Sensors and Actuators A: Physical

Transcription

1 Sensors and Actuators A 57 (7) 98 7 Contents lists available at ScienceDirect Sensors and Actuators A: Physical j ourna l ho me page: A wearable energy harvester unit using piezoelectric electromagnetic hybrid technique Rawnak Hamid, Mehmet Rasit Yuce Department of Electrical and Computer Systems Engineering, Monash University, Melbourne, VIC 38, Australia a r t i c l e i n f o Article history: Received 3 August 6 Received in revised form 9 January 7 Accepted February 7 Available online 4 February 7 Keywords: Piezoelectric energy harvesting Electromagnetic energy harvesting Wearable energy harvesting Wearable sensors a b s t r a c t Wearable sensor electronics require a sustainable electrical power supply to operate. Energy harvesting techniques can be used to convert available nonelectrical energy sources into electrical energy. This paper presents WE-Harvest system, which is a new wearable energy harvesting system that combines piezoelectric and electromagnetic energy harvesters in one unit to generate a combined electrical energy source. Piezoelectric transducers are used to obtain sufficient regulated output voltages while electromagnetic is employed for its high power generation capability. Regular human body motions provide input vibrations for the proposed energy harvester unit. Several conditioning circuit topologies are proposed to efficiently extract energy from the two sources. The experimental results demonstrate that the combined topology enhances the power generation efficiency as well as enables stable output DC voltages. The dependence of energy harvester output on the load and input frequency has also been investigated. 7 Elsevier B.V. All rights reserved.. Introduction Wearable sensor devices are designed to operate around or at close proximity to the human body [,]. A critical challenge for wearable sensor devices is the limited energy source. The required electric energy is supplied by small batteries, which should be regularly monitored for their charge status, and need to be frequently recharged or replaced that may not always be possible. The solution to successfully develop a self-powered wireless wearable sensor relies on energy harvesting techniques. Energy harvesting-based wearable sensors have the capability of extracting energy from the surrounding environment and from the human body using the sources such as motion and heat. Recent advances in low power electronics have reduced the required electrical power for operating sensors devices to several microwatts [3,4]. Hence, the energy generated by harvesting from ambient sources like heat, solar energy, wind, vibration and radio frequency waves has been investigated to power such devices [3,5,6]. There have been attempts utilizing solar harvesting/scavenging techniques using solar photovoltaic (PV) cells for generating energy for wearable sensors. For example, the wear- Corresponding author. addresses: rawnak.hamid@monash.edu (R. Hamid), mehmet.yuce@monash.edu (M.R. Yuce). able energy harvesting technique in [6] harvests 77 W from a flexible PV cell, with the dimension of 6 mm 7 mm. These harvesters may not always be reliable due to the unpredictable light conditions in indoor environment. People with wearable devices are mobile and therefore energy harvesting techniques for wearable technology should be suitable for different environments to generate continuous and reliable energy. Wearable thermoelectric converters can generate similar power levels achievable by solar PV cells in indoor [5]. However, these converters must be in thermal contact with the skin for an efficient energy harvesting. In this paper, a new vibration-based wearable energy harvester that harvests energy from motion and movements of the human body is presented.. Vibration energy harvesting methods Vibration energy harvesting techniques include piezoelectric, electro-magnetic and electrostatic transductions. Capacitive changes of the vibration-dependent capacitors are used for electrostatic energy harvesting. Mechanical energy is converted to electrical energy when vibrations move the initially charged plates [7,8]. A separate voltage source is needed for pre-charging the capacitor plates in electrostatic energy harvesters, which adds more complication for use in wearable applications. In piezoelectric energy harvesting, mechanical stress and strain results in structural deformation, which is then converted to electrical energy. Piezohttp://dx.doi.org/.6/j.sna / 7 Elsevier B.V. All rights reserved.

