electronics

Article A Simple WiFi Harvester with a Switching-Based Power Management Scheme to Collect Energy from Ordinary Routers

Fernando Angulo 1,2, Loraine Navarro 1,* , Christian G. Quintero M. 1 and Mauricio Pardo 1

1 Department of Electrical and Electronics Engineering, Universidad del Norte, Barranquilla 081007, Colombia; [email protected] (F.A.); [email protected] (C.G.Q.M.); [email protected] (M.P.) 2 College of Engineering, Institución Universitaria ITSA, Soledad 083002, Colombia * Correspondence: [email protected]

Abstract: This paper shows the design process of a simplified harvesting circuit for WiFi at the 2.4 GHz frequency band based on the analysis of the environment available signals. Those signals and their power level define an antenna design to maximize captured energy and select the proper number of stages for a voltage multiplier so that an impedance matching network is no longer required. With this, it is possible to maintain the harvester architecture simple without sacrificing performance. The use of supercapacitors is preferred over batteries due to their high-power capacity, the ability to deliver high peak currents, long-life cycle size, and low cost. Hence, supercapacitor availability allows to devise a novel switching scheme that employs two units that favor energy use and speed up the recharging process. The built harvester exhibits a power conversion efficiency greater than 50% under an incident signal of 0 dBm in the rectenna. The tests are carried out in



an academic environment using a multi SSID router, collecting 494 mJ without requiring special
 modifications in the router used as an energy source.

Citation: Angulo, F.; Navarro, L.; Quintero M., C.G.; Pardo, M. A Keywords: energy harvesting; WiFi; rectenna; network traffic; multi SSID; power management Simple WiFi Harvester with a switching scheme Switching-Based Power Management Scheme to Collect Energy from Ordinary Routers. Electronics 2021, 10, 1191. https://doi.org/10.3390/ 1. Introduction electronics10101191 In the last two decades, there has been an increase in research aimed at different techniques for harvesting energy from the environment, and in particular, those associated Academic Editor: Byunghun Lee with radio frequency (RF) signals known as RF energy harvesting (RF-EH), wireless energy harvesting (WEH) or RF energy harvesting wireless communication (RF-EHWC) [1–3]. One Received: 20 April 2021 of the main objectives of such techniques is to use available energy from the environment Accepted: 7 May 2021 Published: 17 May 2021 to be used to support the node power schemes for wireless sensor networks (WSN) and Internet of things (IoT) systems [3,4]. Thus, it is desired to keep the devices running without

Publisher’s Note: MDPI stays neutral the use of batteries. The IoT schemes have seen significant research intending to make them with regard to jurisdictional claims in more energy-efficient, secure, and affordable. Thus, the concept of green information and published maps and institutional affil- communication technology focuses on green RF identification (RFID) technology, green iations. sensor networks technology, green cellular networks, etc. IoT poses a high potential to bolster environmental sustainability and economic with the advancements of enabling technologies [5]. Dedicated sensors can be deployed in selected areas for data gathering and measuring global and local desired variables. A sensor node contains several small, low-cost, and Copyright: © 2021 by the authors. Licensee MDPI, Basel, Switzerland. low-power electronic devices such as a sensing unit, a storage unit, a processing unit, This article is an open access article and a communication unit [5]. However, the footprint and power autonomy for such distributed under the terms and sensors usually are scarce. However, the sensors still need modules that allow them conditions of the Creative Commons the sharing of information and the requirement to protect the information that they are Attribution (CC BY) license (https:// sharing, for example, in a wireless sensor network (WSN) [5,6]. A tendency is to move the creativecommons.org/licenses/by/ complexity and power demand of dedicated communication units to central entities such 4.0/). base stations or even employ mobile agents to visit the IoT sensors using UAVs to assuage

Electronics 2021, 10, 1191. https://doi.org/10.3390/electronics10101191 https://www.mdpi.com/journal/electronics Electronics 2021, 10, 1191 2 of 21

the requirements for power operations [5]. However, the nodes need to have embedded structures to protect the data since electronic attacks are unavoidable. Device-to-device (D2D) protocols can be adopted in the sensor due to the reduced resources required for its implementation [6]. Thus, the extraction and efficient use of energy to run all the processes involved in an IoT becomes very important. In this regard, some of the works have focused on using energy from RF signals such as radio, television, cellular phones, and WiFi due to the massification of these technologies [7,8]. Among the investigations carried out, there are examples of systems that even make simultaneous use of several frequency bands [9–12] to take advantage of a greater spectrum which is reflected in greater energy collected. Similarly, there are works aimed at energy harvesting in indoor and outdoor envi- ronments [7,13–15] where outdoor solutions are generally applied to permanent carrier signals, constant power, or transmission systems of the metropolitan area. In contrast, for indoors, short-range systems and dynamic carrier power are used. This parameter presents the most significant challenges from the design side. Some of the challenges presented in this scenario are low energy levels in the environment, power dependency on the use of the network, and non-linearity in the reception module of a harvester. It is necessary to consider the different parameters of the target signal to develop an energy harvesting system with RF signals, such as modulation, frequency, and power trans- mission. In particular, for WiFi, signal emission regulations limit the power transmission from a common access point (AP) or router to 20 dBm maximum, requiring the use of multiple schemes to take advantage of these reduced levels present in the environment [16]. Besides, evaluating the WiFi signal transmission form establishes strategies that allow better use of the available energy. Thus, Reference [17] presents the design, development, and evaluation of a rectifier circuit prototype to apply the energy collection of 2.45 GHz signals, obtaining an RF/DC conversion efficiency of 68% at −10 dBm in simulation and 33% to −5 dBm of the built prototype. Reference [18] presents the design of an RF energy harvesting system based on three harvesters using three-panel antennas to cover three channels of the 2.4 GHz WiFi band. The output of the three blocks is connected to a voltage multiplier circuit. It uses a supercapacitor as a storage medium, collecting 0.45 V in 6 to 7 h, and then going to a second conversion stage increases the output to 2 V. The work concludes that the proposed system can supply power to electronic devices with a power consumption of less than 1 mW. In [13], the design of a WiFi energy collector for indoor use is described, which operates at low incident power and conversion efficiency of 61%, managing to collect 1.67 mJ of energy in an operation of 38 h on a 1 mF supercapacitor with a voltage of 1.8 V. In [19], a collector is developed to feed a temperature sensor node and a camera at 5 m and 6 m, respectively; but requiring the modification of the router so that it transmits between the silence spaces that occur due to the absence of network traffic, turning the router into a dedicated power generator. From state of the art, it could be stated that WiFi energy harvesting systems are mainly focused on harvester development. From the authors knowledge, the works do not concentrate on the effects of this technology transmission method, which influences the collector performance and the amount of energy collected. Thus, the purpose of this work is to consider both sides of the problem. On the one hand, the nature of the WiFi signal is considered to explore approaches to better utilize the energy available without requiring modifications on routers not considered as standard configurations and operation of this equipment. Hence, in particular, the use of routers or APs with multiple SSID is explored to increase the amount of energy harvested in low traffic conditions given the nature of the WiFi technology. On the other hand, the harvesting circuit is addressed to devise a system as simple as possible while maintaining adequate performance. Thus, the so-called rectenna design and the power management system are developed, emphasizing a simplistic architecture featuring low cost and low power consumption. Therefore, at the end, a working prototype is developed to work in an actual Electronics 2021, 10, x FOR PEER REVIEW 3 of 21

Electronics 2021, 10, 1191 3 of 21 system are developed, emphasizing a simplistic architecture featuring low cost and low power consumption. Therefore, at the end, a working prototype is developed to work in environmentan actual environme (houses andnt (houses buildings) and and buildings) showing theand potential showing to the deliver potential enough to energy deliver toenough operate energy electronic to operate devices electronic such as those devices employed such as forthose IoT-type employed systems for IoT even-type in cases systems of low-trafficeven in cases condition of low- ontraffic a WiFi condition network. on a WiFi network.

2.2. WiFiWiFi Transmission Transmission Signal Signal Behavior Behavior ItIt isis important important toto check check the the WiFi WiFi transmission transmission signal signal behavior behavior toto design design an an energy energy harvesterharvester aimed aimed properly. properly. Thus, Thus, for for a a typical typical AP AP oror router,router, the the WiFi WiFi standard standard periodically periodically sendssends a a signal signal called called beacon beacon [ 20[20],], a a packet packet containing containing router router data data information information that that allows allows thethe users users to to connect connect to to it. it. This This pilot pilot signal signal can can be configured be configured to be to sent be from sent thefrom router the everyrouter 40every to 1000 40 to ms, 100 a0 time ms, knowna time known as the beacon as the beacon interval interval (BI). By (BI). default, By default, the BI is the configured BI is config- to 100ured ms. to If10 there0 ms. are If there no users are connectedno users connected to the router, to the the router, beacon the signal beacon becomes signal thebecomes only signalthe only emitted signal by emitted the equipment. by the equipment. Figure1 shows Figure a1 WiFishows signal a WiFi in signal the time in the domain time wheredomain thewhere only the observed only observed activity correspondsactivity corresponds to the beacon to the signal beacon since signal no userssince areno users connected are con- to thenected router. to the As router. observed, As theobserved, BI is set the to BI 40 is ms, set and to 4 with0 ms, noand traffic, with theyno traffic, are visible they are “periods visible of“periods silence” of in silence” the signal. in Hence, the signal. in terms Hence, of available in terms energy of available to harvest, energy the to amount harvest, of the it recoveredamount of from it recovered the environment from the willenvironment be relatively will low be relatively [19]. low [19].

FigureFigure 1. 1.WiFi WiFi signal signal without without traffic traffic with with BI BI 40 40 ms. ms.

