Article Development of Intelligent Drone Battery Charging System Based on Wireless Power Transmission Using Hill Climbing Algorithm Ali Rohan 1, * , Mohammed Rabah 1 , Muhammad Talha 1 and Sung-Ho Kim 2 1 2 * Department of Electrical, Electronics and Information Engineering, Kunsan National University, Gunsan-Si 573-360, Korea; mohamedmostafamousa1991@gmail.com (M.R.); engrtalha72@gmail.com (M.T.) Department of Control and Robotics Engineering, Kunsan National University, Gunsan-Si 573-360, Korea; shkim@kunsan.ac.kr Correspondence: ali_rohan2003@hotmail.com; Tel.: +82-10-2857-6080 Received: 13 September 2018; Accepted: 5 November 2018; Published: 7 November 2018



 Abstract: In this work, an advanced drone battery charging system is developed. The system is composed of a drone charging station with multiple power transmitters and a receiver to charge the battery of a drone. A resonance inductive coupling-based wireless power transmission technique is used. With limits of wireless power transmission in inductive coupling, it is necessary that the coupling between a transmitter and receiver be strong for efficient power transmission; however, for a drone, it is normally hard to land it properly on a charging station or a charging device to get maximum coupling for efficient wireless power transmission. Normally, some physical sensors such as ultrasonic sensors and infrared sensors are used to align the transmitter and receiver for proper coupling and wireless power transmission; however, in this system, a novel method based on the hill climbing algorithm is proposed to control the coupling between the transmitter and a receiver without using any physical sensor. The feasibility of the proposed algorithm was checked using MATLAB. A practical test bench was developed for the system and several experiments were conducted under different scenarios. The system is fully automatic and gives 98.8% accuracy (achieved under different test scenarios) for mitigating the poor landing effect. Also, the efficiency η of 85% is achieved for wireless power transmission. The test results show that the proposed drone battery charging system is efficient enough to mitigate the coupling effect caused by the poor landing of the drone, with the possibility to land freely on the charging station without the worry of power transmission loss. Keywords: wireless power transfer; unmanned aerial vehicle; automatic charging station; drone station; hill climbing 1. Introduction 1.1. Introduction and Motivation The quadcopter, also known as a quadrotor, is a type of unmanned aerial vehicle (UAV), lifted and propelled by four rotors [2,3]. Quadcopters use two pairs of identical fixed pitched propellers: two clockwise and two counter-clockwise. The quadcopter has high maneuverability, as it can hover, take off, cruise, and land in narrow areas. It also has a simpler control mechanism compared to other UAVs [4] and is equipped with different components such as an inertial measurement unit (IMU), a global positioning system (GPS), an electronic speed control (ESC), a standard radio control (RC), a radio-frequency module (RF) used to transmit live videos to a personal computer (PC), and a flight controller [5,6]. Appl. Syst. Innov. 2018, 1, 44; doi:10.3390/asi1040044 www.mdpi.com/journal/asi Appl. Syst. Innov. 2018, 1, 44 2 of 19 The quadcopter is used for applications such as surveillance, search and rescue, and object detection [7–9]. As mentioned earlier, a quadcopter with two pairs of fixed pitched propellers has a very short operation time because it has to generate lift force all the time to move around, which requires high electrical power. Batteries are used as an electrical power source in quadcopters; however, because of the high electrical power requirement, the normal operation time of a quadcopter is just 20 to 30 min [10]. This limits the quadcopter flight range and operation time drastically, and accordingly, the quadcopter might not be able to fulfill the purpose of its use in a specific application. Generally, to continuously operate quadcopters, the batteries are changed or recharged after the normal operation time. These batteries can be recharged using wired power transmission which requires some physical connection or via wireless power transmission which does not require any physical connection. Even though wired power transmission is more efficient than wireless power transmission, the wireless power transmission technique is currently utilized for its lower maintenance and increased safety (due to no physical connection of wires) for the delivery of power [11,12]. In order to charge the battery using wireless power transmission, the quadcopter is equipped with electromagnetic coils. These coils can be single or multiple depending upon the size and design of the charging system. Generally, a wireless power transmission system is composed of a transmitting and a receiving side. Both the transmitting and receiving sides are equipped with coils to transfer the power from the source to load. The battery with the receiving coil is installed on the quadcopter and the transmitting side is normally a ground station composed of a transmitting coil. For efficient power transmission, it is necessary that the quadcopter land on the ground station in such a way that the receiving and transmitting coils are aligned properly; however, due to the poor landing effect of quadcopters, there is always a chance of misalignment. This misalignment causes power loss and affects the efficiency of the charging system. To eliminate this misalignment issue, there is a need to develop a system which can easily mitigate the poor landing effect. For that, a charging system is proposed in this work which can cope with such issues and increase the power transmission efficiency. 