the wireless sensors, such as a geo-positioningsensor and temperature sensor, without using a battery.

Electromagnet systems have much higher efficiencyin the mechanical-to-electrical energy transfer, while itusually involves coils and magnets (Cho et al., 2016;Ylli et al., 2015), which means a complicated and bulkymechanism is unavoidable. Piezoelectric materials candirectly produce charge when stress is applied on it,and thus, piezoelectric energy harvesting (PEH) is

1Department of Mechanical Engineering, Stony Brook University, Stony

Brook, NY, USA2J. Mike Walker ’66 Department of Mechanical Engineering, Texas A&M

University, College Station, TX, USA3Department of Electrical and Computer Engineering, Stony Brook

University, Stony Brook, NY, USA4School of Information Science and Technology, ShanghaiTech University,

Shanghai, China

Corresponding author:

Ya Wang, J. Mike Walker ’66 Department of Mechanical Engineering,

Texas A&M University, College Station, TX 77843, USA.

Email: ya.wang@tamu.edu

widely studied (Xie and Cai, 2014) for some integratedworking circumstances. In particular, a miniaturepiezoelectric-stack energy-harvesting device is one ofthe best replacements for battery, and it can be installedinside the insole to directly convert human walking andrunning motion into usable electrical form (Ylli et al.,2015). Among existing piezoelectric materials, lead zir-conate titanate (PZT) is one of the most efficient andcost-acceptable materials. Its piezoelectric coefficientd33 is usually two times of d31 and thus can output morepower in d33 mode when exposed to the same stress orstrain. The stand-alone piezoelectric multilayer stackworking at d33 mode is widely researched (Lee et al.,2014; Zhao and Erturk, 2014). However, it has rela-tively low conversion efficiency. Thus, force amplifica-tion frames are designed to improve the efficiencyperformance. There are two different mechanisms ofthe amplification frame: concave (Feenstra et al., 2008)and convex (Liu et al., 2014a) shape. In addition, amultilayer piezoelectric stack (Wang et al., 2016) with aflexure-free convex force amplification frame is used byour group to further improve the conversion efficiency.Besides the force amplification, semi-active vibrationdamping methods such as synchronized switch damp-ing techniques (Ji et al., 2010, 2012) are developed toimprove the conversion efficiency with advanced con-trol strategies.

Besides the energy harvesting structure, a typicalPEH device also needs a power conditioning circuit toconvert the alternate current (AC) power from thepiezoelectric element to directive current (DC) powerfor electronics. A full-bridge rectifier is the most com-monly used power conditioning circuit, and it isregarded as the standard energy harvesting (SEH) cir-cuit (Li et al., 2016). Yet, using SEH, there is a phasedifference between the voltage and current, and thus,they produce negative power, which means that someenergy flows back to the mechanical part of the system(Nechibvute et al., 2014). In order to avoid this phasedifference and boost the output power, synchronouselectric charge extraction (SECE; Lefeuvre et al., 2005)and synchronized switch harvesting on inductor (SSHI;Guyomar et al., 2005; Li and You, 2015; Wu et al.,2017) are developed as two efficient circuit interfaces.

It was proven that such interface circuits canincrease the energy harvesting capability by severalhundred percent. SECE was first addressed by Lefeuvreet al. (2005), and several optimized synchronous electriccharge extraction (OSECE; Wu et al., 2012; Xia et al.,2018) methods have been proposed to reduce powerloss and achieve self-powered switches (Liu et al.,2018a, 2018b) or reduce components with the design ofmechanical switches (Liu et al., 2014b; Wu et al., 2014).SECE has been proved to be better applicable for low-coupling piezoelectric harvesters, which is not applica-ble for strong-coupling piezoelectric stacks. Comparedto SECE, SSHI is suitable for all piezoelectric

harvesters, but sensitive to loads. In SSHI, we need tosense the maximum deforming instants of a piezoelec-tric transducer and simultaneously form an inductiveshortcut for the charge stored in the piezoelectric ele-ment, such that the voltage across the piezoelectricelement can be inverted after half of a resistance–inductance–capacitance (RLC) cycle. For thestand-alone application of SSHI, the sensing, synchro-nization, and switching functions are better to beself-contained without using any external sensor, powersupply, and controller. Given these requirements, sev-eral self-powered SSHI circuits have been invented(Cheng et al., 2016; Lallart and Guyomar, 2008; Liangand Liao, 2012), which consist of the voltage peakdetector, comparator, and electronic switch. Recently,some mechanical self-powered SSHI solutions (Huaet al., 2018; Liu et al., 2015) are proposed to eliminatesome electronic components used in the electronic self-powered SSHI as well as reduce the voltage drop, phaselag, and energy consumption caused by these electroniccomponents. These mechanical switches were designedfor base-excited cantilever beams or specific mechan-isms and therefore cannot be applied directly since thefootstep piezoelectric-stack harvester works in a differ-ent way.