2 R. Hamid, M.R. Yuce / Sensors and Actuators A 57 (7) Fig.. Functional block diagram of the propsed hybrid energy harvesting device. electric harvesters may use one or multiple cantilever beams that vibrate at different frequencies. An efficient piezoelectric energy harvesting technique is to place piezoelectric transducers around walls of a box, and a free moving mass or ball is also placed in the box, which collides with piezoelectric transducers [9,]. A moving ball in a box hits one or multiple piezoelectric transducers generating electrical energy in D (two dimensions) or 3D. Multiple piezoelectric transducers provide a much wider range of resonant frequencies, generating higher energy levels. The electromagnetic induction is based on relative motions between a magnetic field and a coil, which causes electric current to flow in the coil [,]. These vibrational energy harvesters can use a moving permanent magnet like a ball that moves arbitrarily within a unit winded with coils. It is also possible to attach permanent magnet to a piezoelectric cantilever which moves in coil windings [3]. The amount of generated electricity depends on the strength of the magnetic field, the velocity of the relative motion and the number of turns of the coil used []. Some of these vibration energy harvesting techniques have been applied to wearable applications considering the vibration and motion of the human body [,,,4,5]. The ubiquitous presence of vibration energy sources characterized by low acceleration and low frequency in our daily lives provide promising energy sources, which can be harvested by vibration based energy harvesters [,6,7]. An energy harvesting approach that can continuously be used will benefit from a hybrid approach where multiple energy harvesting techniques are combined. The development of hybrid structures that can combine the distinct attributes of individual energy harvesting techniques at the same time, is the part of recent efforts in the design of high efficiency energy harvesters [3,8,9]. These coupled piezoelectric electromagnetic hybrid vibration energy harvesters use a piezoelectric cantilever beam with a magnet attached on its tip end. The studies undertaken in [3,8] present analysis and detailed optimization when such techniques are used. However, the measurement results in these studies did not consider the conditioning circuits. In addition, these studies did not consider wearable applications. This work proposes a hybrid-energy harvesting technique that is suitable for wearable systems. The proposed technique combines a magnet that moves freely with human motion inside a housing winded with coils, and hits piezoelectric transducers, which is housed in the same unit, to facilitate a continuous power source for wearable sensors (Fig. ). An initial prototype has been presented in a conference proceeding [5] with some initial results. This work provides a detailed study of the proposed technique with conditioning circuits and experimental results for wearable scenarios with a completely redesigned and smaller harvester unit. Electromagnetic energy harvesters produce high output currents but low-level output voltages while piezoelectric devices generate high voltages, with the expense of lower-output currents. The design in this paper takes into account these attributes when Fig.. Wearable harvester unit (of dimension 38.5 mm 34 mm 37 mm) designed for harvesting energy from electromagnetic and piezoelectric transducer at the same time. combining these two harvesting techniques in one unit for an efficient power generation method for wearable energy harvesting systems. Vibration-based power generators convert the mechanical energy into electrical energy in the form of alternative current (AC). The AC power is rectified to produce a stable DC (direct current) power using a power conditioning circuit since electronic devices and rechargeable batteries require a DC power supply. Therefore, the performance of energy harvesting systems depends on both the transducer and power conditioning electronics, which has been discussed in detail in this paper. The operation principle of the proposed system, the design procedure and experimental results are described in sections III and IV, respectively. 3. Proposed wearable energy harvester 3.. Electromagnetic and piezoelectric coupled energy harvester Typically energy harvesting systems consist of a transducer that converts non-electrical energy to electrical energy and a power conditioning circuit that converts the AC to DC, and adjusts the output voltage level to a corresponding load. The functional block diagram in Fig. shows the operation principle of the energy harvester presented in this work. The proposed hybrid approach is designed to combine the electromagnetic induction and piezoelectric effect to convert the mechanical vibrations to electrical energy. The transducer and conditioning circuits for each technique are introduced next. 3.. Harvester design The mechanical structure of the proposed energy harvesting transducer is presented in Fig., which features a magnet that can slide freely inside the housing when vibrated. The piezoelectric transducers are placed on the walls, at either end of the housing. The copper wire is wound around the housing to convert the magnetic field changes initiated by the magnet motions to electric current. Human motions provide the vibration, which is needed to move the magnets back and forth, inducing current by electromagnetic induction, according to Faraday s Law. Simultaneously, the magnets hit the piezoelectric transducers when they reach the housing wall at both sides. Kinetic energy of the magnet applies a mechanical force to the piezoelectric transducer unit that is converted to electrical energy. Four T (Lead Zirconate Titanate) diaphragms are electrically connected in parallel and mechanically fixed on a thin polyacetate sheet at two ends of the housing to improve the total amount of energy harvested by the T transducers. The parallel configuration