1.2. Related Works Previously, there were different researches on wireless power transmission. In References [13,14], continuous inductive coupling in radio-frequency identification (RFID) tags were implemented. In References [15,16], a wireless power technique based on an inductive coupling mechanism was used for medical implants. Few wireless power systems for robotic applications were developed using a large array of smaller coils driven by microelectromechanical systems (MEMS) switches and organic field effect transistors to selectively transmit wireless power [17]. Some researchers developed efficient wireless power transmission systems by designing a switched-mode direct current/alternating current (DC/AC) inverter based on Class E topology [17–21]. Recently, there were many researches on power control in wireless power transmission. In References [22–25], the authors proposed a microcontroller-based power control method. In Reference [26], the authors developed a power control scheme using analog feedback circuits. Some researchers used the closed-loop power control technique using a commercial off-the-shelf (COTS) chipset [27–29]. For wireless charging of drones, some authors proposed laser beam systems which deliver the power directly to the drone [30]. Solar energy to support a drone’s long flight time was proposed in Reference [31]. Charging docking stations to recharge the drone battery were proposed in Reference [32]. Some authors applied smart contact arrays [33], whereas some proposed a drone charging station [34]. In Reference [35], the authors presented an approach based on optimal designing of the transmitting and receiving coil with the goal of becoming less sensitive to the misalignment of coils. The system was composed of a charging station with multiple arrays of primary or transmitting coils with a specifically designed secondary or receiving coil. The receiving coil was designed to perfectly fit in the landing skid of the drone. The authors proposed a transmitting coil overlapping scheme to entirely cover the charging area on the charging station. By calculating the impedance of the multiple Appl. Syst. Innov. 2018, 1, 44 3 of 19 transmitting coils and choosing the transmitting coil with maximum impedance, power transmission between the transmitting and receiving coil was achieved. In Reference [36], the authors presented a target detection technique based on image processing. After landing, the center of the coil was aligned with the transmitting coil using a specific color detection and image-processing scheme. The image-processing algorithm worked by taking the images using a drone camera and converted red/green.blue (RGB) color space to hue/saturation/value (HSV) color space. After applying some filters, the red color was detected and considered as a target. In References [37], the authors presented a positioning system using a binary distance laser sensor and ultrasonic sensors. The system was composed of a charging station comprising a single transmitting coil and a drone equipped with a single receiving coil. The system worked by detecting the position of the receiving coil and aligning the transmitting coil with it for wireless power transmission. The positioning system took almost 5 s to detect and align with the receiving coil. Using sensors, the system complexity increased; it required specific places and areas for installation on the drone and the charging station. 1.3. Contribution of the Paper and Road Map To solve the problem of misalignment of coils caused by the drone’s imperfect landing, a battery charging system based on wireless power transmission was developed in this work. The system is composed of a ground charging station and a receiving coil with the load. The charging station is equipped with multiple transmitting coils, whereas, on the receiver side, just one receiving coil is used. Multiple transmitting coils are placed on a movable bed which can move in four directions (positive X-direction, negative X-direction, positive Y-direction, and negative Y-direction). A control technique based on the hill climbing algorithm was implemented to control the alignment between the transmitting and receiving coil. The system was developed in a way allowing the drone to land freely on the charging station without the worry of alignment between the receiving and transmitting coil. The drone can land freely on charging station and the coils will adjust automatically in proper alignment to start the power transmission. Previously, the misalignment issues were solved using different methods, such as the design of a charging station with overlapped coils [35], using a target detection technique based on image processing [36], and using a positioning system based on physical sensors [37]. The proposed system is based on a novel method which uses the hill climbing algorithm on backscattered voltage signals of the transmitting coils. It improves the accuracy of the system and eliminates the flaws found in previous techniques such as the image-processing technique, where there are chances of missing the target. Also, it gives the freedom for the drone to land freely on a charging station and makes the system more adoptable by avoiding the use of physical sensors. The configuration of the proposed system is discussed in detail in Section 2. 2. Configuration of the Proposed Wireless Battery Charging System for a Quadcopter Figure 1 shows the proposed wireless battery charging system for a quadcopter. The system is composed of the following three parts: 1. 2. 3. Wireless power transmitter, comprising transmitting coil array which can move in four directions. Wireless power receiver, a receiving coil, and battery charger. Control unit which can measure the terminal voltage of each transmitting coil, and align the centroid of the transmitting and receiving coil using the hill climbing algorithm. Appl. Syst. Innov. 2018, 1, 44 Appl. Syst. Innov. 2018, 2, x FOR PEER REVIEW 4 of 19 4 of 19 Figure 1. 1. Block Blockdiagram diagramof ofthe the proposed proposed wireless wireless battery battery charging charging system system for for aa quadcopter. quadcopter. Figure The proposed wireless battery charging system consists of multiple transmitting coils which The proposed wireless battery charging system consists of multiple transmitting coils which can can be moved in four directions (positive X-direction, negative X-direction, positive Y-direction, and be moved in four directions (positive X-direction, negative X-direction, positive Y-direction, and negative Y-direction). The use of multiple coils can potentially allow the system to efficiently adapt negative Y-direction). The use of multiple coils can potentially allow the system to efficiently adapt to magnetic field propagation conditions, similar to the way multiple antennas are used to adapt to to magnetic field propagation conditions, similar to the way multiple antennas are used to adapt to channel conditions in wireless communication systems [38]. Also, multiple transmitting coils decrease channel conditions in wireless communication systems [38]. Also, multiple transmitting coils the time for coil alignment, providing a chance for the design of big charging stations for large drones, decrease the time for coil alignment, providing a chance for the design of big charging stations for where coil size and design are limited due to power transmission characteristics. In this proposed large drones, where coil size and design are limited due to