So far, the SSHI technologies are not applied on thepiezoelectric stacks, which have strong coupling coeffi-cient and much higher impedance than piezoelectricpatches, and most of the studies (Guyomar and Lallart,2011; Qiu et al., 2009; Yang et al., 2015) are based onharmonic vibrations, and multiple-frequency excita-tions are seldom explored. On the contrary, in previousfootstep energy harvester study (Xie and Cai, 2014), thepiezoelectric stack outputs the AC power directly to alinear resistor. Due to the complexity of human walkingpattern, most of the footstep energy harvesting devicesin the literature still use a SEH interface (Liang andLiao, 2012), which has relatively low AC–DC energyconversion efficiency.

In this article, a monolithic multilayer piezoelectricstack is used as a transducer, and multi-frequencywaveforms are used to simulate real-time human walk-ing motion. Section 2 introduces the model of multiplefrequency excitations as the input force and the electro-mechanical coupling equations of the piezoelectric-stack harvester. In section 3, the circuit topology andworking principle of SEH and both self-powered series-SSHI (S-SSHI) and self-powered parallel-SSHI (P-SSHI) are introduced. Section 4 includes experimentalprocedures, results, and analyses. Conclusions andfuture work are discussed in section 5.

Experiments show that the parallel-SSHI circuit out-performs the series-SSHI and the SEH interface, in par-ticular, for the stack configuration. However, theparallel-SSHI does not show this obvious advantageover the series-SSHI for the single-layer-piezoelectric-patch configuration, which is commonly used in the

2 Journal of Intelligent Material Systems and Structures 00(0)

literature (Nechibvute et al., 2014; Qiu et al., 2009; Wuet al., 2017). The reasons are detailed throughout thearticle.

2. Work principles of the mechanical part

2.1. Force amplification frame

The piezoelectric stack in d33 configuration has a largerpiezoelectric coefficient and a higher efficiency while itis with greater stiffness and limited strain. The forceamplification frame can be used to multiply the inputforce from human walking. As shown in Figure 1, thepiezoelectric stack is surrounded by the force amplifica-tion frame. The Fin indicates the input force from walk-ing in the vertical direction, and the Fout indicates theoutput force in the horizontal direction. The detailedforce analysis has been done in our previous publica-tions (Chen et al., 2016; Wang et al., 2016).

2.2. Input force from the foot strike

Walking introduces a repetitive impact force, which isexerted on the ground. Although the force of walkinghas anteroposterior and medial–lateral components, thedominant part is the vertical component—often termedas the vertical ground reaction force (VGRF), which isthe easiest to be utilized and quantified (Keller et al.,1996). Studies have also revealed that the VGRF varieswith the gait speed and the body weight. Tongen andWunderlich (2010) had used a force plate to measurethe magnitude of VGRFs during walking of a male sub-ject weighing 77.51 kg. A continuous equation can bederived using curve-fitting method through the experi-mental data of human locomotion with the body weightabout 70 kg (Chen et al., 2016; Masani et al., 2002) andformulated as

y=Mb 3 ½10:53 sin 3:921x+ 0:1854ð Þ+ 1:9 sin 16:5x� 1:1114ð Þ�

ð1Þ

where Mb is the body weight of a person in kilograms.This means there are two peaks on the input forcecurves, as shown in Figure 2, which is different fromthe harmonic vibration excitation and its correspond-ing high-efficiency interface circuit is not well studied.