3 R. Hamid, M.R. Yuce / Sensors and Actuators A 57 (7) 98 7 Fig. 3. (a) Two stage Dickson multiplier used for electromagnetic energy harvester (b) Block diagram of the electrical circuitry inside the LTC3588- chip used [3]. enhances the electrical power since it increases the total output current while keeping the voltage constant. The device consists of a hollow box for the transducers, which can slide into and out of a bobbin wrapped with copper wire, and a key, to keep it in place. This design provides access to the transducers, and it does not require us to unwind the coils every time Electromagnetic () energy harvester The electromagnetic portion of the harvester operates according to Faraday s Law. As the magnet moves within the harvester unit in Fig., it creates a time-varying magnetic flux, inducing a current in the copper coil. In this case, a rare earth neodymium rectangular magnet of strength 4 G was used because of its stable and resistance to demagnetization characteristics []. The housing was then designed to fit the magnet such that it could only slide back and forth; and copper coils were wound around it. Hence the flux density can be considered to be uniform over the area of the coil, and the induced voltage is given by []: V = NA db dt sin( ) () where V is the generated voltage or induced emf, N is the number of turns in the coil, A is the loop area, B is the magnetic field flux density over the area of each turn, is the angle between the coil area and the flux density direction. When the device is moved at a moderate pace, it takes about. s for the magnet to complete the whole length of the box. For the maximum e.m.f., = 9, and is achieved when the magnet comes to the center of the device, and the magnetic field lines are perpendicular to the coil. The number of turns of coils were then calculated to induce an e.m.f. of 3 V (peak). Vt N = AB sin( ) = 3. (.6.35) (.4) sin(9) = 9 () The amplitude of the voltage output from electromagnetic () harvester is not large enough to always overcome the rectifier threshold voltage. Therefore, two cascaded voltage doubler stages are employed in the proposed system to enhance and rectify the output voltage. Due to the voltage multiplier, it was found that 35 turns of coil was sufficient to generate the desired output voltage. The internal resistance of the electromagnetic harvester is measured to be. Fig. 3(a) shows the two stage Dickson multiplier schematic used. The cascaded stage can generate four times larger output. The diodes maximum voltage drop is 4 mv at ma forward current []. The number of multiplier stages required to achieve the Fig. 4. Simulation result for modified Dickson Multiplier for fast and moderate speed. desired output voltage ( 3.3V) is found by simulations as shown in Fig. 4. Two plots were presented, one when the device is moved at a moderate speed (bottom plot), and one when it is moved at a fast speed (top plot), representing walking to running activities. This is to make sure that we will receive sufficient voltage values from the electromagnetic harvesters when worn on the human body Piezoelectric energy harvester The output voltage of a piezoelectric energy harvester is proportional to the frequency of vibration in addition to its amplitude [5]. The frequency dependence can be explained by the capacitive structure of the piezoelectric device. Short time periods of high frequency inputs avoid the capacitive discharge on these devices. Unlike the output from the electromagnetic harvester, the output voltages of piezoelectric transducers were high for the pro-

4 R. Hamid, M.R. Yuce / Sensors and Actuators A 57 (7) 98 7 Piezoelectric Piezoelectric LTC3588- Vpz Vcoup 6.8μF Vem Electromagne c Input (AC) 6.8μF 6.8μF 6.8μF Storage Capacitor Load Fig. 5. Voltage outputs form piezo diaphragms. Fig. 6. Power conditioning circuit for topology (unregulated output). posed application; therefore a multiplier circuit similar to Fig. 3 was not needed. Fig. 5 is an experimental result for voltage outputs obtained from the four piezoelectric transducers connected in parallel (shown in Fig. ), which are sufficiently high for the targeted sensor devices. Each discrete voltage value is due to the hitting of the magnet to piezoelectric sensors during human body motion (e.g. normal walking). Connecting piezoelectric transducers in parallel increases their current output while keeping the output voltage constant. The transducer AC output (due to the high voltage values) is rectified and regulated using the low-power rectifier and regulator circuits in the off-the-shelf chip (LTC3588-) [3] Power conditioning circuits Since the output power generated by both mechanisms is AC, they need to be rectified using a suitable electrical circuitry. Three possible electrical circuit topologies are considered and tested in order to combine the two different energy harvesters- electromagnetic and piezoelectric harvesters. The schematics of the three power conditioning circuits used are explained in this section. Vpz, Vem and Vcoup represents the nodes at which the outputs from the piezoelectric, electromagnetic and coupled energy harvester were measured respectively. In all the topologies, the output voltages of piezoelectric transducers are rectified and regulated by the power electronic circuits in LTC3588- which includes useful circuits such as a bridge rectifier, a buck converter, and regulator [3], as shown in Fig. 3(b). Topology : In this topology, the rectifier and regulator in the off-the-shelf component LTC3588- is used to rectify and regulate the signals generated by the piezoelectric transducers (Fig. 6). The Dickson multiplier in Fig. 3 is used for the voltage output from the electromagnetic harvester. A storage capacitor (6.8 F) is used to store the energy generated by the two harvesters. The reverse leakage of current from each harvester to the other is avoided by using diodes between harvester outputs and large capacitor. The final output of power conditioning circuitry is used to supply various loads. Further details are addressed in the experiment section. Topology : The output voltage of the coupled harvester technique is regulated using the LTC3588- chip. This topology is investigated in case a regulated voltage output is required for operation of a sensor device. The LTC3588- chip has a DC input connection, Vin, which was used to feed the rectified signal from the harvester to the chip (Fig. 7(a)). The rectified signal (DC signal) is then combined internally with the rectified piezoelectric signal before the regulation. The final output is stored in a single storage capacitor (6.8 F). The LTC3588- chip can output signals at selectable voltages of.8,.5, 3.3, and 3.6 V which are operational voltages for many low-power sensor devices. Fig. 7(b) and (c) shows the circuitry used in this topology for the piezoelectric and electromagnetic energy harvesters respectively. Fig. 7. Power conditioning circuit for (a) Topology (regulated output) (b) Piezoelectric harvester (c) Electromagnetic harvester.