power transmission characteristics. In this system, if a drone with a receiving coil lands at any position on the multiple transmitting coils, the proposed system, if a drone with a receiving coil lands at any position on the multiple transmitting voltages across the transmitting coils decrease depending on the coupling of each transmitting coil coils, the voltages across the transmitting coils decrease depending on the coupling of each with the receiver coil, and the controller knows where the drone lands by observing the change in transmitting coil with the receiver coil, and the controller knows where the drone lands by observing terminal voltages of the multiple coils. When the power is transferred from the transmitting coil to the the change in terminal voltages of the multiple coils. When the power is transferred from the receiving coil, the terminal voltage of the transmitting coil tends to decrease due to the phenomenon transmitting coil to the receiving coil, the terminal voltage of the transmitting coil tends to decrease called signal backscattering. observing the backscattered signal each transmitting coil,each the due to the phenomenon calledBy signal backscattering. By observing thefrom backscattered signal from controller automatically knows which transmitting coil is which nearest transmitting to the receiving the controller transmitting coil, the controller automatically knows coilcoil. is If nearest to the pinpoints the nearest transmitting coil, the multiple transmitting coil array is moved to automatically receiving coil. If the controller pinpoints the nearest transmitting coil, the multiple transmitting coil align the centroid the detected transmitting coil with thatdetected of the receiving coil using climbing array is moved toof automatically align the centroid of the transmitting coil the withhill that of the algorithm. The practical system used to implement the proposed system comprises a receiver circuit receiving coil using the hill climbing algorithm. The practical system used to implement the proposed for battery recharging and a circuit power for transmission station. and a power transmission station. system comprises a receiver battery recharging 2.1. Transmitter and Receiver Circuit Design and Description 2.1. Transmitter and Receiver Circuit Design and Description Figure 2 shows a simple wireless power transmission circuit. The circuit comprises a transmitter Figure 2 shows a simple wireless power transmission circuit. The circuit comprises a transmitter circuit, a receiving circuit, and the control part (voltage divider and analog-to-digital converter (ADC) circuit, a receiving circuit, and the control part (voltage divider and analog-to-digital converter (ADC) of the microcontroller). The power amplifier drives the transmitting coil Lt . When the receiving coil of the microcontroller). The power amplifier drives the transmitting coil 𝐿 . When the receiving coil is brought near to the transmitting coil, the voltage at receiving coil Lr is induced and it is processed is brought near to the transmitting coil, the voltage at receiving coil 𝐿 is induced and it is processed by a bridge rectifier to convert AC to DC voltage. To achieve the optimum performance, values of by a bridge rectifier to convert AC to DC voltage. To achieve the optimum performance, values of Ct , Lo , Co , and Cr are calculated using Equations (1) to (4). 𝐶 , 𝐿 , 𝐶 , and 𝐶 are calculated using Equations (1) to (4). Appl. Syst. Innov. 2018, 1, 44 Appl. Syst. Innov. 2018, 2, x FOR PEER REVIEW 5 of 19 5 of 19 Figure Figure 2. 2. Basic Basic circuit circuit diagram diagram for for wireless wireless power power transmission. transmission. C𝐶r == 11 R𝑅0 L𝐿r ± ± 2 ωM 𝜔𝑀2 2 ;; 2 2 R𝑅0 ω 𝜔 L𝐿r (1) (1) −1 ω 𝜔 C𝐶0 == ; 2 ) + R ( Q + 1 − sec( ϕ )) ; ωL 1 − K ( 𝜔𝐿t (1 − 𝐾 ) + 𝑅0 (𝑄 + − (𝜑))
2ω −1 ) (2𝜔 C𝐶t = ;; = π2 𝜋 11 + + 4 R𝑅 4 L0 = ω −1 QR0 . 𝐿 = 𝜔 𝑄𝑅 . (2) (2) (3) (3) (4) (4) In inductance, K In the the above above equations, equations, M 𝑀 is is the the mutual mutual inductance, 𝐾 is is the the coupling coupling factor, factor, Q 𝑄 isis the the quality quality ◦ to 70◦ . factor between the transmitting and receiving coils, and ϕ is the phase angle ranging from 40 factor between the transmitting and receiving coils, and 𝜑 is the phase angle ranging from 40° to 70° . The The derivation derivation of of the the equations equations for for aa Class-E Class-E amplifier amplifier can can be be found found in in Reference Reference [39]. [39]. Transmitting Transmitting coils the receiving receivingcoil coilare aremade madeofofcopper copper wire with 40 turns, respectively (Tables 1 coils and and the wire with 15 15 andand 40 turns, respectively (Tables 1 and and 2). According to the power transmission range capabilities, wireless power transmission can be 2). According to the power transmission range capabilities, wireless power transmission can be categorized categorized into into three three types types [40,41]. [40,41]. First First is is the the inductive inductive power power transmission transmission (IPT) (IPT) and and capacitive capacitive power transmission (CPT), used for short-range distances; second is the resonant inductive power transmission (CPT), used for short-range distances; second is the resonant inductive coupling coupling power widely used for medium-range distances; and third is the laser or microwave powertransmission, transmission, widely used for medium-range distances; and third is beam the laser beam or power transmission, used for long-range order toIn achieve high efficiency, inductive microwave power transmission, used for distances. long-range In distances. order to achieve high efficiency, coupling close coupling betweenbetween the transmitting and receiving coil. coil. Whereas, in inductiverequires couplingvery requires very close coupling the transmitting and receiving Whereas, resonant inductive coupling, efficient power transmission can be achieved with some distance between in resonant inductive coupling, efficient power transmission can be achieved with some distance the transmitting and receiving coil via the use ofthe resonant resonant coupling between the transmitting and receiving coil via use of circuits. resonantAlso, circuits. Also, inductive resonant inductive has better tolerance than inductive coupling. Therefore, the resonant inductive coupling is considered coupling has better tolerance than inductive coupling. Therefore, the resonant inductive coupling is an effective an technique coping with the coil misalignment issues, andissues, for drone charging considered effectivefor technique for coping with the coil misalignment and battery for drone battery systems. Therefore, in this work, a resonance inductive coupling-based wireless power transmission charging systems. Therefore, in this work, a resonance inductive coupling-based wireless power technique wastechnique used for was charging the charging drone battery. Thebattery. detailedThe block diagram of diagram the circuit transmission used for the drone detailed block of for the wireless power transmission is shown in Figure 3. At the transmitter side, Class-E power amplifiers circuit for wireless power transmission is shown in Figure 3. At the transmitter side, Class-E power are used to are generate AC voltages AC withvoltages a resonance of 240 kHz forof each amplifiers used amplified to generate amplified withfrequency a resonance frequency 240coil. kHz for each coil. Appl. Syst. Innov. Innov. 