2.3. Modeling

In this article, the piezoelectric-stack energy harvester issimply equivalent to single degree of freedom (SDOF;Lee et al., 2014). According to the detailed modeling inour previous paper (Wang et al., 2016), the governingequations can be expressed as

€u+v2nu� nv2

nd33v=Fin

Meff

ð2Þ

_V +1

RLCp

V +nd33Meff v2

n

Cp

_u= 0 ð3Þ

where u is the axial displacement, n is the number ofpiezoelectric layers in one stack, Meff is the effectivemass of piezoelectric stack, d33 is the piezoelectric con-stant, Fin is the input force on the stack, v is the outputvoltage, RL is the external load, and vn is the naturalfrequency, which is defined as

vn =

ffiffiffiffiffiffiffiffiffikeff

Meff

sð4Þ

where keff is the effective stiffness of the piezoelectricstack, and Cp is the internal capacitance, which isdefined as

Cp =neT

33A

hð5Þ

where eT33 is the dielectric permittivity constant, A is the

cross-sectional area of each piezoelectric layer, and h isthe thickness of each piezoelectric layer.

3. Work principle of the conditioningcircuits

3.1. Conditioning circuits

The AC power generated by the piezoelectric stack can-not be directly used by many electric devices. Therefore,an AC–DC energy conditioning circuit between the

Figure 1. Diagram of the force amplification frame.Fin: input force to the force amplification frame; Fout: amplified output

force to piezoelectric stack.

Figure 2. Input force from human locomotion.

Liu et al. 3

piezoelectric electrical generator and the energy storageis necessary. In order to reach higher energy conversionefficiency, there are many different schemes of electricinterfaces. In this work, three interfaces including SEH,series-SSHI, and parallel-SSHI are studied.

Figure 3(a) shows the schematic diagram of SEH,which includes a full-bridge rectifier, a filter capacitorCL, and a resistive load RL. Since footstep energy har-vester usually works at low-frequency range (around1.5–2.3 Hz) and this working frequency is far from theresonance of the stack, the piezoelectric element can bemodeled as an equivalent current source in parallelwith a capacitor Cp and the internal leakage resistanceRp. The rectifier is in a blocked state when piezoelectricoutput voltage is lower than the voltage across thecapacitor, and when the piezoelectric output voltage islarger than the voltage across the capacitor, the currentwill flow through the diodes and power the load. Theshortage of SEH circuit is that the output voltage andcurrent cannot keep the same phase and would producesome negative power, which means the harvestedenergy may return to the mechanical part, losing someof the harvested power (Liang and Liao, 2009).

Figure 3(b) and (c) illustrates the schematic diagramof the parallel-SSHI and the series-SSHI, respectively.The SSHI overcomes the energy return issue by addinga switch path, which contains a switch S1 and an induc-tor L1. The switch S1 turns on at the short instant whenoutput voltage reaches a maximum or minimum, whichensures the output voltage and current have the samephase and thus prevents the harvested power fromturning back. The only difference between the parallel-SSHI and series-SSHI is whether the switch path isconnected in parallel with the rectifier or it is connectedin series with the rectifier. Both of these SSHI interfacescan increase the output voltage from the piezoelectricstack due to a voltage inversion process (Lallart andGuyomar, 2008).

3.2. Self-powered SSHI circuit

In SSHI circuits, the switch S1 needs to turn on/off atthe desired time instant, which involves sensing the

voltage extreme point, switching on/off, and invertingthe voltage across the piezoelectric elements. Electronicbreaker can sense the voltage extreme point and thuscan form a self-powered SSHI interface (Lallart andGuyomar, 2008). Without the requirement of anyexternal power supply, it can automatically performswitching actions once the output voltage from thepiezoelectric stack reaches its extremum. Figure 4shows the schematic of self-powered series-SSHI, andTable 1 demonstrates related component models andvalues. One breaker can only allow the current flow inone direction, and therefore, each SSHI circuit employstwo such breakers.

In the self-powered series-SSHI, the current comesfrom the piezoelectric transducer. When the output vol-tage across the piezoelectric stack increases, C2 will becharged, while T1 and T2 are blocked; when the outputvoltage reaches an extreme point, and starts decreasing,a voltage difference appears between the emitter andthe base of T1. When this difference is greater than thethreshold voltage of T1, T1 will conduct. At the sametime, C2 provides the base voltage for T2 because D1 isblocked. The conduction of these two transistors T1

and T2 initiates the inductance–capacitance (LC) reso-nant circuit, so that Cp will be quickly dischargedthrough D3, T2, RL, D12, and L1. With the presence ofinductor L1, the voltage across Cp will be inverted. Onthe other hand, when the voltage reaches the negativeextremum, similarly, it will also be inversed due to sym-metrical circuit topology.

The schematic of self-powered parallel-SSHI isshown in Figure 5. The self-powered series-SSHI has a

Figure 3. Schematic diagram of (a) the standard energy harvesting circuit, (b) the parallel-SSHI circuit, and (c) the series-SSHIcircuit.