OnOn thethe contrary,contrary, FigureFigure2 2shows shows the the WiFi WiFi signal signal with with the the traffic traffic demand demand by by users users connected.connected. InIn thisthis case,case, thethe beaconbeacon signalsignal isis identified identified by by the the arrows, arrows, where where the the BI BI is is kept kept equalequal to to 40 40 ms. ms.If If thethe BIBI isis takentaken asas thethe operationoperation window window for for harvesting, harvesting, it it is is clear clear that that the the WiFiWiFi signal signal can can be be viewed viewed as as a a PWM-like PWM-like signal; signal; therefore, therefore, it it can can be be concluded concluded that that more more energyenergy can can be be harvested harvested with with a a significant significant amount amount of of network network traffic. traffic.

Electronics 2021, 10, 1191 4 of 21 Electronics 2021, 10, x FOR PEER REVIEW 4 of 21

FigureFigure 2. 2.WiFi WiFi signal signal with with traffic. traffic.

AnotherAnother important important parameterparameter in the operation of of WiFi WiFi technology technology is is that that it itworks works by bychannels. channels. There There are are14 channels 14 channels for the for 2. the4 G 2.4Hz GHzband, band, spaced spaced 5 MHz 5 apart, MHz apart,which whichdistrib- distributeute them from them 2.0 from4 GHz 2.04 to GHz 2.72 G toHz. 2.72 According GHz. According to the geographical to the geographical area, 11 or area,14 channels 11 or 14are channels used. Thus, are used.the configuration Thus, the configuration of 11 channels of is 11used channels in America, is used from in America,which Channels from which1, 6, and Channels 11 are 1,primarily 6, and 11 used. are primarily Usually, used.working Usually, with workingadjacent withchannels adjacent is not channels preferred is notbecause preferred each because channel each occupies channel a occupiesBW of 22 a M BWHz; of therefore, 22 MHz; therefore, overlap generates overlap generates between betweenthem. them. ThereThere isis aa metricmetric known as RSSI RSSI ( (receivedreceived s signalignal strength strength indicator) indicator) in in terms terms of ofre- receivedceived power. power. An An RSSI RSSI is is defined defined as excellentexcellent toto goodgood forfor datadata connectivityconnectivity if if its its value value is is betweenbetween− −3300 to −−6060 d dBm;Bm; m medium,edium, between between −6−060 to to −75− 75dBm; dBm; low low or v orery very low, low, between between −75 −to75 −85 to −dBm;85 dBm; and andno c noonnection connection if the if theRSSI RSSI is below is below −85− d85Bm. dBm. However, However, even even excellent excellent and andgood good RSSIs RSSIs may may not not be besufficient sufficient from from an anenergy energy-harvesting-harvesting perspective perspective since since a asignal signal of of−40−40 dBm dBm is isequivalent equivalent to toonly only 10 1000 nW, nW, which which cannot cannot be be efficiently efficiently harvested. harvested. The The latter latter is issignific significantlyantly more more critical critical for for WiFi WiFi technology, technology, where where the the radiated radiated power power is isvariable variable ac- accordingcording to to the the transmitted transmitted packets packets and and not not a constant a constant or orstable stable radiation radiation source source (See (See Fig- Figuresures 1 1and and 2).2). AsAs shown shown in in Figure Figure3 ,3, the the WiFi WiFi spectrum spectrum is is packed packed for for urban urban environments, environments and, and it it cancan be be thought thought that that the the several several WiFi WiFi signals signals would would add,add, allowingallowing harvestingharvesting moremore energy.energy. However,However, this this would would be be only only possible possible if if such such signals signals had had considerable considerable power power levels levels and and have the same frequency (i.e., constructive interference). It is important to note that levels have the same frequency (i.e., constructive interference). It is important to note that levels below −40 dBm or −60 dBm only represent 100 nW to 1 nW. Thus, even if the signals add below −40 dBm or −60 dBm only represent 100 nW to 1 nW. Thus, even if the signals add to each other, it would still be very little energy, a scenario that is illustrated in Figure4 to each other, it would still be very little energy, a scenario that is illustrated in Figure 4 with APs 1, 2, and 3. This is the typical scenario in WiFi energy harvesting since it is with APs 1, 2, and 3. This is the typical scenario in WiFi energy harvesting since it is un- unlikely to find APs installed at a single point or very close, making it difficult to take likely to find APs installed at a single point or very close, making it difficult to take ad- advantage of energy from all the equipment. Likewise, it must be considered that the vantage of energy from all the equipment. Likewise, it must be considered that the anten- antennas with higher gain (suitable for energy harvesting) are directive; therefore, not all nas with higher gain (suitable for energy harvesting) are directive; therefore, not all the the signals will fall within their aperture unless there is an arrangement or configuration of signals will fall within their aperture unless there is an arrangement or configuration of several antennas for this purpose. several antennas for this purpose.

Electronics 2021, 10, 1191 5 of 21 ElectronicsElectronics 2021 2021, 10, 10, x, FORx FOR PEER PEER REVIEW REVIEW 5 5of of 21 21

Figure 3. Presence of WiFi signal in urban environment. FigureFigure 3.3.Presence Presence ofof WiFiWiFi signal signal in in urban urban environment. environment.

FigureFigureFigure 4. 4.4. EHEH EH WiFi WiFiWiFi Scenario ScenarioScenario with withwith Single SingleSingle SSID SSID SSID and and and Multiple. Multiple. Multiple. SSID SSIDSSID APs. APs.APs. The above provides a significant opportunity to target routers with multiple SSID TheThe above above provides provides a asignificant significant opportunity opportunity to to target target routers routers with with multiple multiple SSID SSID characteristics that can simultaneously broadcast several wireless networks, perceiving characteristicscharacteristics that that can can simultaneously simultaneously broadcast broadcast several several wireless wireless networks, networks, perceiving perceiving themselves as 2, 3, or 4 carriers or different RF signals, as shown in Figure4 for the AP4. themselvesthemselves as as 2, 2, 3, 3, or or 4 4carrie carriersrs or or different different RF RF signals, signals, as as shown shown in in Figure Figure 4 4for for the the AP4. AP4. Thus, multiple SSID routers can be used to achieve more energy because signals effectively Thus,Thus, multiple multiple SSID SSID routers routers can can be be used used to to achieve achieve more more energy energy because because signals signals effec- effec- add up, and it is possible to maintain a level of energy harvesting even when there is no tivelytively add add up, up, and and it itis is possible possible to to maintain maintain a alevel level of of energy energy harvesting harvesting even even when when there there traffic on the network. Thus, if the operation window for harvesting is kept as the BI for a isis no no traffic traffic on on th the enetwork. network. Thus, Thus, if ifthe the operation operation window window for for harvesting harvesting is is kept kept as as the the BI BI single SSID router (i.e., the operation window is kept the same for comparison); then, the BI forfor a asingle single SSID SSID router router (i.e., (i.e., the the operation operation window window is is kept kept the the same same for for comparison); comparison); then, then, for the other SSIDs becomes energy available for harvesting. Figure5 shows the harvester thethe BI BI for for the the other other SSIDs SSIDs becomes becomes energy energy available available for for harvesting. harvesting. Figure Figure 5 5shows shows the the output signal exposed to WiFi signals with 1 and 2, 3, and 4 SSIDs, highlighting that the haharvesterrvester output output signal signal exposed exposed to to WiFi WiFi signals signals with with 1 1and and 2, 2, 3, 3, and and 4 4SSIDs, SSIDs, highlighting highlighting

Electronics 2021, 10, 1191 6 of 21 Electronics 2021, 10, x FOR PEER REVIEW 6 of 21 Electronics 2021, 10, x FOR PEER REVIEW 6 of 21

beaconthat the signal beacon for signal 4 SSIDs for is 4 equivalentSSIDs is equivalent to having to a constanthaving a trafficconstant rate traffic a single rate SSID a single AP. ThisSSIDthat factthe AP. beacon can This be fact evidencedsignal can forbe evidenced4 in SSIDs measurements is inequivalent measurements using to having a spectrum using a constanta spectrum analyzer, traffic analyz as observedrateer, a assingle ob- in FigureservedSSID AP.6 ,in where ThisFigure fact a denser6, canwhere be frequency evidenceda denser spectrumfrequency in measurements can spectrum be observed using can bea atspectrum observed the 2.4 GHzanalyz at the bander, 2. 4as G for ob-Hz multibandserved SSIDsfor in mFigureulti routers. SSIDs 6, where The routers. signal a denser The originating signal frequency originat from spectrum aing router from can with a berouter multiple observed with SSIDs multipleat the indicates 2. 4SSIDs GHz that more power can be obtained for the same bandwidth. indicatesband for mthatulti more SSIDs power routers. can Thebe obtained signal originat for the ingsame from bandwidth. a router with multiple SSIDs indicates that more power can be obtained for the same bandwidth.

Figure 5. Exposed harvester output WiFi signal with 1 and 2 SSID without traffic. Figure 5. Exposed harvester output WiFi signal with 1 and 2 SSID without traffic. Figure 5. Exposed harvester output WiFi signal with 1 and 2 SSID without traffic.

1 SS ID 4 SS ID 1 SS ID 4 SS ID

Figure 6. Spectrum analyzer WiFi signal measurement for 1 and 4 SSIDs routers. FigureFigure 6. 6.Spectrum Spectrum analyzeranalyzer WiFiWiFi signal signal measurement measurement for for 1 1 and and 4 4 SSIDs SSIDs routers. routers. 3. Typical RF-EH Harvester System 3.3. TypicalTypical RF-EH RF-EH Harvester Harvester System System Figure 7 shows the essential components that comprise an RF energy harvesting sys- tem:Figure Figureantenna,7 7shows shows rectifier, thethe and essential power components components management that that circuit. comprise comprise an an RF RF energy energy harvesting harvesting sys- system:tem: antenna, antenna, rectifier, rectifier, and and power power management management circuit. circuit.