5 R. Hamid, M.R. Yuce / Sensors and Actuators A 57 (7) 98 7 Fig. 8. Power conditioning circuit for topology 3 (unregulated output using separate LTC3588- chips). Fig. 9. Wearable piezoelectric electromagnetic energy harvester, attached to a person s wrist, ankle and just under the knee cap. Topology 3: Two separate LTC3588- chips are used in this topology (one for each energy harvesting method) as shown in Fig. 8. The regulated voltages are fed to two capacitors connected in series through diodes, which are used to prevent reverse leakage of current. Although the output of this power conditioning circuit is not regulated, the output can be arranged to produce a range of stable DC voltage levels. It is useful for sensors that can operate with different voltage supplies. Further details are addressed in the experiment section. 4. Experiments The system was integrated using three different circuit topologies that have been explained in the previous section. In this section, different experiments are performed for each topology, and the results are presented. Fig. 9 shows the prototype designed in this work that integrates electromagnetic and piezoelectric energy harvesting techniques in one unit, and the system was integrated using three different circuit topologies that have been explained in the previous section. In this section, different experiments are performed for each topology, and the results are presented. Experiments were repeated for each technique, individually and their combination to show the performances after combining the two techniques. The device using the electrical circuitry of topology 3 was also attached to a person s body to demonstrate its energy harvesting feasibility for some normal daily activities and also while exercising. They were asked to walk, jog and ride a bicycle, and the energy generated calculated. Optimum orientation and position are considered for the device for on-body placement. While walking it was attached just below the knee, it was attached to the wrist while jogging and while riding a bicycle it was attached to the ankle, as shown in Fig. 9. Voltage (V) voltage voltage power power Resitance (kω) Fig.. Maximum power obtained from individual harvesters. The output of the LTC3588- chip can be selected to be at four different levels. Experiments were performed using two levels, 3.3 and.8 V. These are common supply voltages for the most wireless and wearable sensor devices. The individual harvesters were tested individually before they were used in the three possible conditioning circuit topologies provided in this work. The power results are plotted in Fig. in terms of maximum achievable power for each harvester, when they were moved at fast speed. The LTC3588- is used with piezoelectric harvester and its regulation voltage is set at.8 V. Maximum voltages for the selected loads range from.8 to.8 V for the piezoelectric harvester and.95 to 3.4 V for the electro-magnetic harvester, respectively. Although the maximum power will be higher for lower voltages as the load becomes smaller, the voltage obtained will not be sufficient enough to be used with electronics of a sensor device. It is important for a wear-

6 R. Hamid, M.R. Yuce / Sensors and Actuators A 57 (7) Voltage (V) 3 Voltage (Volts) (a) 4 6 (b) Fig.. The output voltage from the energy harvesting device along with the individual methods, electromagnetic () and piezoelectric () approaches for Topology with the LTC3588- chip set at (a) 3.3V and (b).8v Resistor (kω) (a) Resistor (kω) (b) Fig.. Average power versus various load resistors for (a).8 V regulation for piezoelectric, (b) 3.3 V regulation for piezoelectric. able energy harvesting device to be tested with various restive loads, as a wearable sensor device may operate in different modes of current requirements, and thus leading several different loads. In our experiment, the loads were selected such that an output voltage higher than.8 V was obtained for most of the times. The maximum power is.5 mw for the harvester and 93 W for the piezoelectric harvester. For all the three topologies, due to the design of the device, both the harvesters would not start generating energy at exactly the same time as the electromagnetic harvester starts harvesting energy as soon as the magnet starts moving, whereas the piezoelectric harvester can only harvest when the magnet hits the piezoelectric transducers placed at either end of the device. With the output of LTC3588- set at 3.3 and.8 V in topology, the output voltages of the individual and coupled harvester were recorded without any load, and the results are shown in Fig. (a) and (b) respectively. In this experiment, the outputs from all the harvesters were recorded across a 6.8 F storage capacitor. Fig. (a) demonstrates that the coupled harvester follows the output from the electromagnetic harvester, and further explanation on this has been given in the next paragraph. From Fig. (b) it can be seen that the coupled harvester generates greater voltage than the individual harvesters, and this has been further proved when the power generated was measured against a range of different load resistances. When the two harvesters are integrated, there is some energy lost as integration requires the use of additional circuitry. Also, as explained previously, the two harvesters do not generate