Table 1. Components for the self-powered series-SSHI andparallel-SSHI.

Component Model/value

Dx 1N4004T1 and T3 TIP32CT2 and T4 TIP31CL1 120 mH 97 OC2 and C4 23 nFCL 47 mF

4 Journal of Intelligent Material Systems and Structures 00(0)

similar working principle as the self-powered parallel-SSHI. One minor difference is that in the parallel-SSHI, when Cp is discharging, the current only passesD3, T2, and L1 (or D6, T3, and L1), which means that RL

does not influence the performance of the voltageinverting process.

4. Experiments and analyses

Figure 6 shows the experiment setup used to measurethe output voltage and output power from the piezo-electric stack. The monolithic multilayer piezoelectricstack is SP505 manufactured by Smart Material. Itconsists of 300 layers of piezoelectric patches, and eachlayer is 0.1 mm in thickness. Its dimension is 7 mm3 7 mm 3 32 mm, and the internal capacitance Cp is2.1 mF. As shown in Figure 6(b), the piezoelectric stackis placed in the amplification frame. The simulatedinput force signal is sent to NI DAQ and amplified by

power amplifier to drive the shaker (APS 420). Theshaker is used to provide excitations to the amplifica-tion frame on one side through its moving armature,and the other side of the frame is fixed. A force trans-ducer is placed between the amplification frame andthe shaker to measure the input force. Another forcetransducer is placed between the interface of thepiezoelectric-stack harvester and the amplificationframe to measure the amplified force. Both force trans-ducers connect to a charge amplifier to send amplifiedsignals to NI DAQ. When different interface circuitsconnect to the harvester, the output voltage from thecircuit is measured with the oscilloscope. By alternatingthe load (resistance value), we can test the outputpower for further analysis.

In order to simulate the force excitation of humanlocomotion, the input force adopts the curve-fittedfunction in equation (1), where Mb is the body weightof a person in kilograms. Considering the shaker

Figure 4. Schematic diagram of the self-powered series-SSHI circuit.

Figure 5. Schematic diagram of the self-powered parallel-SSHI circuit.

Liu et al. 5

cannot output the same level of excitations as humanlocomotion, the amplitude of the input force is set toaround 114 N, so that the maximum output force ofthe shaker is used in the testing. Figure 7 shows thewaveforms of one cycle of the input force applied onthe force amplification frame and input force appliedon the piezoelectric stack. These data are measured bya force transducer with a charge amplifier. From the

figure, we can see that the force applied on the frame isabout 114 N, and the amplified force on the stack isabout 846 N; thus, the amplification ratio calculatedfrom one cycle is about 7.4.

With the amplified force transferred to the piezoelec-tric stack, the voltage will be generated across thestack. In the experiments, the stack is connected toa SEH circuit, a self-powered series-SSHI circuit, or aself-powered parallel-SSHI circuit, respectively, using aseries of load resistance RL. Figure 8 shows the outputvoltage across piezoelectric stack under different loadresistance. From the waveforms, we can see that for theself-powered series-SSHI circuit, when the value of loadresistance increases, the voltage amplitude decreases,and the performance of voltage inversion processbecomes worse; for the self-powered parallel-SSHI cir-cuit, when the value of load resistance increases, thevoltage amplitude increases, and the value of load resis-tance RL does not affect its voltage inversion process.

The output power can be calculated by measuringthe voltage U across the different load resistance RL

P=U2

RL

ð6Þ

Figure 6. Experiment setup of (a) voltage and power measurement—ffi shaker, ffl oscilloscope,� charge amplifier, Ð circuitboard,ð resistance box, Þ NI DAQ, þ power amplifier, ¼ differential probe; and (b) the tested stack and frame; the spacer andforce transducers—ffi force transducer, ffl spacer, � frame amplification, Ð piezoelectric stack,ð shaker.

Figure 7. Force input waveform from experimentalmeasurement.