2018, 2018, 2, 1, x 44FOR PEER REVIEW Appl. Syst. of 19 19 66of Figure 3. 3. Block Block diagram diagram of of the the circuit circuit for for wireless wireless power power transmission transmission with with multiple multiple transmitting transmitting coils coils Figure and a receiving coil. and a receiving coil. Table 1. Excitation circuit component values. Table 1. Excitation circuit component values. Component Value Component Number of Turns Lt 𝐿 Lc 𝐿 Lo 𝐿 Ct 𝐶 Co 𝐶 R1 R2 𝑅 𝑅 Number of Turns Value 15 15 14.5µH µH 14.5 1 mH 1 mH 9.9 µH 9.9 µH 15 nF 15 nF 28 nF 28 nF 100 Ω 100 Ω 30 Ω 30 Ω Table 2. Receiver circuit component values. Valuevalues. Table Component 2. Receiver circuit component Number of Turns 40 Component Value L 733.39 r Number of Turns 40 µH Cr 680 pf 𝐿 733.39 µH CL 0.05 µF 𝐶 680 pf RLoad 2 kΩ 0.05 µF 𝐶 2 kΩ 𝑅 These amplifiers are driven by a high-frequency clock signal. The clock signal is provided to the gate driver of each amplifier of the respective transmitting coil simultaneously. A microcontroller is used for generating a high-frequency clock signal and it keeps measuring the output voltages of each transmitting coil using an ADC, and performs envelop detection of the measured voltages to identify which excitation coil is the closest to the the receiving receiving coil coil of of the the drone. drone. coilcoil is installed on the and itand is connected to a full At the the receiving receivingside, side,the thereceiving receiving is installed on drone the drone it is connected tobridge a full rectifier, which which is usedisto convert the AC DC. Finally, a battery charger (DC(DC to DC converter) is bridge rectifier, used to convert thetoAC to DC. Finally, a battery charger to DC converter) used for battery charging. The system works by detecting the change in the voltage of any of the is used for battery charging. The system works by detecting the change in the voltage envelope-detected voltage voltage signals signals of of the transmitting transmitting coils, coils, which which (changes (changes in voltage) refer to the envelope-detected presence of the receiving coil on on the the charging charging station. station. After detecting the receiving coil, a control algorithm is activated to align the coils and start wireless power power transmission. transmission. 2.2. Four-Way Four-Way Directional Directional XY Table 2.2. XY Table A four-way four-way directional directional XY table was to control control the coil A XY table was constructed constructed to the position position of of the the transmitting transmitting coil array, as shown in Figure 4. The XY table is controlled by two stepper motors that are driven by two array, as shown in Figure 4. The XY table is controlled by two stepper motors that are driven by two stepper motor Motor 11 is is responsible responsible for for moving moving the the transmitting transmitting coil Y-direction, stepper motor drivers. drivers. Motor coil array array in in the the Y-direction, while motor 2 is responsible for moving it in the X-direction. When the quadcopter approaches the Appl. Syst. Innov. 2018, 1, 44 Appl. Syst. Innov. 2018, 2, x FOR PEER REVIEW 7 of 19 7 of 19 while motor 2 is responsible for moving it in theexcited X-direction. the quadcopter approachesa charging station, the transmitting coil, already havingWhen a constant voltage, experiences the charging station, the transmitting coil, already excited having a constant voltage, experiences a change in voltage. This change in voltage refers to the presence of the receiving coil and load (battery). change in voltage. change in voltage to the presence of the and is load (battery). By measuring theThis voltage change from refers each transmitting coil, the receiving charging coil station moved in a By measuring theone voltage from coils each istransmitting coil, the charging station is moved a direction where of the change transmitting aligned to the receiving coil. The controller sendsinthe direction where one of the transmitting coils is aligned to the receiving coil. The controller sends the required pulse width modulation (PWM) signals to the stepper motor driver to move the transmitting required pulse width modulation (PWM) signals to the stepper driver to move the transmitting coil array to the proper position to initiate the wireless power motor transmission and battery charging. coil array to the proper position to initiate the wireless power transmission and battery charging. Figure 4. Four-way directional XY table. Figure 4. Four-way directional XY table. 2.3. Control Part 2.3. Control Part The control part of the proposed system performs specific tasks such as generating the clock signal The control part ofcoils, the proposed system performs specific tasksand suchcontrolling as generating the clock for driving transmitting monitoring each coil’s terminal voltage, the four-way signal for driving transmitting coils, monitoring each coil’s terminal voltage, and controlling the fourXY table to align the centroids of the transmitting and receiving coils. wayGenerally, XY table to centroids the transmitting receiving coils. station precisely. In most it align is notthe easy to land aofdrone at a specificand point on the drone Generally, it is not easy to land a drone at a specific point on the drone station in precisely. cases, the centroids of the transmitting and receiving coil are misaligned, as shown Figure 5.In most cases, the centroids of the transmitting coil are the misaligned, as shown Figure power 5. When the misalignment betweenand thereceiving coils happens, efficiency of the in wireless When the misalignment between the coils happens, the efficiency of the wireless power transmission deteriorates. In order to solve this problem, an intelligent automatic alignment algorithm transmission deteriorates. In order to solve this