Electronics 2021, 10, 1191 7 of 21 Electronics 2021, 10, x FOR PEER REVIEW 7 of 21

A nte nna

Im pedance Pow er Volta ge Matc hing Managem ent Multiplier N etw ork U nit

A C /DC

R ecte nna Load FigureFigure 7. RFRF h harvesterarvester b blocklock d diagram.iagram.

TheThe antenna antenna is is the the component component that that allows allows the the capture capture of of RF RF energy energy from from the the environ- environ- ment,ment, which makesmakes itit thethe interfaceinterface element element between between the the harvester harvester and and the the environment environment on onwhich which the the RF RF energy energy is captured.is captured. For For RF-EH RF-EH systems, systems, the the predominant predominant type type of of antenna antenna is issector sector type. type. The The most most extensive extensive use use of these of these antennas antennas is for is mobile for mobile communications, communications, such suchas WiFi, as WiFi, and theyand arethey used are used for limited-range for limited-range distances. distances. The antennaThe antenna is designed is designed taking tak- as ingreference as reference parameters parameters such as suchgain, asfrequency, gain, frequency, and bandwidth and bandwidth that define that the defineamount the of amountpotential of energy potential that energy could bethat harvested. could be harvested. AA block block called called impedance impedance matching matching network network follows, follows, whose whose function function is matching is matching the antennathe antenna and andrectifier/multiplier rectifier/multiplier impedances impedances looking looking for reducing for reducing the losses the due losses to duereflec- to tionreflection (i.e., satisfy (i.e., satisfy the maximum the maximum transfer transfer power power theorem). theorem). Next,Next, a rectifierrectifier isis employedemployed to to single-polarize single-polarize the the perceived perceived signal signal for for DC DC conversion. conver- sion.Since Since the obtained the obtained DC componentDC component is low, is low, the the rectifier rectifier is complemented is complemented with with a voltage a volt- agemultiplier multiplier (charge (charge pump). pump). The The multiplier multiplier is is described described based based onon sensitivity,sensitivity, numbernumber of stages,stages, output output voltage, voltage, and and conversion conversion efficiency. efficiency. There There are areseveral several configurations configurations of rec- of tifierrectifier-multipliers.-multipliers. Among Among the most the most widely widely used used are the are Villard, the Villard, Greinacher, Greinacher, Dickson, Dickson, and Cockroftand Cockroft-Walton-Walton configurations configurations [21–23]. [21– 23]. TheThe antenna and rectifier/multiplier rectifier/multiplier set set are are usually usually referred referred to to as as rectennarectenna within the RFRF-EH-EH literature. literature. Proper Proper design design and and interoperation interoperation of ofthe the rectenna rectenna blocks blocks are are essential essential to ensureto ensure high high energy energy conversion conversion efficiency efficiency (i.e., (i.e.,input input and output and output powers powers as close as as close possi- as ble).possible). Finally,Finally, aa powerpower managementmanagement circuit circuit (PMU) (PMU) guarantees guarantees suitable suitable power power conditions conditions for the load given the energy harvested. For example, if the load requires a regulated or specific for the load given the energy harvested. For example, if the load requires a regulated or voltage or the load requires specific running times. According to the design constraints, a specific voltage or the load requires specific running times. According to the design con- commercial or a tailored power management system can be employed. Finally, in terms of straints, a commercial or a tailored power management system can be employed. Finally, the energy storage units, supercapacitors are preferred in EH WiFi systems due to their in terms of the energy storage units, supercapacitors are preferred in EH WiFi systems power density characteristics associated with high loading and unloading speed, long-life due to their power density characteristics associated with high loading and unloading cycle, size, and low cost compared to batteries [24,25]. speed, long-life cycle, size, and low cost compared to batteries [24,25]. 3.1. Antenna Design Considerations 3.1. Antenna Design Considerations A basic antenna used in RF-EH systems based on WiFi is the patch type mainly becauseA basic of its antenna ease of PCBused construction in RF-EH systems and its based adaptability on WiFi to is different the patch applications type mainly [26 ,be-27]. causeOther of types its ease of usable of PCB antennas construction are dipole, and its periodic adaptability log, andto different strip-loop applications [26,27]. An [26 array,27]. Otherof patch types antennas of usable can antennas be used to are improve dipole, the periodic harvester log, antennaand strip performance.-loop [26,27]. TheAn patcharray ofantenna patch isantenn consideredas can forbe used the present to improve work the due harvester to its simplicity antenna of performance. design and manufacture. The patch antennaFor is the considered construction for ofthe patch present antennas, work due Equations to its simplicity (1)–(4) are of used design to determine and manufac- their ture.dimensions according to the required parameters [28]. For the construction of patch antennas, Equations (1–4) are used to determine their r dimensions according to the required parametersλ 2 [28]. W = (1) 2 εr + 1 휆 2 푊 = √  − 1 (1) ε + 1 ε2− 1휀푟 + 1 h 2 ε = r + r . 1 + 12 (2) r,e f ec 2 2 W

Electronics 2021, 10, 1191 8 of 21

vo Le f ec = √ (3) 2 fr. εr,e f ec 1 L = √ √ − 2∆L (4) 2 fr εre f f µ0e0 Equally important are the return loss parameters (from the S-parameters), VSWR, and impedance adaptation. On the other hand, it is worth highlighting the effect of the incident power value within the performance of the system, which is defined for a transmission system according to Equation (5):

GtGrλ2 Pr = Pt , (5) (4πd)2L

where Pr is the power that reaches the receiver, Pt is the power of the transmission system, L is the losses associated with propagation in free space, Gt is the gain of the transmitting antenna, Gr is the gain of the receiving antenna (in this case the rectenna), λ is the wave- length at the operating frequency, and d is the distance between the transmitter and the harvester. Materials such as Rogers RO4350, RO3206, RO6010, and FR-4 [29] are typically used to manufacture the antenna, the latter preferred due to its low cost and acceptable performance in the 2.4 GHz band [28].

3.2. Voltage Multiplier Approach Voltage multipliers are electronic circuits that allow to rectify an AC signal and convert it into a DC signal while increasing its voltage by a factor related to the number of stages of its construction, based on the combination of rectifier diodes and capacitors. The multiplier voltage output is given by Equation (6), where Vi corresponds to the input voltage to the multiplier and n to the number of stages that comprises it,

Vo = 2nVi. (6)

In RF-EH WiFi, the diodes must generate the lowest possible voltage drop to reduce losses. Thus, metal-semiconductor type (i.e., Schottky) diodes are preferred. This diode type is also widely used in RF-EH systems due to its working frequency of up to 24 GHz. Equation (7) is used to evaluate the antenna and voltage multiplier performance, which corresponds to the power conversion efficiency and is given by the relationship between the incident power on the antenna and the power that can deliver at its output,

PRL ηPCE = . (7) PRxHV Along with the efficiency, the sensitivity is defined as the minimum incident power required for system operation, which is also adjusted to the load requirements [30].

3.3. Power Management Circuit An adequate PMU becomes crucial for proper sensor operation. As the total equivalent load changes, the supply system must provide a sufficient amount of power. DC-DC converters are circuits that convert a DC voltage to a different DC regulated level with high efficiency (>80%). According to the application, the DC-DC converters can increase or reduce the output voltage compared to the input voltage. Some are also designed to find the maximum power point (MPP) [31]. Given the WiFi signal nature, it is expected to require high charging/discharging rates, which imposes speed requirements for a storage unit. Batteries can be an option; however, the load probably could be inoperative for an extended period before use it. In this case, the speed requirement for the storage unit can be satisfied with supercapacitors (SCs), if Electronics 2021, 10, x FOR PEER REVIEW 9 of 21 Electronics 2021, 10, 1191 9 of 21

this case, the speed requirement for the storage unit can be satisfied with supercapacitors (SCs),the loads if the to loads be fed to withbe fed the with harvested the harvested energy energy are selected are selected to be ultra-lowto be ultra power.-low power. Thus, Thus,defining defining the energy the energy stored stored by a supercapacitor by a supercapacitor as as 11 W푊== CV퐶푉22,, ( (8)8) 22 andand considering considering a aload load working working a a3. 3.33 V V and and a a SC SC equal equal to to 22 2200 m mF,F, it it is is possible possible to to feed feed a a loadload requiring requiring 300 300 μµWW constantly constantly for for an an hour. hour. Thus, Thus, it it is is perfectly perfectly feasible feasible to to feed feed an an IoT IoT temperaturetemperature sensor sensor with with the the SC SC stored stored energy. energy. It It is is important important to to determine determine if if that that amount amount ofof energy energy can bebe obtainedobtained from from a a WiFi WiFi signal signal radiated radiated from from an ordinaryan ordinary router router (multi (multi SSID, SSID,as commented). as commented).

4.4. System System Description Description 4.1.4.1. Antenna Antenna Design Design Concept Concept GivenGiven the the low low power power received received from from an an ordinary ordinary router router and and the the ease ease of of building building patch patch antennas,antennas, it it is is decided decided to to work work with with a a 2 2 ×× 1 1array array to to keep keep the the footprint footprint as as small small as as possible. possible. TheThe dimensions dimensions and and geometry geometry of of the the array array are are shown shown in in Figure Figure 8.8 The. The total total are areaa for for the the 2 2 ×2 1× array1 array equals equals 88.8 88.8 cm2 cm. .

13.5 cm

4.29 cm 1.5 cm

0.23 cm 9.23 cm

1.5 cm

m

c

8

5 .