6 Journal of Intelligent Material Systems and Structures 00(0)

Figure 9 compares the output power of thethree interface circuits at different load resistances.Figure 9(a) uses the resistance as x axis, named asresistance–power curve, and can directly indicate theoptimum resistance; Figure 9(b) uses the voltage acrossthe load resistance as x axis, named as voltage–powercurve, and can be used to verify the optimum power

being correctly selected. Figure 9 shows that the SEHand the series-SSHI have the similar performance,while parallel-SSHI is the best under different loadresistances. The values of the optimal power output ofthe three interface circuits are compared in Table 2. Inparticular, the SEH circuit provides the optimal poweroutput 1:35 mW at RL = 50 kO, which is similar to theseries-SSHI circuit (1:33 mW at RL = 60 kO). However,the parallel-SSHI circuit outputs the power of 2:35 mWat RL = 80 kO, and thus, the harvesting efficiencyincreases by 74%, in comparison to that of the SEHcircuit.

Also, Figure 10 shows the output voltage waveformsof the two interface circuits with the optimal load. Atthe resistive load of 80 kO, the parallel-SSHI has theoptimal output power, while at the resistive load of60 kO, the series-SSHI has the optimal output power.

5. Discussion

According to many previous studies (Niazi andGoudarzi, 2015; Pang et al., 2016), series-SSHI andparallel-SSHI should both bring a significant improve-ment and the performance is almost the same. Why theseries-SSHI circuit does not work well in this footsteppiezoelectric-stack energy harvester? Some factors maycause this result: first, the optimum load resistance forthe two SSHI interfaces is around 50–70 kO (given inTable 2), which is determined by the piezoelectric stackand its working condition; from Figure 8, we can seethat the voltage is not inverted well for series-SSHI atthis range of RL, which undermines the performance ofthe SSHI technique, while parallel-SSHI can invert thevoltage well at high load resistance and thus shows bet-ter performance. In series-SSHI circuit, RL is locatedwithin the discharge loop. Thus, when the value of theload resistance RL increases, the discharging speeddecreases, which degrades the voltage inversion

Figure 8. Output voltage across piezoelectric stack with(a) series-SSHI circuit and (b) parallel-SSHI circuit underdifferent load resistance: 0 O, 10 kO, 100 kO, and 1 MO.

Figure 9. Comparison of the output power of the three interface circuits at different load resistance: (a) resistance–power curveand (b) voltage–power curve.

Liu et al. 7

process. Second, the input force from human locomo-tion is not a simple harmonic wave, as usually studied.The experimental results show that the series-SSHIcannot invert the voltage at some small peaks of thevoltage waves, as shown in Figure 8, which means somenegative power would be produced and thus underminethe performance of the SSHI, decreasing the efficiencyto around the SEH counterpart. Another test was car-ried out to rule out the input dependence on the perfor-mance superior of the parallel-SSHI in comparisonwith its series-SSHI counterpart. Results are shown inFigure 11, indicating the power output of the three

interface circuits at their respective optimal load resis-tance, under a harmonic excitation. Their waveformsshow that the parallel-SSHI still performs better thanthe series-SSHI under a harmonic excitation.

6. Conclusion

Aiming to take advantage of the great energy sourceof human locomotion, the high-efficiency energy har-vesting circuits, that is, self-powered series-SSHI andself-powered parallel-SSHI, are discussed and experi-mentally tested for the footstep piezoelectric-stackenergy harvester. According to the unique characteris-tics of the footstep piezoelectric-stack energy harvester,including low frequency, multiple frequency, strongcoupling, with a high internal resistance, and a largecapacitance, the parameters of the self-powered SSHIare redesigned. Contrary to the most studied piezoelec-tric patch energy harvester under a base excitation,experiments show that the self-powered parallel-SSHIperforms much better than the self-powered series-SSHI. Both of them are compared with the SEH circuitin terms of harvested energy. The results show that theparallel one can increase the harvested power by a fac-tor of 174%, while the series one outputs almost thesame power as the SEH circuit. This big difference isdue to the unique characteristics of the footsteppiezoelectric-stack energy harvester. These results aremeaningful for the harvesting circuit selection and thefurther real-life applications of the footsteppiezoelectric-stack energy harvester.

Acknowledgements

The authors are thankful to their colleague SiddharthBalasubramanian who provided the structure design of theframes.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest withrespect to the research, authorship, and/or publication of thisarticle.

Funding

The author(s) disclosed receipt of the following financial sup-port for the research, authorship, and/or publication of this

Table 2. Output power using different interface circuits.

Circuit type Optimal load resistance (kO) Voltage across the storage capacitor (V) Optimized power output (mW)

SEH 50 8.22 1.35Series-SSHI 60 8.92 1.33Parallel-SSHI 80 13.7 2.35

SEH: standard energy harvesting.