problem, an intelligent automatic alignment based on the hill climbing algorithm is proposed and implemented. algorithm based onalignment the hill climbing algorithm is proposed andaimplemented. The automatic algorithm was implemented inside microcontroller. The microcontroller The automatic alignment algorithm was implemented inside a microcontroller. The keeps measuring the terminal voltages of each transmitting coil simultaneously, and finds out which microcontroller keeps measuring the terminal voltages of each transmitting coil simultaneously, and transmitting coil has the lowest voltage. Generally, when the receiving coil of the drone is near to outtransmitting which transmitting has the lowest voltage. Generally, when the receiving of the afinds certain coil, the coil corresponding transmitting coil’s terminal voltage tends tocoil decrease. drone is near to a certain transmitting coil, the corresponding transmitting coil’s terminal voltage Therefore, the microcontroller can detect the transmitting coil with a voltage difference and, from this tends toon, decrease. Therefore, the microcontroller can detect transmitting a voltage moment the microcontroller tries to align the centroid of thethe transmitting coilcoil withwith the receiving difference and, from this moment on, the microcontroller tries to align the centroid of the transmitting coil by moving the four-way directional XY table. coil with the receiving coil by moving the four-way directional XY table. Appl. Syst. Innov. 2018, 2, x FOR PEER REVIEW Appl. Syst. Innov. 2018, 1, 44 Appl. Syst. Innov. 2018, 2, x FOR PEER REVIEW 8 of 19 8 of 19 8 of 19 Figure 5. 5. Misalignment of coils coils for for wireless wireless power power transmission. transmission. Figure Figure 5. Misalignment of coils for wireless power transmission. 2.3.1. 2.3.1. Signal Signal Backscattering Backscattering 2.3.1. Signal Backscattering Signal backscattering backscattering is is aa phenomenon phenomenon basically basically used used in in wireless wireless communication communication for for RFID RFID (radio-frequency identification). The transmitting side voltage tends to decrease by increasing the (radio-frequency identification). The transmitting side used voltage to decrease by increasing the Signal backscattering is a phenomenon basically in tends wireless communication for RFID coupling between the thetransmitting transmittingand and receiving coils. When the drone lands on the charging coupling between receiving coils. When the drone lands on the charging station, (radio-frequency identification). The transmitting side voltage tends to decrease by increasing the station, XYstarts tablethe starts the aligning process and, during theoftime of movement, the voltage the XY the table aligning process and, coils. during the the time voltage at the coupling between the transmitting and receiving When dronemovement, lands on thethe charging station, at the transmitting side decreases rapidly. This change in voltage produces a natural backscattered transmitting side decreases rapidly. This change in voltage produces a natural backscattered signal. the XY table starts the aligning process and, during the time of movement, the voltage at the signal. This backscattered signal at ahigh very high sampling frequency of 84isMHz is continuously This backscattered signal at a very of 84 MHz continuously readsignal. byread the transmitting side decreases rapidly. This sampling change in frequency voltage produces a natural backscattered by the ADC of the microcontroller. Figure 6 shows the sampling process of the voltage signal during ADCbackscattered of the microcontroller. 6 shows the sampling of the signalread during one This signal at a Figure very high sampling frequencyprocess of 84 MHz is voltage continuously by the one period of time for aligned (transmitting coil voltage atand load) and misaligned (transmitting coil period of time for aligned (transmitting coil voltage at load) misaligned (transmitting coil voltage ADC of the microcontroller. Figure 6 shows the sampling process of the voltage signal during one voltage at no load) coils.ofInside of the microcontroller, an envelope detection algorithm is usedthe to at no load) coils. Inside the microcontroller, an envelope detection algorithm is used to detect period of time for aligned (transmitting coil voltage at load) and misaligned (transmitting coil voltage detect the envelope of the backscattered signal. After the detection of an envelope, the peak value of envelope the backscattered signal. After the an detection of an envelope, the peakisvalue voltage at no load)ofcoils. Inside of the microcontroller, envelope detection algorithm used of to the detect the the voltage signal isand selected and this peak value is observed continuously to provide the information signal is of selected this peak value is the observed continuously to provide thevalue information about envelope the backscattered signal. After detection of an envelope, the peak of the voltage about aligned or misaligned coils. In the case of aligned coils, the peak value will be very low (almost aligned misaligned coils.peak In the case is of observed aligned coils, the peak value will bethe very low (almostabout 25 V; signal is or selected and this value continuously to provide information 25 V; Figure 6) in and, incase the of case of misalignment, the peak value will be high (almost 70 V; Figure 6). Figure 6) and, the misalignment, the peak value will be high (almost 70 V; Figure 6). aligned or misaligned coils. In the case of aligned coils, the peak value will be very low (almost 25The V; The climbing algorithm processes this voltage informationdata dataand andfinds findsthe the optimum solution solution by hill hill climbing this voltage by Figure 6) and, algorithm in the caseprocesses of misalignment, the information peak value will be high (almostoptimum 70 V; Figure 6). The moving the XY table in a specific direction and aligning the transmitting and receiving coils. moving the XY table in processes a specific direction andinformation aligning thedata transmitting hill climbing algorithm this voltage and findsand the receiving optimumcoils. solution by moving the XY table in a specific direction and aligning the transmitting and receiving coils. Figure 6. 6. Sampling Sampling of of voltage voltage signal signal during during one one time time period. period. Figure Figure 6. Sampling of voltage signal during one time period. Appl. Syst. Innov. 2018, 2, x FOR PEER REVIEW Appl. Syst. Innov. 