energy simultaneously, which also contributes to the power generated from the coupled harvester to be less than the sum of the power generated from the individual harvesters. The power generated were measured against a range of different load resistances, as shown in Fig. (a) and (b). When the output of the LTC3588- chip is set at.8 V for the piezoelectric harvester, the coupled (i.e. combined) energy harvester produces a power from to 65 W for the selected loads. The output voltage is not regulated and ranges from.3 V (for low resistance) to 3.5 V (for the highest value of resistance). When the output of the LTC3588- chip is set at 3.3 V for the piezoelectric harvester, the coupled harvester produces a power up to 3 W. The output voltage for this case was changing from to 3.6 V. For the 3.3 V case, the harvester and the coupled harvester generate similar power levels, and this can be explained by the fact that the harvester is being regulated, whereas the is not. With the proposed connection of diodes, only one of the voltages at any time higher than the other would forward bias the diode. The other diode will be reverse-biased, hence no current can flow from that particular harvester. Both the piezoelectric and electromagnetic energy harvesters were combined and energy is harvested using the circuit topology given in Fig. 7. The combined signal from both harvesters is regulated using the same LTC3588- chip (set to generate 3.3 and.8 V respectively). The output energy is stored on a 6.8 F capacitor without any load. From Fig. 3(a) and (b), it can be seen that the coupled harvester can hold the charge for a longer time than the individual harvesters, which will result in more energy harvesting, therefore higher average power. Some spikes in the output voltages appear due to the response of the regulator in the LTC3588- chip to regulate higher input voltages, mainly from the harvester. These spikes are ignored for power calculations. The power for all the experiments were calculated using Ohm s law, P = V R. Hence the power from topology is lower than that of topology, but it can be explained by the fact that it holds the charge longer than the former one, and the related voltage at this

7 4 R. Hamid, M.R. Yuce / Sensors and Actuators A 57 (7) Voltage (Volts) Voltage (Voltage) (a) (b) Fig. 3. The output voltage from the energy harvesting device along with the individual methods, electromagnetic (), and piezoelectric approaches for Topoogy with the output is regulated at 3.3 V (a) and.8 V (b), respectively Resistor (kω) (a) Resistor (kω) (b) Fig. 4. Average power measurements for Topology, regulated at and 3.3 V (a) and.8 V (b) respectively. time is also at a usable level. So the circuit topologies can be chosen according to the application. Average power values are calculated for various load conditions when the output of the LTC3588- chip is set at 3.3 V as well as at.8 V. Fig. 4 demonstrate measured average power levels obtained for.8 and 3.3 V regulated outputs, respectively. We show the individual harvesters and as well as the combined scenario. As can be seen, the output power is dominated by the harvester, especially for the higher regulation voltage of 3.3 V, which is expected. The output voltages were between.8.9 V, and thus regulated for Fig. 4. The output voltage for 3.3 V case was from.5 V (lowest load) to 3.3 V (highest load). At around k load, the output becomes 3.3 V (regulated), the desired regulated output. The output of the regulator was not generating sufficient current before this point to produce a regulated output voltage for 3.3 V. In the last configuration, two separate LTC3588- chips were utilized, one for each energy harvesting method. The chip is set to generate.8 V for each individual harvester since a voltage value more than 3.6 V will be very high for low-power electronics in wearable devices. The voltage across the harvester can be found by deducting the voltage across the harvester from the voltage across the coupled harvester since the two harvesters are connected in parallel. From Fig. 5 it can be seen that the voltage across the coupled device is double to that of the individual harvesters without any load, which is expected when both harvesters produce a regulated output. Fig. 6(a) demonstrates that the coupled device generates more power than the individual harvesters and Fig. 6(b) demonstrates that the voltage generated by the coupled Voltage (Volts) Fig. 5. The output voltage across the piezoelectric () and coupled energy harvester for Topoogy 3 (with no load), with the chip set to generate.8 V. harvester is approximately (since the additional circuitry needed to integrate the two harvesters introduces some energy loss) the sum of the voltages generated from the individual harvesters. Fig. 6(a) shows that the coupled harvester generates more power than the sum of the power generated from the individual harvesters, but Fig. 6(b) proves that the coupled harvester does not actually generate more than what is generated from the individual harvesters. Hence the coupled energy harvester is more efficient than using only one harvesting technique. Fig. 6 shows the average power measured against various loads. When the voltage becomes completely regulated as shown in Fig. 5, the average power stays between W with a stable DC voltage over 3 V.