Figure 10. Voltage output of two interface circuits at theirrespective optimal resistive loads: parallel-SSHI at 80 kO andseries-SSHI at 60 kO.

Figure 11. The power output of three interface circuits undera harmonic excitation.

8 Journal of Intelligent Material Systems and Structures 00(0)

article: The study was supported by the DOE ARPA-E (grantno. DE-AR0000945).

ORCID iD

Ya Wang https://orcid.org/0000-0003-4353-4638

References

Chen W, Wang Y and Deng W (2016) Deformable force

amplification frame promoting piezoelectric stack energy

harvesting: parametric model, experiments and energy

analysis. Journal of Intelligent Material Systems and Struc-

tures 28: 827–836.Cheng CC, Chen ZS, Xiong YP, et al. (2016) A high-effi-

ciency, self-powered nonlinear interface circuit for bi-stable

rotating piezoelectric vibration energy harvesting with

nonlinear magnetic force. International Journal of Applied

Electromagnetics and Mechanics 51: 235–248.Cho HS, Yang JH, Park SH, et al. (2016) An exploratory

study on the feasibility of a foot gear type energy harvester

using a textile coil inductor. Journal of Electrical Engineer-

ing & Technology 11: 1210–1215.Feenstra J, Granstrom J and Sodano H (2008) Energy har-

vesting through a backpack employing a mechanically

amplified piezoelectric stack. Mechanical Systems and Sig-

nal Processing 22: 721–734.Giakas G and Baltzopoulos V (1997) Time and frequency

domain analysis of ground reaction forces during walking:

an investigation of variability and symmetry. Gait & Pos-

ture 5: 189–197.Guyomar D and Lallart M (2011) Recent progress in piezo-

electric conversion and energy harvesting using nonlinear

electronic interfaces and issues in small scale implementa-

tion. Micromachines 2: 274–294.Guyomar D, Badel A, Lefeuvre E, et al. (2005) Toward

energy harvesting using active materials and conversion

improvement by nonlinear processing. IEEE Transactions

on Ultrasonics Ferroelectrics and Frequency Control 52:

584–595.Hua R, Liu H, Yang H, et al. (2018) A nonlinear interface

integrated lever mechanism for piezoelectric footstep

energy harvesting. Applied Physics Letters 113: 053902.Ji H, Qiu J, Cheng J, et al. (2010) Application of a negative

capacitance circuit in synchronized switch damping tech-

niques for vibration suppression. In: Proceedings of SPIE:

active and passive smart structures and integrated systems,

San Diego, CA, 12 April.Ji H, Qiu J, Xia P, et al. (2012) Analysis of energy conversion

in switched-voltage control with arbitrary switching fre-

quency. Sensors and Actuators A: Physical 174: 162–172.Keller T, Weisberger A, Ray J, et al. (1996) Relationship

between vertical ground reaction force and speed during

walking, slow jogging, and running. Clinical Biomechanics

11: 253–259.Lallart M and Guyomar D (2008) An optimized self-powered

switching circuit for non-linear energy harvesting with low

voltage output. Smart Materials and Structures 17: 035030.Lee AJ, Wang Y and Inman DJ (2014) Energy harvesting of

piezoelectric stack actuator from a shock event. Journal of

Vibration and Acoustics 136: 011016.

Lefeuvre E, Badel A, Richard C, et al. (2005) Piezoelectric

energy harvesting device optimization by synchronous

electric charge extraction. Journal of Intelligent Material

Systems and Structures 16: 865–876.Li B and You JH (2015) Experimental study on self-powered

synchronized switch harvesting on inductor circuits for

multiple piezoelectric plates in acoustic energy harvesting.

Journal of Intelligent Material Systems and Structures 26:

1646–1655.Li Q, Kuang Y and Zhu M (2016) System-level modelling and

validation of a strain energy harvesting system by directly

coupling finite element and electrical circuits. In: IEEE sen-

sors, Orlando, FL, 30 October–3 November, pp. 1–3. New

York: IEEE.Liang J and Liao WH (2009) Piezoelectric energy harvesting

and dissipation on structural damping. Journal of Intelli-

gent Material Systems and Structures 20: 515–527.Liang J and Liao WH (2012) Improved design and analysis

of self-powered synchronized switch interface circuit for

piezoelectric energy harvesting systems. IEEE Transactions

on Industrial Electronics 59: 1950–1960.Liu H, Liang J and Ge C (2015) A mechatronic power boost-

ing design for piezoelectric generators. Applied Physics Let-

ters 107: 141902.Liu J, O’Connor WJ, Ahearne E, et al. (2014a) Electromecha-

nical modelling for piezoelectric flextensional actuators.