6 0.776 cm

0.67 cm

m c

0.4 cm

8

. 2

3.7 cm

0.58 cm

Figure 8. 2 × 1 array dimensions. Figure 8. 2 × 1 array dimensions. The design of the antenna array is carried out using the CST software. This tool containsThe design a module of the that antenna allows array adjusting is carried antenna out dimensionsusing the CST until software. it centers This the tool desired con- tainsfrequency a module of operation. that allows The adjusting design antenna process starteddimensions from until a rectangular it centers patchthe desired obtaining fre- quencya gain of 3.7operation. dB; then, The configuring design process an array, started a totalfrom of a rectangular 6.9 dB is reached, patch obtaining allowing a higher gain ofenergy 3.7 dB; harvesting then, configuring than the an single array, patch a total (see of Figure 6.9 dB9). is Even reached, though allowing more patch higher antennas energy harvestingcan be added than to the configure single patch a bigger (see Figure array, extra9). Even patches though increase more patch the antenna antennas footprint, can be addedwhich to cannot configure be attractive a bigger for array, IoT modules. extra patches As observed, increase the the patches antenna are linedfootprint, up vertically which cannotso that be it agreesattractive with for signal IoT modules. transmission As observed, polarization. the patches are lined up vertically so that itThe agrees simulated with signal S11 parameter transmission shows polarization.−43.19 dB, which guarantees the highest incident RF signal capture. It can be observed a 78% coverage of the 2.4 GHz WiFi band defining a −10 dB bandwidth, which allows having optimal behavior in 2.437 MHz equivalent to Channel 6. Figure 10 summarizes the simulated performance of the 2 × 1 array.

Electronics 2021, 10, x FOR PEER REVIEW 10 of 21

Electronics 2021, 10, 1191 10 of 21 Electronics 2021, 10, x FOR PEER REVIEW 10 of 21

Figure 9. Radiation diagram and simulated 2 × 1 array gain.

The simulated S11 parameter shows −43.19 dB, which guarantees the highest incident RF signal capture. It can be observed a 78% coverage of the 2.4 GHz WiFi band defining a −10 dB bandwidth, which allows having optimal behavior in 2.437 MHz equivalent to FigureFigure 9. 9.Radiation Radiation diagram diagram and and simulated simulated 2 2× × 11 arrayarray gain.gain. Channel 6. Figure 10 summarizes the simulated performance of the 2 × 1 array.

The simulated S11 parameter shows −43.19 dB, which guarantees the highest incident RF signal capture. It can be observed a 78% coverage of the 2.4 GHz WiFi band defining a −10 dB bandwidth, which allows having optimal behavior in 2.437 MHz equivalent to Channel 6. Figure 10 summarizes the simulated performance of the 2 × 1 array.

Figure 10. Parameters S simulated 2 × 1 array. Figure 10. Parameters S1111 simulated 2 × 1 array. 4.2. Multiplier Design Concept 4.2. Multiplier Design Concept From the literature, it is possible to design the rectenna without an impedance coupling networkFrom between the literature, antenna it is and possible multiplier. to design the rectenna without an impedance cou- plingFigure Whennetwork 10. Parameters there between are non-linearS11 antenna simulated circuitsand 2 × multiplier. 1 array. likethe rectifier, the impedance matching becomes dynamicWhen since there the are impedance non-linear ofcircuits these circuitlike the changes rectifier, with the impedance the input power matching and frequency.becomes dynamicTherefore,4.2. Multiplier since the case theDesign impedancecan Concept be further of complicated these circuit if changes the load with circuit the is also input non-linear power and since fre- its quency.impedanceFrom Therefore, wouldthe literature, the change case it withcan is possiblebe the further same to complicated parameters.design the rectenna Inif the general load without circuit terms, anis the alsoimpedance load non coupling-linear cou- sinceprocesspling its network isimpedance influenced between would by component-value antenna change and with multiplier. the tolerances, same parameters. errors associated In general with terms, manufacturing the load couplingprocesses,When process among there isare others influenced non [-32linear]. bycircuits component like the- valuerectifier, tolerances, the impedance errors matching associated becomes with mdynamicanufacturingA determining since processes, the factor impedance among in the voltage ofothers these multiplier[32]. circuit changes corresponds with to the the inputnumber power of stages and that fre- makequency.A itdetermining up. Th Theerefore, output factor the voltagecase in canthe and voltagebe further input multiplier sensitivity complicated corresponds are if defined the load to as circuit the devicenumber is also can ofnon harveststages-linear thatenergysince make its from impedanceit up. the The WiFi outputwould signal. voltagechange Therefore, withand performance inputthe same sensitivity parameters. evaluation are defined In is general required as the terms, according device the can load to harvestthecoupling frequency energy process offrom use, is the incidentinfluenced WiFi signal. power, by componentThere and load.fore, - Basedperformancevalue on tolerances, the analysisevaluation errors developed is associated required in [ac- with23], cordingfrommanufacturing three to the multiplier frequency processes, topologies of amonguse, (Dickson,incident others power,[32]. Cockroft-Walton, and load. andBased Greinachert), on the analysis the Cockroft- devel- opedWalton inA typedetermining [23], multiplier from three factor is chosen.multiplier in the voltage topologies multiplier (Dickson, corresponds Cockroft to- Walton,the number and of Grein- stages achert),that Experimentalmake the Cockroft it up. The tests-Walton output are carried type voltage multiplier out withand voltageinput is chosen. sensitivity multipliers are of 4,defined 5, 6, 7, andas the 8 stages, device and can theharvest performanceExperimental energy from of tests each the are one WiFicarried is evaluatedsignal. out withThere to voltage definefore, performance themultipliers number of evaluation stages.4, 5, 6, 7, As andis commented, required 8 stages, ac- andSchottkycording the performance diodesto the frequency are selected of each of tooneuse, reduce isincident evaluated voltage power, losses.to define and load. the numberBased on of the stages. analysis As com-devel- mented,opedAccording in Schottky [23], from to diodes the three simulations, are multiplier selected the to topologieshigher reduce the voltage number (Dickson, losses. of stages, Cockroft the-Walton, higher the and voltage Grein- output;achert),According however, the Cockroft to the in practice, simulations,-Walton from type the amultiplier determined higher the is chosen. number ofof stages,stages, nothe improvement higher the voltage in the output;outputExperimental however, will be obtained in practice,tests due are to fromcarried the littlea determined out amount with voltage of number available multipliers of energystages, of atno the4, improvement 5, input 6, 7, affectingand 8 instages, the the conversionand the performance efficiency. Therefore,of each one the is number evaluated of stages to define with the the number best possible of stages. performance As com- mustmented, be determined Schottky diodes via testing. are selected to reduce voltage losses. According to the simulations, the higher the number of stages, the higher the voltage output; however, in practice, from a determined number of stages, no improvement in the

Electronics 2021, 10, x FOR PEER REVIEW 11 of 21

Electronics 2021, 10, 1191 output will be obtained due to the little amount of available energy at the input affecting11 of 21 the conversion efficiency. Therefore, the number of stages with the best possible perfor- mance must be determined via testing.

4.3.4.3. PMUPMU ProposedProposed ArchitectureArchitecture ForFor the the present present application, application, a a commercial commercial device device could could be be used used as as a a PMU. PMU. These These availableavailable unitsunits havehave functionsfunctions suchsuch asas managingmanaging thethe chargingcharging andand dischargingdischarging of of storage storage devices,devices, providingproviding aa regulatedregulated supply supply voltage voltage to to the the components,components, and and even even generating generating an an alarmalarm whenwhen the the energy energy storage storage element element approaches approaches the the end end of itsof usefulits useful life [life31]. [31]. However, How- givenever, thegiven WiFi the signal WiFi behavior,signal behavior, the present the workpresent proposes work proposes a switching a switching unit using unit SCs using that canSCs maximize that can maximize the energy the from energy the PWM-like from the PWM waveform-like waveform of WiFi and of is WiFi sensitive and is enough sensitive to respondenough to reducedrespond timeto reduced intervals time such intervals as the BI.such Thus, as the the BI. energy Thus, available the energy for available the load for or athe following load or PMU,a following if required, PMU, isif improved.required, is improved. TheThe switchingswitching scheme scheme proposed proposed is is shown shown in in Figure Figure 11 11.. The The operation operation of theof the system system is basedis based on on the the charge-discharge charge-discharge cycle cycle of of the the two two SCs. SCs. An An SC, SC, identified identified as asCH CHor or harvester harvester capacitor,capacitor, remains remains connected connected to to the the rectenna rectenna charging charging with with the the harvested harvested energy. energy. A second A sec- SC,ond labeled SC, labeled as CF as orCF feeder or feeder capacitor, capacitor, supplies supplies power power to to the the load load (connected (connected atat S1A).S1A). TheThe twotwo SCsSCs areare controlledcontrolled byby anan SPDT-type SPDT-type electronic electronic switch switch (ADG884) (ADG884)governed governed bybya a microcontrollermicrocontroller (12F675). (12F675). TheThe microcontrollermicrocontroller compares compares the the two two capacitors capacitors voltage voltage values values toto generategenerate aa switchingswitching command to to the the SPDT SPDT to to allow allow the the transfer transfer of of energy energy from from CHCH to toCFCF whenwhen a sufficient a sufficient amount amount of energy of energy Δe is∆ availablee is available at CH at. ForCH recharging,. For recharging, CF is discon-CF is disconnected from the load during a time t (time necessary to guarantee the passage of nected from the load during a time t (time necessary to guarantee the passage of energy) energy) and connected to CH. After the time t has expired, the elements return to their and connected to CH. After the time t has expired, the elements return to their previous previous connections. Thus, CH receives power from the rectenna again; however, the connections. Thus, CH receives power from the rectenna again; however, the charging will charging will not start from zero but the prior-step residual charge. CF returns to supply not start from zero but the prior-step residual charge. CF returns to supply energy to the energy to the load but with a higher ∆e value than the beginning of the commutation. load but with a higher Δe value than the beginning of the commutation.

A nte nna A D G 884

S1A VCC Volta ge S1B D1 Multiplier IN1 CH CF RL R ecte nna

CTRL AN0 AN1 VCC 12F675 32kHz CAUX

FigureFigure 11.11.Proposed Proposed systemsystem blockblockdiagram. diagram.