8 R. Hamid, M.R. Yuce / Sensors and Actuators A 57 (7) Fig. 6. Average (a) power and (b) voltage for Topology 3, regulated at.8 V Table Energy generated while riding a bicycle at speed of (35 38) km/h. Time (s) Voltage (mv) Energy (mj) Resistor (kω) Walking 3-5 km per hr Riding a bike Walking 4-6 km per hr Fig. 7. Maximum power generated from the Topology 3 across various load resistors. Table Enegy generated while walking at two different speeds. Speed: 3 5 km/h Time (s) Voltage (mv) Energy (mj) Speed: 4 6 km/h Energy (μj) Fig. 8. Comparison of the energy generated from different activities over time using Topology 3. generated was calculated using Eqs. (3) and has been recorded in Tables and. E = CV (3) We also plotted the maximum power obtained from this topology, which is shown in Fig. 7. The maximum power obtained is.55 mw, and the maximum average power obtained is 5 W. The output voltage levels for selected loads range from.8 to 3.6 V. 3.6 V is obtained when both and harvesters produced a regulated output voltage of.8 V at higher loads. While exercising, the device was attached to the part of the body which moves the device most in back and forth motion, which will ensure that the magnet inside the device has hit the piezoelectric transducers placed at both ends, generating maximum energy from both the electromagnetic and piezoelectric energy harvesters. For example, it was attached under the knee while walking, on the wrist while jogging and ankle while riding a bicycle. A F capacitor was used as the storage capacitor and the voltage across this capacitor was measured after exercising at specific intervals. The energy It is also important to note that the low voltages obtained was due to the large storage capacitor used, since we wanted to store the charge for a longer time. It is easier to achieve higher voltages if a small capacitor is used, which was done in the previous experiments. Fig. 8 compares the energy generated from the different activities over time. From the two tables it can be concluded that the faster the device is shaken (higher vibration frequency), the quicker it will charge a capacitor/battery as expected. Table 3 compares studies very related to this work. The PV based wearable energy harvesting technique in [6] harvests about 68 W and the efficiency drops significantly when the PV is bended during motion of the human body. The wearable harvester work in [] uses a ball that hits and bounces from piezoelectric diagrams to obtain energy from low-frequency vibrations. It uses six diagraphs. The energy converter is tested during walking and running actives

9 6 R. Hamid, M.R. Yuce / Sensors and Actuators A 57 (7) 98 7 Table 3 The State-of-art comparison with wearable harvesting techniques. Harvesting Techniques Voltage levels Power This work: Vibration Combined -Piezoelectric Flexible solar panel [6].8 V, 3.3 V 5 3 W (Average).55 mw (Max) Dimension of device is 38.5 mm 34 mm 37 mm. V W (PV cell with dimension of 6 mm 7 mm) Thermoelectric [5]. V,.3 V W/cm (Maximum, indoor) mw/cm (at C, ambient) Dimension is 3 mm 3 mm 3 mm Vibration Piezoelectric-only ] Vibration- Piezoelectric only [9] Vibration- Magnetic only [] 3.3 V 7 W (from 6 piezo diagraphs, 9 cm 3 ) Dimension is 5 mm 5 mm 3 mm mv 5nW ( mm steel ball in box of side length. mm) 8 nw/cm 8 mv to 7 mv.5 mw/cm 3. (Calculated) Dimension of device not given Piezoelectric [7].9 mv W (Max) Dimension of device not given Electromagnetic generator [7] 98. mv W (Max) Dimension of device not given with an average power of 7 W. A similar study is investigated in [9] by placing a ball in a very small box, generating 8 nw/cm. The work in [] uses permanent magnet ball rolling arbitrarily in a spherical cavity wrapped with copper coil windings. This wearable harvester is tested on different locations of human body with different activities walking (4 km/h) and running (4.5 km/h). The voltage levels achieved by this harvester are not sufficient enough to be used with wearable electronics. In addition the power level provided is calculated rather than measuring a conditioning circuit. Another interesting wearable energy work that used two separate harvesters-piezoelectric and electromagnetic for generating energy from movements of the torso during breathing is presented in [7]. The maximum power levels obtained from this work is without any regulation. In our work we combined piezoelectric and electromagnetic in one unit. Piezoelectric is mainly utilized because of its high voltage output to generate several stable voltage values. Meanwhile the electromagnetic harvesting technique is used mainly due its high power generation ability. This combination