Smart Materials and Structures 23: 025005.

Liu W, Badel A, Formosa F, et al. (2018a) Comparative case

study on the self-powered synchronous switching harvest-

ing circuits with BJT or MOSFET switches. IEEE Trans-

actions on Power Electronics 33: 9506–9519.Liu W, Badel A, Formosa F, et al. (2018b) A comprehensive

analysis and modeling of the self-powered synchronous

switching harvesting circuit with electronic breakers. IEEE

Transactions on Industrial Electronics 65: 3899–3909.Liu W, Formosa F, Badel A, et al. (2014b) Self-powered non-

linear harvesting circuit with a mechanical switch structure

for a bistable generator with stoppers. Sensors and Actua-

tors A: Physical 216: 106–115.Masani K, Kouzaki M and Fukunaga T (2002) Variability of

ground reaction forces during treadmill walking. Journal

of Applied Physiology 92: 1885–1890.Nechibvute A, Luhanga P and Chawanda A (2014) Experi-

mental study of energy harvesting using a self-powered

SSHI interface circuit. International Journal of Engineering

and Technology 4: 455–462.Niazi K and Goudarzi M (2015) Pyroelectric effect in PZT

and PMN-0.25 PT ceramics for hybrid energy harvesting

in different interface circuits. German Journal of Renewable

and Sustainable Energy Research 1: 23–29.Pang S, Li W and Kan J (2016) Optimization analysis of inter-

face circuits in piezoelectric energy harvesting systems.

Journal of Power Technologies 96: 1–7.Paradiso JA and Starner T (2005) Energy scavenging for

mobile and wireless electronics. IEEE Pervasive Computing

4: 18–27.Qiu J, Jiang H, Ji H, et al. (2009) Comparison between four

piezoelectric energy harvesting circuits. Frontiers of

Mechanical Engineering in China 4: 153–159.Starner T and Paradiso J (2004) Human generated power for

mobile electronics. Low-Power Electronics Design 45: 1–35.

Liu et al. 9

Tongen A and Wunderlich RE (2010) Biomechanics of run-ning and walking. Mathematics and Sports 43: 1–12.

Wang Y, Chen W and Guzman P (2016) Piezoelectric stackenergy harvesting with a force amplification frame: model-ing and experiment. Journal of Intelligent Material Systems

and Structures 27: 2324–2332.Wu L, Do XD, Lee SG, et al. (2017) A self-powered and opti-

mal SSHI circuit integrated with an active rectifier forpiezoelectric energy harvesting. IEEE Transactions on Cir-

cuits and Systems I: Regular Papers 64: 537–549.Wu Y, Badel A, Formosa F, et al. (2012) Piezoelectric vibra-

tion energy harvesting by optimized synchronous electriccharge extraction. Journal of Intelligent Material Systems

and Structures 24: 1445–1458.Wu Y, Badel A, Formosa F, et al. (2014) Nonlinear vibration

energy harvesting device integrating mechanical stoppersused as synchronous mechanical switches. Journal of Intel-

ligent Material Systems and Structures 25: 1658–1663.

Xia H, Xia Y, Ye Y, et al. (2018) Analysis and simulation ofsynchronous electric charge partial extraction techniquefor efficient piezoelectric energy harvesting. IEEE Sensors

Journal 18: 6235–6244.Xie LH and Cai MJ (2014) Increased piezoelectric energy har-

vesting from human footstep motion by using an amplifi-cation mechanism. Applied Physics Letters 105: 143901.

Yang YYW, Shi X, Lan HR, et al. (2015) Investigation onbehavior of the vibration-based piezoelectric energy har-vester array in ultracapacitor charging. Applied Physics

Letters 106: 173902.Ylli K, Hoffmann D, Willmann A, et al. (2015) Energy har-

vesting from human motion: exploiting swing and shockexcitations. Smart Materials and Structures 24: 025029.

Zhao S and Erturk A (2014) Deterministic and band-limitedstochastic energy harvesting from uniaxial excitation of amultilayer piezoelectric stack. Sensors and Actuators A:

Physical 214: 58–65.

10 Journal of Intelligent Material Systems and Structures 00(0)