AA microcontrollermicrocontroller is is chosen chosen to to govern govern the the switching switching process process of theof the two two SCs SCs to ease to ease the switchingthe switching instant instant selection, selection, allowing allowing greater greater versatility versatility and future and improvements future improvements in the load in exchangethe load exchange scheme programming. scheme programming. According According to requirements, to requirements, the 12F675 the microcontroller 12F675 micro- iscontroller selected is due selected to its lowdue powerto its low consumption, power consumption, availability availa of ability standby of a mode,standby voltage mode, comparators,voltage comparators, reduced footprint,reduced footprint, and pin configuration. and pin configuration. The idea is The that idea the useis that of harvested the use of energyharvested is minimally energy is spentminimally on the spent operation on the of operation the microcontroller. of the microcontroller. In low power In modelow power (LP), themode 12F675 (LP), draws the 12F675 a current draws of abouta current 8.5 ofµA about at 2 V 8. voltage5 μA at 2 when V voltage set to when LP mode. set to Therefore, LP mode. sinceTherefore, the ADG819 since the switch ADG819 has aswitch consumption has a consumption of 500 nA, the of consumption500 nA, the consumption estimate for esti- the proposedmate for the power proposed management power management circuit is approximately circuit is approximately 18µW. 18μW. Thus, according to the typical charging/discharging process, Figure 12 describes the time-domain behavior of the two SCs (CH and CF). As observed, there are two previous processes before the proposed scheme can start normal operation, Charging CH and Building up in CF.

Electronics 2021, 10, x FOR PEER REVIEW 12 of 21

Thus, according to the typical charging/discharging process, Figure 12 describes the time-domain behavior of the two SCs (CH and CF). As observed, there are two previous Electronics 2021, 10, 1191 12 of 21 processes before the proposed scheme can start normal operation, Charging CH and Building up in CF.

FigureFigure 12. 12. TimeTime domain domain sequence sequence for for the the operation operation of of CHCH andand CFCF. . Once the Normal Switching Operation is reached, it can be observed that four pa- Once the Normal Switching Operation is reached, it can be observed that four pa- rameters need to be defined. Vi_harvesting corresponds to the maximum voltage that it is rameters need to be defined. Vi_harvesting corresponds to the maximum voltage that it is ex- expected to have during charge in CH; Vf_harvesting is the final voltage observed in the CH pected to have during charge in CH; Vf_harvesting is the final voltage observed in the CH after after transferring energy. Similarly, Vi_feeder identifies the minimum voltage that can appear transferring energy. Similarly, Vi_feeder identifies the minimum voltage that can appear in in CF after a load supplying cycle is carried out, and Vf_harvesting is the voltage reached CF after a load supplying cycle is carried out, and Vf_harvesting is the voltage reached after after charging from CH. The selection of those parameters is important so that losses in the charging from CH. The selection of those parameters is important so that losses in the energy transfer process are minimized. energy transfer process are minimized. According to [33], the fractional losses in the process of energy transfer between According to [33], the fractional losses in the process of energy transfer between ca- capacitors can be expressed as pacitors can be expressed as 2  Vi_harversting  2 CF 푉𝑖_ℎ푎푟푣푒푟푠푡𝑖푛푔− 1 퐶퐹 ( Vi_ f eeder − 1) Normalized Fractional Losses = 푉𝑖_푓푒푒푑푒푟  , (9) 푁표푟푚푎푙𝑖푧푒푑 퐹푟푎푐푡𝑖표푛푎푙 퐿표푠푠푒푠 = V 2 , + i_harversting +2 CF (9) (CF CH) 푉𝑖V_ℎ푎푟푣푒푟푠푡𝑖푛푔 CH퐶퐹 (퐶퐹 + 퐶퐻) ( i_ f eeder + ) 푉𝑖_푓푒푒푑푒푟 퐶퐻 and Equation (9) applies when the voltage across both capacitors finally is the same. and Equation (9) applies when the voltage across both capacitors finally is the same. Alt- Although the proposed scheme is not expected to reach the same voltage in both capacitors, hough the proposed scheme is not expected to reach the same voltage in both capacitors, Equation (9) still provides a design criterium for minimizing losses in the process even if Equationthe energy (9) transfer still provides is partial. a design Reference criterium [33] for shows minimizing the generic losses plot in ofthe Equation process even (9) and if thestates energy that fractionaltransfer is lossespartial. are Reference minimized [33 when] shows both the capacitors generic plot are of the Equation same. However, (9) and statessince that the same fractional voltage losses is not are reached minimized in this when proposed both capacitors case, this statementare the same. does However, not apply sincethe same; the same therefore, voltage it would is not reached be shown in that this different proposed values case, ofthis capacitors statement actually does not produced apply thebetter same; results. therefore, it would be shown that different values of capacitors actually pro- ducedTesting better results. will be carried out to evaluate the operation of the scheme, and circuits includingTesting LC-like will be circuits carried could out to be evaluate employed the for operation improved of efficiency. the scheme, For and now, circuits the circuit in- cludingis kept asLC simple-like circuits as possible could for be demonstration employed for purposes.improved efficiency. For now, the circuit is kept as simple as possible for demonstration purposes. 5. Tests and Results 5.1. 2 × 1 Array Antenna Initially, a single patch antenna is built for testing; so that the performance of the antenna could be corroborated with the geometry reported for the CST software. Given that there are tolerances involved during the manufacturing process, an iterative trial and error process is followed to tune the definitive antenna dimensions finely. Once the single patch is obtained, the 2 × 1 array is developed, and a Keysight E5071C Vector Network Analyzer (VNA) is employed for the verification of array parameters. Figure 13 summarizes the results. The measured S11 corresponds to −39.30 dB at a frequency of 2.43 GHz. Another test for the validation of the antenna gain is demonstrated with the experiment shown in Electronics 2021, 10, x FOR PEER REVIEW 13 of 21 Electronics 2021, 10, x FOR PEER REVIEW 13 of 21

5. Tests and Results 5. Tests and Results 5.1. 2 × 1 Array Antenna 5.1. 2 × 1 Array Antenna Initially, a single patch antenna is built for testing; so that the performance of the antennaInitially, could a besingle corroborated patch antenna with theis built geometry for testing; reported so thatfor thethe CSTperformance software. ofGiven the antennathat there could are tolerancesbe corroborated involved with during the geometry the manufacturing reported forprocess, the CST an iterativesoftware. tr Givenial and thaterror there process are tolerances is followed involved to tune theduring definitive the manufacturing antenna dimensions process, finely. an iterative Once thetrial single and errorpatch process is obtained, is followed the 2 to× 1tune array the is definitive developed, antenna and a dimensions Keysight E5071C finely. VectorOnce the Network single patchAnalyzer is obtained, (VNA) isthe employed 2 × 1 array for isthe developed, verification and of aarray Keysight parameters. E5071C F Vectorigure 13 Network summa- Electronics 2021, 10, 1191 Analyzer (VNA) is employed for the verification of array parameters. Figure 13 summa-13 of 21 rizes the results. The measured S11 corresponds to −39.30 dB at a frequency of 2.43 GHz. rizesAnother the results.test for Thethe validationmeasured Sof11 thecorresponds antenna gain to −3 is9.3 demonstrated0 dB at a frequency with the of experiment 2.43 GHz. Anothershown in test Figure for the 14. validation In this test, of the the measurement antenna gain ofis freedemonstrated-space power with losses the experimentat distances shown in Figure 14. In this test, the measurement of free-space power losses at distances Figureof 0.3 14m. and In this 0.6 test, m is the performed, measurement obtaining of free-space results power equivalent losses to at distancesthose of the of 0.3simulated m and ofgain. 0.3 Freem and-space 0.6 mlosses is performed, are calculated obtaining with: results equivalent to those of the simulated 0.6gain. m Free is performed,-space losses obtaining are calculated results equivalent with: to those of the simulated gain. Free-space losses are calculated with:  4πd  FSL = 20 log 4휋푑, (10)(10) FSL = 20 푙표푔 (4휋푑 ), (10) FSL = 20 푙표푔λ( 휆 ), 휆 andand the the test test results results are are summarized summarized in in Table Table1. 1. and the test results are summarized in Table 1.

Figure 13. Parameter S11 array 2 × 1 measured. FigureFigure 13.13. Parameter SS1111 array 2 × 11 measured. measured.

TX Antenna R F Generator A ntenna Under Test Spectrum Analyzer TX Antenna R F Generator A ntenna Under Test Spectrum Analyzer

Figure 14. Testing with fixed power emission. Figure 14. Testing with fixed power emission. Figure 14. Testing with fixed power emission.

Table 1. 2 × 1 Array gain validation.

Power RX Incident EIRP (dBm) Distance (m) FSL (dB) Gain (dB) (dBm) Power (dBm)

19 0.3 29.7 −3.6 −10.7 7.1 19 0.6 35.7 −9.99 −16.71 6.72

5.2. Voltage Multiplier The multipliers have been implemented using discrete PCB developed using Proteus software. The selected diodes are Schottky SMS 7630-005LF, and the connection interface with the antenna is given through an RF SMA connector (impedance of 50 Ω and operating range up to 18 GHz), see Figure 15. ElectronicsElectronics 2021 2021, 10, ,10 x ,FOR x FOR PEER PEER REVIEW REVIEW 14 14of of21 21

TableTable 1. 1.2 ×2 1× Array1 Array gain gain validation. validation.

EIRPEIRP (dBm) (dBm) Distance Distance (m) (m) FSL FSL (dB) (dB) Power Power RX RX (dBm) (dBm) Incident Incident Power Power (dBm) (dBm) Gain Gain (dB) (dB) 1919 0.30.3 29.729.7 −3−3.6 .6 −1−10.70.7 7.17.1 1919 0.60.6 35.735.7 −9−9.99.99 −1−16.716.71 6.726.72

5.2.5.2. Voltage Voltage Multiplier Multiplier TheThe multipliers multipliers have have been been implemented implemented using using discrete discrete PCB PCB developed developed using using Proteus Proteus Electronics 2021, 10, 1191 software.software. The The selected selected diodes diodes are are Schottky Schottky SMS SMS 7630 7630-005LF,-005LF, and and the the connection connection interface interface14 of 21 withwith the the antenna antenna is isgiven given through through an an RF RF SMA SMA connector connector (impedance (impedance of of 50 50 Ω Ω and and operat- operat- inging range range up up to to 18 1 G8 Hz),GHz), see see Figure Figure 15. 15.