facilitates a harvester producing several regulated output voltages with power levels generated as much as or higher than other existing techniques. We managed to bring our output voltages to.8 and 3.3 V to highlight that electronics of existing sensors operate with such voltage levels. All the relevant harvesters found in the literature mentioned in Table 3 have unregulated voltages. Their achieved voltages are included in the Table 3 to highlight this. The comparison with the existing harvesters shows that it is possible to achieve a higher output voltage with our device for a wearable harvester. It is important to note that the voltage values obtained from existing harvesters are not useful unless they are high enough and regulated. Measurement results from our work are comparable with thermoelectric and solar PV based techniques. However the motions and movements of the human body will affect the performance of PV cells and thermoelectric transducers significantly due to the bending effect of PV cells and loose contact of thermoelectric transducers, respectively. Our results and existing wearable electromagnetic harvesters indicate that it is possible to generate power from harvesters up to.5 mw. This work by combining electro-magnetic energy harvester with a piezoelectric harvester enables solid regulated voltages required for wearable sensors and sufficient output power levels that can operate with many wearable sensors. We tested the proposed harvester against various loads to study its suitability for various wearable devices while the existing harvesters provided results mainly for one optimum load. The proposed technique is also tested to power a temperature sensor, MCP97, with a memory to store values. In this the devices was consuming 45 W of power which was obtained when the frequency of vibration was about 3 Hz. More of studies similar to this are required in the future to enable self-powered wearable sensors. Combined energy harvesting methods will play an important role in the design of future wearable electronics as power levels of electronic devices in wearable devices are changing from mw to W power ranges. 5. Conclusion This paper presents an energy harvesting technique that harvests energy from low frequency vibrations like human motion to enable an alternative power supply in wearable devices. Electromagnetic and piezoelectric energy harvesting techniques were combined, and it was shown that the resulting energy harvested is significantly improved over individual energy harvesters. The proposed technique was tested to charge a capacitor when it was worn by a person while walking and exercising. The proposed harvester is able to generate 75.6 J of energy when the speed of riding a bicycle was around (35 38) km/h. It is shown that the combined system is capable of producing higher energy levels. In addition, this work demonstrates that combining both piezoelectric and electromagnetic based harvesting techniques provide stable higher regulated output voltages. Such techniques can be a part of electronics of future wearable devices and can potentially enable autonomous wearable sensing applications. From Eq. () it can also be deduced that the desired output depends on the number of turns or coil, the area at which the magnetic flux affects the coil, and the rate of change of magnetic flux density with respect to time. Hence a greater output can be achieved at a greater speed, by increasing the number of turns of coil or by using a magnet which has greater magnetic field flux density. Likewise, to generate the same output, a smaller device can be designed by using a stronger magnet, or increasing the number of turns of coils, and the number of piezoelectric transducers would also have to be changed accordingly. Acknowledgment Mehmet R. Yuce s work was supported by Australian Research Council Future Fellowships Grant FT343. References [] M.R. Yuce, Implementation of wireless body area networks for healthcare systems, Sens. Actuators A: Phys. 6 () 6 9, July. [] M.R. Yuce, J. Khan, Wireless Body Area Networks: Technology, in: Implementation and Applications, Pan Stanford Publishing, Singapore,, ISBN [3] M.H. Ghaed, et al., Circuits for a cubic-millimeter energy-autonomous wireless intraocular pressure monitor, IEEE Trans. Circuits Syst. I 6 () (3) 35 36, December. [4] N. Desai, J. Yoo, A.P. Chandrakasan, A Scalable,.9 mw, Mb/s e-textiles body area network transceiver with remotely-powered nodes and bi-directional data communication, IEEE J. Solid-State Circuits 49 (9) (4) 995 4, Sept. [5] V. Leonov, Thermoelectric energy harvesting of human body heat for wearable sensors, IEEE Sens. J. 3 (6) (3) 84 9.