2018, 1, 44 9 of 19 9 of 19 2.3.2. Hill Climbing Algorithm Hill climbing is anAlgorithm optimization technique that is used to find an optimum solution to a 2.3.2. Hill Climbing computational problem. It starts off with a solution that is normally very poor compared to the Hill climbing is an optimization technique thatItisdoes usedthis to by find an optimum optimal solution and then iteratively improves from there. generating other solution solutionsto a computational problem. It starts off with a solution that is normally very poor compared to the which are better than the current solution. It repeats the process until it finds the optimal solution optimal and then improves from there. It does this by generating other solutions where it can solution no longer find anyiteratively improvements. which are better than the current solution. repeats process until battery it finds charging the optimal solution Generally, drones tend to land at any placeItupon the the drone’s wireless station. where it can no longer find any improvements. That means the centroid of the receiving coil on the drone may not be aligned with any of the Generally, drones tend to land at anycoils placeare upon the drone’s wireless battery station. transmitting coils. In this case, transmitting moved using the XY table in charging an arbitrary That means the centroid of the receiving coil on the drone may not be aligned with any of the direction to get closer to the receiving coil. By measuring terminal voltages of each transmitting coil, transmitting coils. In this case, transmitting coils are moved using the XY table in an arbitrary direction the controller can detect whether the receiving coil gets closer to one of the transmitting coils. This closermovement to the receiving coil. Bycontinues measuring terminal voltages of each coil, the kindtoofget arbitrary of the XY table until it detects the decrease in transmitting terminal voltages controller can detect whether the receiving coil gets closer to one of the transmitting coils. This of the transmitting coils, as shown in Figure 7. When one transmitting coil, which correspondskind to of arbitrary coil movement of the 7, XYistable continues until it detects the decrease in terminal of the transmitting 2 in Figure chosen, the hill climbing algorithm is activated to voltages move the transmitting coils, as shown in Figure 7. When one transmitting coil, which corresponds to transmitting transmitting coil to the proper position where the voltage measured from that transmitting coil coil its 2 inminimum. Figure 7, is chosen, the hill climbing algorithm is activated to move the transmitting coil to the reaches proper the voltage measured that transmitting coil9reaches Figureposition 8 showswhere the overall flowchart of the from proposed method. Figure shows its theminimum. flowchart of Figure 8 shows the overall flowchart of the proposed method. Figure 9 shows the flowchart the hill climbing algorithm used in this system. Previously, authors tried to solve a different kinds of of the hill climbing algorithm used in this system. Previously, authors tried to solve a different kinds of control problems using the hill climbing algorithm [42,43]. In this work, the hill climbing algorithm control problems using thePWM hill climbing [42,43]. this work, thethe hillXY climbing algorithm starts by sending the required value toalgorithm the stepper motorIn driver to move table. At the starts by sending the required PWM value to the stepper motor driver to move the XY table. At the start, when there is no change in the voltage, i.e., the drone is not on the charging station, there is no start, when there is no change in the voltage, the droneWhen is notthe ondrone the charging station, there is no movement and the voltage of the transmitting coili.e., is constant. lands on the charging movement and the voltage of the transmitting coil is constant. When the drone lands on the charging station, there is a change in the voltage value, and the voltage decreases in the presence of a drone station, there is a change in the voltage value, and the voltage decreases in the presence of drone with a receiving coil. The hill climbing algorithm is activated and one transmitting coil with athe with a change receiving coil. The hill climbing algorithm is activated and one transmitting coil with maximum in the voltage value is selected. The current terminal voltage 𝑉 of the selected coil the maximum change in the voltage value is selected. The current terminal voltage V of the selected n is measured and compared with the previous measured voltage 𝑉 . The minimum value is stored coil is measured and compared with the previous measured voltage V . The minimum n − 1 as 𝑉 . After that, the controller sends the required PWM to move the XY table to the positionvalue of is stored as V . After that, the controller sends the required PWM to move the XY table to the position min the minimum value. This kind of process keeps repeating until it reaches the minimum voltage. In of the This kind of process keeps until it reaches the minimum voltage. Figure 9, 𝑛minimum = 1 to 4 isvalue. the number of transmitting coils, repeating and 𝑘 is the number of the sample. In Figure 9, n = 1 to 4 is the number of transmitting coils, and k is the number of the sample. Figure 7. (a) Receiving coil’s initial position. (b) Receiving coil gets close to a transmitting coil after Figure 7. (a) Receiving coil’s initial position. (b) Receiving coil gets close to a transmitting coil after arbitrary movement. arbitrary movement. Appl. Syst. Innov. 2018, 2, x FOR PEER REVIEW Appl. Syst. Innov. 2018, 1, 44 Appl. Syst. Innov. 2018, 2, x FOR PEER REVIEW Figure 8. Overall flowchart of the proposed method. Figure 8. Overall flowchart of the proposed method. Figure 8. Overall flowchart of the proposed method. Figure 9. Flowchart of the hill climbing algorithm. Figure 9. Flowchart of the hill climbing algorithm. Figure 9. Flowchart of the hill climbing algorithm. 10 of 19 10 of 19 10 of 19 Appl. Syst. Innov. 2018, 1, 44 Appl. Syst. Innov. 2018, 2, x FOR PEER REVIEW 11 of 19 11 of 19 3. Explanation Explanation of of the the Test Test Bench 3. Bench In order order to of the the proposed proposed algorithm, algorithm, we we made made aa test test bench bench of of aa wireless In to verify verify the the feasibility feasibility of wireless power transmission and battery charging station for the drone. The developed test bench is power transmission and battery charging station for the drone. The developed test bench is shown shown in Figure Figure 10. 