Figure 15. 4, 5, 6, 7, and 8 stage multipliers. FigureFigure 15. 15.4, 4, 5, 5, 6, 6, 7, 7, and and 8 8 stage stage multipliers. multipliers.

ComparedComparedCompared to to to LTSpice LTSpice LTSpice simulations, simulations, simulations, the the the physical physical physical circuits circuits circuits present present present similar similar similar values, values,values, show- show-show- inginging empirically empirically empirically that thatthat the thethe s seven-stageeven seven-stage-stage multiplier multipliermultiplier is is isthe thethe one one one with with with the the the best best best performance, performance, performance, as as as observedobservedobserved in inin Figures FiguresFigures 16 16 16 and and and 17, 17 17, ,reaching reaching reaching a a ahigher higher higher percentage percentage percentage of of of agreement agreement agreement between between between the the the theoreticaltheoreticaltheoretical and andand measured measuredmeasured values valuesvalues at atat 85% 85%85% with withwith an anan input input input voltage voltagevoltage of ofof 40 400 400 0m mV mV Vaccording according according to to to EquationEquationEquation 6. (6). 6.For For Forproper proper proper testing, testing, testing, a controlleda controlled a controlled power power power generator generator generator source source source is isused used is given used given the given the men- men- the tionedmentionedtioned nature nature nature of of the the of WiFi theWiFi WiFisignal. signal. signal.

Electronics 2021, 10, x FOR PEER REVIEW 15 of 21

Figure 16. Performance multipliers 4, 5, 6, 7, and 8 stages with a load of 10 kΩ. FigureFigure 16. 16. Performance Performance multipliers multipliers 4, 4,5, 5,6, 6,7, 7,and and 8 stages8 stages with with a loada load of of10 1 k0Ω. kΩ.

FigureFigure 17. 17. Theoretical,Theoretical, simulated, simulated, and and measured measured open open circuit circuit voltage voltage fo forr 7 7 stages. stages.

AsAs commented, commented, considering considering that that impedance impedance matching matching is is a akey key factor factor for for the the power power transfer between circuits, the input impedance measurement is performed for each multi- transfer between circuits, the input impedance measurement is performed for each mul- plier. The results are summarized in Tables2–4. Measurements are developed using the tiplier. The results are summarized in Tables 2 to 4. Measurements are developed using Keysight E5071C VNA. the Keysight E5071C VNA. Given its best performance, the seven-stage multiplier presents an impedance of 51.297 + j11.922 Ω for the 2.412 GHz frequency (Channel 1 of the 2.4 GHz WiFi band), while for Channels 6 and 11, the value increases to 62 Ω and 72 Ω in the real part. It is clear how impedance is affected by frequency due to the non-linear nature of the multi- plier.

Table 2. Impedance of multipliers of 5, 6, 7, and 8 stages for Channel 1.

CH GHz 5 Stages 6 Stages 7 Stages 8 Stages REAL IMAG REAL IMAG REAL IMAG REAL IMAG 1 2.412 32.96 −34.79 29 0 51.29 11.92 32.85 −0.18

Table 3. Impedance of multipliers of 5, 6, 7, and 8 stages for Channel 6.

CH GHz 5 Stages 6 Stages 7 Stages 8 Stages REAL IMAG REAL IMAG REAL IMAG REAL IMAG 6 2.437 27.44 −23.5 47 32 62.43 11.10 35.17 5.256

Table 4. Impedance of multipliers of 5, 6, 7, and 8 stages for Channel 11.

CH GHz 5 Stages 6 Stages 7 Stages 8 Stages REAL IMAG REAL IMAG REAL IMAG REAL IMAG 11 2.462 24.9 −14.2 97 5 72.67 0.95 38.86 10.14

Another parameter validated is the antenna array input impedance, as summarized in Table 5. Here, the measured impedance of the 7× multiplier is listed observed the prox- imity in values for the associated impedances. Thus, for this particular case, it can be ob- served empirically that an impedance matching network could be eliminated from the proposed prototype simplifying the receiver design.

Table 5. 7× multiplier and array impedance for WiFi Channels 1, 6, and 11.

Multiplier Array Channel Z Real Z Imaginary Z Real Z Imaginary 2.412 51.297 11.922 49.252 6.835 2.437 62.437 11.108 60.46 −0.659 2.462 72.674 0.951 57.352 −7.467

Electronics 2021, 10, 1191 15 of 21

Table 2. Impedance of multipliers of 5, 6, 7, and 8 stages for Channel 1.

CH GHz 5 Stages 6 Stages 7 Stages 8 Stages REAL IMAG REAL IMAG REAL IMAG REAL IMAG 1 2.412 32.96 −34.79 29 0 51.29 11.92 32.85 −0.18

Table 3. Impedance of multipliers of 5, 6, 7, and 8 stages for Channel 6.

CH GHz 5 Stages 6 Stages 7 Stages 8 Stages REAL IMAG REAL IMAG REAL IMAG REAL IMAG 6 2.437 27.44 −23.5 47 32 62.43 11.10 35.17 5.256

Table 4. Impedance of multipliers of 5, 6, 7, and 8 stages for Channel 11.

CH GHz 5 Stages 6 Stages 7 Stages 8 Stages REAL IMAG REAL IMAG REAL IMAG REAL IMAG 11 2.462 24.9 −14.2 97 5 72.67 0.95 38.86 10.14

Given its best performance, the seven-stage multiplier presents an impedance of 51.297 + j11.922 Ω for the 2.412 GHz frequency (Channel 1 of the 2.4 GHz WiFi band), while for Channels 6 and 11, the value increases to 62 Ω and 72 Ω in the real part. It is clear how impedance is affected by frequency due to the non-linear nature of the multiplier. Another parameter validated is the antenna array input impedance, as summarized in Table5. Here, the measured impedance of the 7 × multiplier is listed observed the proximity in values for the associated impedances. Thus, for this particular case, it can be observed empirically that an impedance matching network could be eliminated from the proposed prototype simplifying the receiver design.

Table 5. 7× multiplier and array impedance for WiFi Channels 1, 6, and 11.

Multiplier Array Channel Z Real Z Imaginary Z Real Z Imaginary 2.412 51.297 11.922 49.252 6.835 2.437 62.437 11.108 60.46 −0.659 2.462 72.674 0.951 57.352 −7.467

Once the rectenna is assembled, Table6 summarizes the obtained DC component from the power setting from the controlled generator source. The listed efficiency employs the definition of Equation (7) and a load equal to 10 kΩ.

Table 6. Harvester efficiency for WiFi channels 1 and 6 RL 10k.

Pot_In Vout IRL Ef % Vout IRL Ef % −5 dBm 0.72 V 72 µA 16.39 0.6 V 63 µA 12.55 −4 dBm 0.93 V 94 µA 21.96 0.85 V 84 µA 17.72 −3 dBm 1.19 V 120 µA 28.49 1.10 V 110 µA 24.14 −2 dBm 1.49 V 148 µA 34.94 1.43 V 140 µA 31.06 −1 dBm 1.84 V 185 µA 42.85 1.74 V 175 µA 38.55 Electronics 2021, 10, 1191 16 of 21

Table 6. Cont.

Pot_In Vout IRL Ef % Vout IRL Ef % −0 dBm 2.23 V 225 µA 50.18 2.15 V 216 µA 46.22 1 dBm 2.7 V 273 µA 58.55 2.61 V 264 µA 54.73 2 dBm 2.24 V 327 µA 66.85 3.15 V 318 µA 63.20 3 dBm 3.85 V 389 µA 75.06 3.75 V 379 µA 71.23 4 dBm 4.54 V 459 µA 82.96 4.44 V 449 µA 79.37 5 dBm 5.31 V 530 µA 89.0 5.21 V 527 µA 86.83

5.3. PMU Core Circuit The scheme proposed in Figure 11 is called the PMU core since another commercial PMU can be added for higher functionality. Thus, an experiment is developed to validate the operation of the PMU core. Moreover, the selection of CH and CF is presented based on the claim that different values of capacitors actually minimize the energy transfer losses. After preliminary tests, Vi_harvesting is found to be as high as 3.3 V (100%). At the same time, Vi_feeder is selected to be equal to 1.5 V, which is the lowest power supply for the microcontroller and switch to operate adequately. It is necessary to consider the possible losses generated during energy transfer [32], the potential energy harvested, and the load operating time to define CF and CH values. The latter must not be longer than the energy harvest time sufficient to recharge from CH to CF before the energy in CF is insufficient to keep the load operating for at least 1 h. Here, the load includes the PMU circuit containing the microcontroller and switch. Thus, CF is selected by evaluating the amount of energy required to power the load estimating the target time using   Vf _ f eeder − Vi_ f eeder t = CF , (11) ILmax

where t corresponds to the operation target, and IL_max corresponds to the maximum current consumption of the system. Hence, revising IoT development boards such as the TI eZ430-RF2500, it can be considered an IL_max equal to 50 µA; then a suitable value for CH becomes 360 mF, rounding Vi_feeder and Vf_feeder to 1.5 V and 2.0 V, respectively. Once CF has been selected, a parametric simulation of Equation (9) is carried out to identify a proper value for CH. For this work, the software LTSpice is employed for the simulations using the feature of behavioral sources to implement the expression for the fractional losses. The different parameters and variables are also implemented via voltages sources, using the DC analysis (DC Sweep) to find a suitable value for CH. Figure 18 shows the DC analysis results, and although LTSpice reports the y-axis as voltage, from the simulation setup, it should be clear that this axis represents the percentage loss. The parametric simulations show the results for possible and available values of CH, ranging from 50 mF to 350 mF. The loss profiles obtained are highly dependent on the selected voltage values. Thus, there is an advantage in using a microcontroller for setting the switching thresholds to keep the losses as low as possible. ElectronicsElectronics2021 2021, 10, 10, 1191, x FOR PEER REVIEW 17 ofof 2121 Electronics 2021, 10, x FOR PEER REVIEW 17 of 21

Figure 18. Parametric simulations of the normalized fractional losses for the harvester. FigureFigure 18. 18.Parametric Parametric simulations simulations of of the the normalized normalized fractional fractional losses losses for for the the harvester. harvester.