10 R. Hamid, M.R. Yuce / Sensors and Actuators A 57 (7) [6] W.Y. Toh, Y.K. Tan, W.S. Koh, L. Siek, Autonomous wearable sensor nodes with flexible energy harvesting, IEEE Sens. J. 4 (7) (4) 99 36, July. [7] D. Zhu, Vibration energy harvesting: machinery vibration, human movement and flow induced vibration, in: Y.K. Tan (Ed.), Sustainable Energy Harvesting Technologies Past, Present and Future, In-Tech, Croatia,. [8] Y. Zhu, S.O.R. Moheimani, M.R. Yuce, A -DOF MS ultrasonic energy harvester, IEEE Sens. J. () 55 6, January. [9] E. Simon, Y. Hamate, S. Nagasawa, H. Kuwano, 3D vibration harvesting using free moving ball in T microbox, Proc. Power MS () [] D. Alghisi, S. Dalolaa, M. Ferrari, V. Ferrari, Triaxial ball-impact piezoelectric converter for autonomous sensors exploiting energy harvesting from vibrations and human motion, Sens. Actuators A: Phys. (5) , September. [] B.J. Bowers, D.P. Arnold, Spherical, rolling magnet generators for passive energy harvesting from human motion, J. Micromech. Microeng. 9 (9) (9), August. [] D.F. Berdya, D.J. Valentinoc, D. Peroulis, Kinetic energy harvesting from human walking and running using a magnetic levitation energy harvester, Sens. Actuators A: Phys. () (5) 6 7, February. [3] H. Xia, R. Chen, L. Ren, Analysis of piezoelectric electromagnetic hybrid vibration energy harvester under different electrical boundary conditions, Sens. Actuators A: Phys. 34 (5) 87 98, October. [4] M. Wahbah, M. Alhawari, B. Mohammad, H. Saleh, M. Ismail, Characterization of human body-based thermal and vibration energy harvesting for wearable devices, IEEE J. Emerg. Sel. Top. Circuits Syst. 4 (3) (4) [5] R. Hamid, A. Mohammadi, and M.R. Yuce, WE-Harvest: a wearable piezoelectric-electromagnetic energy harvester, Proceedings of the International Conference on Body Area Networks (BodyNets 5), September, 5. [6] P.D. Mitcheson, Energy harvesting for human wearable and implantable bio-sensors, Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society (BC),, pp [7] E. Shahhaidar, et al., Piezoelectric and electromagnetic respiratory effort energy harvesters, Proceedings of the 35th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (BC), 3 7 July 3, pp [8] P. Li, et al., An analysis of the coupling effect for a hybrid piezoelectric and electromagnetic energy harvester, Smart Mater. Struct. 3 (6) (4) 7. [9] T.T. Nguyen, T. Feng, P. Hafliger, S. Chakrabartty, Hybrid CMOS rectifier based on synergistic RF-piezoelectric energy scavenging, IEEE Trans. Circuits Syst. I 6 (4) , December. [] mm Block (Rare Earth) x.7mm-Block- %8Rare-Earth%9.html. [] S. Priya, D.J. Inman (Eds.), Energy Harvesting Technologies,, Springer, New York, 9, pp [] P. Nintanavongsa, et al., Design optimization and implementation for RF energy harvesting circuits, IEEE J. Emerg. Sel. Top. Circuits Syst. () 4 33, March. [3] LTC3588-:Piezoelectric Energy Harvesting Power Supply. Biographies Mehmet Rasit Yuce is an associate professor in the Department of Electrical and Computer Systems Engineering, Monash University, Australia. He received the M.S. degree in Electrical and Computer Engineering from the University of Florida, Gainesville, Florida in, and the Ph.D. degree in Electrical and Computer Engineering from North Carolina State University (NCSU), Raleigh, NC in December 4. He was a post-doctoral researcher in the Electrical Engineering Department at the University of California at Santa Cruz in 5. He was an academic member in the School of Electrical Engineering and Computer Science, University of Newcastle, New South Wales, Australia until Jul. In July, he joined Monash University where he is an Associate Professor and an Australian Research Council (ARC) Future Fellow. His research interests include wearable devices, wireless implantable telemetry, wireless body area network (WBAN), bio-sensors, integrated circuit technology dealing with digital, analog and radio frequency circuit designs for wireless, biomedical, and RF applications. Dr. Yuce has published more than 4 technical articles in the above areas and received a NASA group achievement award in 7 for developing an SOI transceiver. He received a best journal paper award in 4 from the IEEE Microwave Theory and Techniques Society (MTTS). He received a research excellence award in the Faculty of Engineering and Built Environment, University of Newcastle in. He is an author of the books: Wireless Body Area Networks published in and Ultra-Wideband and 6 GHz Communications for Biomedical Applications published in 3. He is a senior member of IEEE. He is a member of the technical committee on wearable biomedical sensors and systems for the IEEE Engineering in Medicine and Biology Society. He is one of Founding Technical Committee Co-Chairs of Interactive and Wearable Computing and Devices, IEEE Systems, Man, and Cybernetics Society. He is an editor for IEEE Sensors Journal and a guest editor for IEEE Journal of Biomedical and Health Informatics in 5. Rawnak Hamid is research assistant working in the Department of Electrical and Computer Systems Engineering, Monash University. She received bachelor degree from Monash University in 5. Her research work is in the area of wearable sensors for healthcare applications.