10.The Thetest test bench was composed of four transmitting which are mounted the in bench was composed of four transmitting coilscoils which are mounted on theonfourfour-way XY table, one receiving coil connected to an electrical load, and a controller which performs way XY table, one receiving coil connected to an electrical load, and a controller which performs the the measurement climbing algorithm. measurement andand hill hill climbing algorithm. Figure 10. Practical Practical test test bench for drone battery charging station. 3.1. Battery Battery Charging Charging Station Station 3.1. The battery battery charging charging station station was was built The XY XY table table is is used used to to move move the the The built using using an an XY XY table. table. The transmitting coil coil array array to to the the position position where where the the centroid centroid of of the coil is is aligned aligned with with the the transmitting the transmitting transmitting coil centroid of the receiving coil using the hill climbing algorithm. The dimensions of the XY table were centroid of the receiving coil using the hill climbing algorithm. The dimensions of the XY table were 394 mm (l (𝑙 ×w ). The frame frame of the XY wastable madewas of aluminum 394mm mm××414 414mm mm××93.6 93.6 mm ×× 𝑤 h× ℎ ). main The main of table the XY made of and the rectangular plate, where the coils are placed, was made of a plastic sheet with a thickness aluminum and the rectangular plate, where the coils are placed, was made of a plastic sheet withof a 2.5 mm. Four transmitting coils were placed topplaced of the on rectangular within the XYwithin table and thickness of 2.5 mm. Four transmitting coils on were top of theplate rectangular plate the theytable wereand controlled by controlled two stepper for positioning. XY they were bymotors two stepper motors for positioning. In this work, four excitation circuits for the four transmitting coils werecoils developed. Each transmitting In this work, four excitation circuits for the four transmitting were developed. Each coil was connected to the excitation circuit. A gate voltage of 15 V and a supply voltage of 12 voltage V were transmitting coil was connected to the excitation circuit. A gate voltage of 15 V and a supply applied to theapplied IRF510to metal-oxide-semiconductor field-effect transistor (MOSFET), and it was driven of 12 V were the IRF510 metal-oxide-semiconductor field-effect transistor (MOSFET), and by a low-power 240 kHz clock signal, generated by the microcontroller. A voltage divider circuit was it was driven by a low-power 240 kHz clock signal, generated by the microcontroller. A voltage also builtcircuit to decrease the built output into voltage voltage that wasinto readable by the controller. Based by on the the divider was also tovoltage decrease thea output a voltage that was readable proposed method the equations in Reference [40], the optimumin values of the electrical controller. Based and on the proposed presented method and the equations presented Reference [40], the components for the excitation circuits were calculated. Table 1 shows the component in optimum values of the electrical components for the excitation circuits were calculated.values Table 1used shows the excitation circuit, and Figure a closer view and of theFigure excitation circuitsa and controller. the component values used in 11 theshows excitation circuit, 11 shows closer view of the excitation circuits and controller. Appl. Syst. Innov. 2018, 1, 44 12 of 19 Appl. Syst. Innov. 2018, 2, x FOR PEER REVIEW 12 of 19 Figure 11. Closer Closer view of the excitation circuits and controller. 3.2. Receiving Coil Coil and and Electric Electric Load 3.2. Receiving Load When When the the drone drone lands lands on on the the battery battery charging charging station station and and the the XY XY table table starts starts moving moving the the transmitting coil array arraytotothe theproper properposition position using climbing algorithm, transmitting transmitting coil using thethe hillhill climbing algorithm, the the transmitting coil coil starts to couple inductively with the receiving coil.inductive This inductive coupling coils starts to couple inductively with the receiving coil. This coupling betweenbetween the coilsthe induces induces voltage in the receiving coil. The receiving coil is connected to the receiving circuit as shown voltage in the receiving coil. The receiving coil is connected to the receiving circuit as shown in Figure in Figure 3. Thevoltage inducedatvoltage at Lr is processed by a full-wave bridge rectifier to convert AC In to this DC. 3. The induced 𝐿 is processed by a full-wave bridge rectifier to convert AC to DC. In this system, rectification is achieved via four insulated-gate bipolar transistors (IGBTs), designed system, rectification is achieved via four insulated-gate bipolar transistors (IGBTs), designed to work to on high-frequency AC input signals. The voltage output voltage is obtained across load resistor onwork high-frequency AC input signals. The output is obtained across the loadthe resistor 𝑅 . R . Table 2 shows the component values used in the receiving circuit, and Figure 12 shows a closer Load Table 2 shows the component values used in the receiving circuit, and Figure 12 shows a closer view view the receiving and receiving of theofreceiving circuitcircuit and receiving coil. coil. Figure 12. Closer view of the receiving circuit and receiving receiving coil. coil. 3.3. Controller 3.3. Controller In order to perform the control tasks, an STM32f4 discovery kit featuring a 32-bit ARM Cortex-M4 In order to perform the control tasks, an STM32f4 discovery kit featuring a 32-bit ARM Cortexwas used. It can generate up to a 168-MHz clock signal. It also supports ADC with 19 channels and M4 was used. It can generate up to a 168-MHz clock signal. It also supports ADC with 19 channels 12-bit resolution. The controller was used to generate 240-kHz clock signals for the four excitation and 12-bit resolution. The controller was used to generate 240-kHz clock signals for the four excitation circuits and it read the terminal voltage of transmitting coils simultaneously. circuits and it read the terminal voltage of transmitting coils simultaneously. When the drone lands on the battery charging station, the controller sends different PWM values When the drone lands on the battery charging station, the controller sends different PWM values to the two stepper motor drivers to move the XY table in random directions. At the same time, the to the two stepper motor drivers to move the XY table in random directions. At the same time, the controller the output voltage of of the coils. Once Once there there is is aa drop drop in controller keeps keeps measuring measuring the output voltage the four four transmitting transmitting coils. in the voltage across one of the transmitting coils, the controller triggers the hill climbing algorithm and sends the required PWM to the stepper motor drivers until it measures the minimum voltage value Appl. Syst. Innov. 