Thus,Thus,Thus, thethe the twotwo capacitors’ capacitors’capacitors’ values values are are are finally finally finally 4 47 477 m m mFFF and and and 33 330 3300 m mF mFF for for forCH CHCH and andand CF CF,, CFrespec- respec-, re- spectively,tively,tively, due due due toto SC SC to valueSCvalue value commercialcommercial commercial availability. availability. availability. Therefore,Therefore, Therefore, givengiven given thethe the conditionsconditions conditions statedstated stated forfor Equation (11), maintaining the IL_max equal to 50 mA; then, the energy autonomy is re- forEquation Equation (11 (11),), maintaining maintaining the the IL_maxIL_max equalequal to 5 to0 50mA; mA; then, then, the the energy energy autonomy autonomy is isre- restrictedstrictedstricted toto to aboutabout about 1111 11 min min whichwhich still still is is attractive attractiveattractive toto IoTIoT modules.modules. FigureFigure 19 1919 shows showsshows the thethe built builtbuilt PMUPMUPMU core. core.core.

H arvester SC H arvester SC

Feeder SC Feeder SC

M icrocontroller M icrocontroller & & C rystal C rystal

M ultiplier M ultiplier C onnector C onnector

Figure 19. Protoype of the built harvester. FigureFigure 19. 19.Protoype Protoype of of the the built built harvester. harvester. 5.4. Test of the Energy Harvester with a Multi SSID Router 55.4..4. TestTest ofof thethe EnergyEnergy HarvesterHarvester withwith aa MultiMulti SSIDSSID RouterRouter Tests are carried out in a real scenario placing the harvester exposed to a multi TestsTests areare carriedcarried outout inin aa realreal scenarioscenario placingplacing thethe harvesterharvester exposedexposed toto aa multimulti SSIDSSID SSIDrouter router to verify to verify the harvester the harvester performance. performance. The multi The multiSSID SSIDrouter router is located is located in the in Engi- the Engineeringrouter to verify Laboratories the harvester of the performance. ITSA. Figure The20 shows multi theSSID test router set up. is located in the Engi- neeringneering LaboratoriesLaboratories ofof thethe ITSA.ITSA. FigureFigure 2020 showsshows thethe testtest setset up.up.

Electronics 2021, 10, x FOR PEER REVIEW 18 of 21 Electronics 2021, 10, x 1191 FOR PEER REVIEW 1818 of of 21 21

H arvester H arvester

d d

M ulti SS ID Router M ulti SS ID Router

Figure 20. Test set up for the energy harvester with a multi SSID router. FigurFiguree 20. 20. TestTest set set up up for the energy harvester with a m multiulti SSID router. Figure 21 shows the test results. In this experiment, the harvester is left for 12 h, and FigureFigure 2121 showsshows the the test test results. results. In In this this experiment, experiment, the the harvester harvester is is left for 12 h h,, and and the time selected is the daily hours where mainly PCs and smartphones are present. The thethe time time selected selected is is the the daily daily hours hours where where mainly mainly PCs PCs and and smartphones are are present. The The key idea is to create variable traffic in the WiFi network. As can be seen, there are four keykey idea idea i iss to to create variable traffic traffic in the WiFi network. As As can can be be seen, seen, there there are are four intervals marked on the graph, from 0 to 120, 120 to 400, 400 to 560, and 560 to 800 min. intervalsintervals marked marked on on the the graph, graph, from from 0 0 to to 120, 120, 120 120 to to 400, 400, 400 400 to to 560, 560, and and 560 560 to to 800 800 min min.. This behavior is directly associated with network traffic being video streaming the main ThisThis behavior behavior is is directly directly associated associated with with network network traffic traffic being video streaming the m mainain application that makes the channel busier with transfer rates between 0.5 Mbps to 25 applicationapplication that that makes makes the the channel channel busier busier with with transfer transfer rates rates between between 0.5 Mbps 0.5 M tobps 25 Mbps. to 25 Mbps. There is an increase in the SC voltage for the first segment up to about 120 min; MTherebps. isThere an increase is an increase in the in SC the voltage SC voltage for the for first the segment first segment up to up about to about 120 min; 120 later,min; later, there is a valley where the voltage remains relatively the same, which is due to a later,there there is a valley is a valley where where the voltage the voltage remains remains relatively relatively the same, the which same, is which due to is a due decrease to a decrease in network traffic from applications with higher data consumption, such as decreasein network in traffic network from traffic applications from applications with higher with data higher consumption, data consumption, such as YouTube such for as YouTube for streaming/watching videos for classes and tutoring sessions. After 360 min, YouTubestreaming/watching for streaming/watching videos for classes videos and for tutoring classes and sessions. tutoring After sessions. 360 min, After an increase 360 min is, an increase is presented again due to the rise in traffic carried, and then it returns to have anpresented increase again is presented due to theagain rise due in trafficto the rise carried, in traffic and thencarried, it returns and then to haveit returns low growth.to have low growth. For this test, the distance is maintained at 0.25 m throughout the energy har- lowFor growth. this test, For the this distance test, the is maintained distance is maintained at 0.25 m throughout at 0.25 m throughout the energy harvesting.the energy har- It is vesting. It is observed the prototype can harvest about 494 mJ, which is equivalentµ to 137 vesting.observed It theis observed prototype the can prototype harvest can about harvest 494 mJ, about which 494 ismJ, equivalent which is equivalent to 137 Wh. to 137 For μexample,Wh. For example, suppose ansuppose IoT temperature an IoT temperature sensor is available,sensor is available, and it reports and it its reports measure its duringmeas- μWh. For example, suppose an IoT temperature sensor is available, and it reports its meas- ure5 s, during working 5 s at, working 3.3 V power at 3.3 supply. V power In thatsupply. case, In there that iscase, available there is about available 30 mA about to complete 30 mA ure during 5 s, working at 3.3 V power supply. In that case, there is available about 30 mA toits complete task, which its task, enough which for enough this application. for this application. to complete its task, which enough for this application.

FigureFigure 21. 21. ChargingCharging time time for for a a220 220 mF mF SC SC under under the the testing testing environment environment.. Figure 21. Charging time for a 220 mF SC under the testing environment. As an evaluation of the proposed harvester, Table7 compares the results obtained for similar state-of-art works. It is important to recall that the efficiency refers to Equation (7), which is used in the literature.

Electronics 2021, 10, 1191 19 of 21

Table 7. Comparison of results vs. other similar studies.

Input Power Frequency Reference Load Efficiency S (dB) (dBm) (GHz) 11 [28] 8 57.8 2.42 −27 [29] 5 k 5 68 2.45 −37 [30] 2 k 0 30 2.4 −20 [31] 1 M 4, 6, 8 45.6–53.8 61.6 2.4 [32] 0.5 k 2 59 2.45 −20 [23] 12 k 0 45 2.4 −13 [this work] 10 k 0 50.18 2.46 −39

6. Conclusions This work presents an energy harvester for a 2.4 GHz WiFi MultiSSID router, which exhibits an estimated efficiency of 50.18% over an incident power of 0 dBm. The harvester has been designed based on the analysis of WiFi signals, and, through empirical testing, it has been possible to remove an impedance matching network due to the measured impedances for the proposed versions of the antenna and the voltage multiplier. Thus, the system complexity has been reduced without sacrificing performance. The harvester has been completed with a PMU module that simultaneously provides charge to the load and harvests energy from WiFi via a switching scheme. Such scheme is controlled by a low-power microcontroller that can work with supply values as low as 1.5 V and current drawing of about 18 µA. The switching scheme adjusts to the PWM-like nature of the WiFi signal and provides rapid power availability for load operation. The advantage of using a microcontroller-based switching scheme lies in the possibility of adjusting through programming the switching threshold to minimize losses in the charging losses between two SCs that serve as storage units. The built prototype shows that thanks to a basic analysis of the WiFi signals, it is possible to have the required amount of energy for a sensor operation from a multiple SSID router. It is important to note that the router employed does not require additional modification but the BI to be used as a source generator. As future work, it is possible to optimize the antenna to reduce its dimensions, increase its gain, and improve the conversion efficiency on power less than −5 dBm. Regarding the PMU, the objective is to continue improving its performance in terms of minimizing losses. This work can be done on three fronts; one, to optimize the selection of capacitors CH and CF; second, to work with adjustable and adaptive switching thresholds; and finally, to explore LC circuits to increase energy transfer efficiency. Besides, it is also possible the implementation of D2D encryption mechanisms such as the one proposed in [6] thanks to the availability of an on-board microcontroller.

Author Contributions: Formal analysis, F.A.; investigation, F.A. and M.P.; methodology, F.A., C.G.Q.M. and M.P.; supervision, C.G.Q.M. and M.P.; validation, F.A. and L.N.; writing—original draft, F.A., L.N., C.G.Q.M. and M.P.; writing—review and editing, F.A., L.N., C.G.Q.M. and M.P. All authors have read and agreed to the published version of the manuscript. Funding: This research received no external funding. Acknowledgments: This work was supported by Universidad del Norte, and by the College of Engineering, Institución Universitaria ITSA Barranquilla—Colombia. Conflicts of Interest: The authors declare that there are no conflicts of interest regarding the publica- tion of this paper. Electronics 2021, 10, 1191 20 of 21

References 1. Lu, X.; Wang, P.; Niyato, D.; Kim, D.I.; Han, Z. Wireless Networks with RF Energy Harvesting: A Contemporary Survey. IEEE Commun. Surv. Tutor. 2015, 17, 757–789. [CrossRef] 2. Kamalinejad, P.; Mahapatra, C.; Sheng, Z.; Mirabbasi, S.; Leung, V.C.M.; Guan, Y.L. Wireless energy harvesting for the Internet of Things. IEEE Commun. Mag. 2015, 53, 102–108. [CrossRef] 3. Fan, P.M.-Y.; Wong, O.-Y.; Chung, M.-J.; Su, T.-Y.; Zhang, X.; Chen, P.-H. Energy harvesting techniques: Energy sources, power management and conversion. In Proceedings of the 2015 European Conference on Circuit Theory and Design (ECCTD), Trondheim, Norway, 24–26 August 2015; Volume 1, pp. 1–4. [CrossRef] 4. Panahi, F.H.; Hajimirzaee, P.; Erfanpoor, S.; Panahi, F.H.; Ohtsuki, T. Smart Image-Processing based Energy Harvesting for Green Internet of Things. In Proceedings of the 2018 Smart Grid Conference (SGC), Sanandaj, Iran, 28–29 November 2018; pp. 1–5. 5. Alsamhi, S.; Afghah, F.; Sahal, R.; Hawbani, A.; Al-Qaness, M.A.; Lee, B.; Guizani, M. Green internet of things using UAVs in B5G networks: A review of applications and strategies. Ad Hoc Netw. 2021, 117, 102505. [CrossRef] 6. Tan, H.; Song, Y.; Xuan, S.; Pan, S.; Chung, I. Secure D2D Group Authentication Employing Smartphone Sensor Behavior Analysis. Symmetry 2019, 11, 969. [CrossRef] 7. Muncuk, U.; Alemdar, K.; Sarode, J.D.; Chowdhury, K.R. Multiband Ambient RF Energy Harvesting Circuit Design for Enabling Batteryless Sensors and IoT. IEEE Internet Things J. 2018, 5, 2700–2714. [CrossRef] 8. Sidhu, R.K.; Ubhi, J.S.; Aggarwal, A. A Survey Study of Different RF Energy Sources for RF Energy Harvesting. In Proceedings of the 2019 International Conference on Automation, Computational and Technology Management (ICACTM), London, UK, 24–26 April 2019; pp. 530–533. [CrossRef] 9. Liu, J.; Zhang, X.Y. Compact Triple-Band Rectifier for Ambient RF Energy Harvesting Application. IEEE Access 2018, 6, 19018– 19024. [CrossRef] 10. Sarma, S.S.; Akhtar, M.J. A dual band meandered printed dipole antenna for RF energy harvesting applications. In Proceedings of the 2016 IEEE 5th Asia-Pacific Conference on Antennas and Propagation (APCAP), Kaohsiung, Taiwan, 26–29 July 2016; Volume 3, pp. 93–94. 11. Lopez-Yela, A.; Segovia-Vargas, D. A triple-band bow-tie rectenna for RF energy harvesting without matching network. In Proceedings of the 2017 IEEE Wireless Power Transfer Conference (WPTC), Taipei, Taiwan, 10–12 May 2017. [CrossRef] 12. Kurvey, M.; Kunte, A. Tri-Stepped Rectangular Antenna for Efficient RF Energy Harvesting. J. Commun. Inf. Netw. 2018, 3, 86–90. [CrossRef] 13. Ermeey, A.K.; Hu, A.P.; Biglari-Abhari, M.; Aw, K.C. Indoor 2.45 GHz Wi-Fi Energy Harvester With Bridgeless Converter. IEEE J. Sel. Areas Commun. 2016, 34, 1536–1549. [CrossRef] 14. Gudan, K.; Shao, S.; Hull, J.J.; Ensworth, J.; Reynolds, M.S. Ultra-low power 2.4 GHz RF energy harvesting and storage system with 25 dBm sensitivity. In Proceedings of the 2015 IEEE International Conference on RFID (RFID), San Diego, CA, USA, 15–17 April 2015; pp. 40–46. 15. Nechibvute, A.; Chawanda, A.; Taruvinga, N.; Luhanga, P. Radio Frequency Energy Harvesting Sources. Acta Electrotech. Inform. 2017, 17, 19–27. [CrossRef] 16. Song, C.; Huang, Y.; Zhou, J.; Carter, P. Recent advances in broadband rectennas for wireless power transfer and ambient RF energy harvesting. In Proceedings of the 2017 11th European Conference on Antennas and Propagation (EUCAP), Paris, France, 19–24 March 2017; pp. 341–345. 17. Mansour, M.M.; Kanaya, H. Compact RF rectifier circuit for ambient energy harvesting. In Proceedings of the 2017 IEEE International Symposium on Radio-Frequency Integration Technology (RFIT), Seoul, Korea, 30 August–1 September 2017; pp. 220–222. 18. Kadir, E.A.; Hu, A.P.; Biglari-Abhari, M.; Aw, K.C. Indoor WiFi energy harvester with multiple antenna for low-power wireless applications. In Proceedings of the 2014 IEEE 23rd International Symposium on Industrial Electronics (ISIE), Istanbul, Turkey, 1–4 June 2014; pp. 526–530. [CrossRef] 19. Talla, V.; Kellogg, B.; Ransford, B.; Naderiparizi, S.; Smith, J.R.; Gollakota, S. Powering the next Billion devices with wi-fi. Commun. ACM 2017, 60, 83–91. [CrossRef] 20. Milani, I.; Colone, F.; Bongioanni, C.; Lombardo, P. Impact of Beacon Interval on the performance of WiFi-based passive radar against human targets. In Proceedings of the 2018 22nd International Microwave and Radar Conference (MIKON), Poznan, Poland, 14–17 May 2018; pp. 190–193. [CrossRef] 21. Chen, X.; Huang, L.; Xing, J.; Shi, Z.; Xie, Z. Energy harvesting system and circuits for ambient WiFi energy harvesting. In Proceedings of the 2017 12th International Conference on Computer Science and Education (ICCSE), Houston, TX, USA, 22–25 August 2017; pp. 769–772. [CrossRef] 22. Al-Azawy, M.M.; Sari, F. Analysis of Dickson Voltage Multiplier for RF Energy Harvesting. In Proceedings of the 2019 1st Global Power, Energy and Communication Conference (GPECOM), Nevsehir, Turkey, 12–15 June 2019; pp. 10–14. [CrossRef] 23. Hong, S.S.B.; Ibrahim, R.; Khir, M.H.M.; Daud, H.; Zakariya, M.A. Rectenna architecture based energy harvester for low power RFID application. In Proceedings of the 2012 4th International Conference on Intelligent and Advanced Systems (ICIAS2012), Lumpur, Malaysia, 12–14 June 2012; Volume 1, pp. 382–387. [CrossRef] 24. Newell, D.; Duffy, M. Review of Power Conversion and Energy Management for Low-Power, Low-Voltage Energy Harvesting Powered Wireless Sensors. IEEE Trans. Power Electron. 2019, 34, 9794–9805. [CrossRef] Electronics 2021, 10, 1191 21 of 21

25. Kim, S.; Chou, P.H. Energy harvesting: Energy harvesting with supercapacitor-based energy storage. In Smart Sensor and System; Springer: Cham, Switzerland, 2015; pp. 215–241. 26. Waguaf, A.; Alvernhe, R.; Fadel, L.; Grzeskowiak, M. Energy Harvesting with 2.45 GHz Rectenna for urban application. In Proceedings of the 2018 25th IEEE International Conference on Electronics, Circuits and Systems (ICECS), Bordeaux, France, 9–12 December 2018; pp. 345–348. 27. Saeed, W.; Shoaib, N.; Cheema, H.M.; Khan, M.U. RF Energy Harvesting for Ubiquitous, Zero Power Wireless Sensors. Int. J. Antennas Propag. 2018, 2018, 1–16. [CrossRef] 28. Chaour, I.; Fakhfakh, A.; Kanoun, O. Patch Antenna Array for RF Energy Harvesting Systems in 2.4 GHz WLAN Frequency Band. In Proceedings of the 2018 15th International Multi-Conference on Systems, Signals & Devices (SSD), Hammamet, Tunisia, 19–22 March 2018; pp. 179–183. [CrossRef] 29. He, H.; Li, T. Design of an indoor RF Energy Harvesting Module using Dual-Band Rectenna Array for IoT Applications. In Proceedings of the 2019 IEEE 4th Advanced Information Technology, Electronic and Automation Control Conference (IAEAC), Chengdu, China, 20–22 December 2019; Volume 1, pp. 1407–1410. 30. Tran, L.-G.; Cha, H.-K.; Park, W.-T. RF power harvesting: A review on designing methodologies and applications. Micro Nano Syst. Lett. 2017, 5, 14. [CrossRef] 31. Charris, D.; Gomez, D.; Ortega, A.R.; Carmona, M.; Pardo, M. A Thermoelectric Energy Harvesting Scheme with Passive Cooling for Outdoor IoT Sensors. Energies 2020, 13, 2782. [CrossRef] 32. Song, C.; Huang, Y.; Zhou, J.; Carter, P.; Yuan, S.; Xu, Q.; Fei, Z. Matching Network Elimination in Broadband Rectennas for High-Efficiency Wireless Power Transfer and Energy Harvesting. IEEE Trans. Ind. Electron. 2017, 64, 3950–3961. [CrossRef] 33. Levine, R.C. Apparent Nonconservation of Energy in the Discharge of an Ideal Capacitor. IEEE Trans. Educ. 1967, 10, 197–202. [CrossRef]