2018, 1, 44 13 of 19 the across of the transmitting coils, the controller triggers the hill climbing algorithm and Appl.voltage Syst. Innov. 2018, one 2, x FOR PEER REVIEW 13 of 19 sends the required PWM to the stepper motor drivers until it measures the minimum voltage value transmitting and receiving coils are aligned; then, the hill climbing where both the centroids of the transmitting algorithm stops working, and the wireless power transmission begins until the the drone drone is is fully fully charged. charged. 4. Experiments and Results Several experiments are conducted conducted with a test bench in order to verify the feasibility of the proposed scheme. The The system system was was tested tested under under different test scenarios depending on the drone landing position on the charging station. The receiving coil attached with load was placed at different positions of of thethe system was observed. In one testtest scenario, the positions on onthe thecharging chargingstation stationand andthe theresponse response system was observed. In one scenario, system was tested for the misalignment of x = 100 mm and y = 50 mm, whereas in other test scenarios, the system was tested for the misalignment of x = 100 mm and y = 50 mm, whereas in other test the misalignment was x = 75was mm,x y= = mmyand = 50and mm, = 10 mm. of the position scenarios, the misalignment 7530 mm, = 30xmm x =y 50 mm, y =Regardless 10 mm. Regardless of the of the receiving coil, the charging wasstation able towas perfectly align the centroid of the transmitting position of the receiving coil, thestation charging able to perfectly align the centroid of the and receivingand coilreceiving with an accuracy Figure 13 shows one13ofshows the test scenarios. Initially, we transmitting coil with of an98.8%. accuracy of 98.8%. Figure one of the test scenarios. assumed thatassumed the receiving coil and fourcoil transmitting coils were positioned shown inasFigure Initially, we that the receiving and four transmitting coils were as positioned shown13a. in At the start, the XY table was moved randomly until the receiving coil got closer to one of the four Figure 13a. At the start, the XY table was moved randomly until the receiving coil got closer to one transmitting coils. During coils. this process, microcontroller kept measuring the terminal the voltages of the four transmitting Duringthe this process, the microcontroller keptfour measuring four simultaneously. Bysimultaneously. detecting the voltage drop between all four transmitting microcontroller terminal voltages By detecting the voltage drop between allcoils, fourthe transmitting coils, activated the hill climbing algorithm the XY table was moved to align the microcontroller activated the hill and climbing algorithm and the XY table the wasnearest movedtransmitting to align the coil withtransmitting a receiving coil coil.with The distance and position of the nearest transmitting coil were judged on nearest a receiving coil. The distance and position of the nearest transmitting the the voltage The thedrop. transmitting and coil, and the larger the coil, voltage coil basis were of judged on the drop. basis of thecloser voltage The closer thereceiving transmitting receiving the drop would be due to power transmission between the two coils. As shown in Figure 13b,c, after the larger the voltage drop would be due to power transmission between the two coils. As shown in random movement of the XY table, transmitting coiltable, 1 was found as the nearest coil to as thethe receiving Figures 13b,c, after the random movement of the XY transmitting coil 1 was found nearest coil. Then, the hill climbing algorithm was activated by the microcontroller and the XY table moved to coil to the receiving coil. Then, the hill climbing algorithm was activated by the microcontroller and align transmitting coil withtransmitting the receivingcoil coil. Figure shows the recorded waveforms for the XY table moved to 1align 1 with the14receiving coil. Figure voltage 14 shows the recorded the scenario shown infor Figure 13. It canshown be seeninclearly voltage of transmitting decreased, voltage waveforms the scenario Figurethat 13.the It can be seen clearly thatcoil the1 voltage of red in the case of Figure 13a, green in the case of Figure 13b, and blue in in the case of Figure 13c. transmitting coil 1 decreased, red in the case of Figure 13a, green in the case of Figure 13b, and blue Figure 15 shows voltage of Figure other transmitting coils under of load (wireless power transmission) and in in the case ofthe Figure 13c. 15 shows the voltage other transmitting coils under load no load (normal) condition. and no load (normal) condition. (wireless power transmission) Figure 13. (a) (a)Initial Initialposition positionofofthe thecoils. coils.(b)(b) table movement toward nearest transmitting Figure 13. XYXY table movement toward nearest transmitting coil.coil. (c) (c) Transmitting and receiving coil aligned for wireless power transmission. Transmitting and receiving coil aligned for wireless power transmission. Appl. Syst. Innov. 2018, 1, 44 Appl. Syst. Innov. 2018, 2, x FOR PEER REVIEW Appl. Syst. Innov. 2018, 2, x FOR PEER REVIEW 14 of 19 14 of 19 14 of 19 Figure coil 11 voltage under wireless power transmission. (For scenario Figure Figure 14. 14.Transmitting Transmitting 1 voltage under wireless transmission. case in scenario Figure 14. Transmitting coilcoil voltage under wireless powerpower transmission. (For case case(For scenario in Figure 13). in Figure 13). 13). Figure under load (normal) and load (wireless power transmission) Figure 15. 15. Voltage under no for transmitting Figure 15. Voltage Voltage under no no load load (normal) (normal) and and load load (wireless (wireless power power transmission) transmission) for for transmitting transmitting coil 2. coil 2. 2. coil The 𝑃 = W. was designed designed for for the the transmission transmission power power of of P 60 output The system system was was designed for the transmission power of 𝑃tx = = 60 60 W. W. Figure Figure 16 16 shows shows the the output output under different test scenarios. distance between power at the load resistance 𝑅 power R under different test scenarios. The misalignment power at the load resistance 𝑅L under different test scenarios. The misalignment distance between the observed. The maximum load power 𝑃 the coils coils was waschanged changedand andthe theload loadpower power𝑃 responsewas was observed. The maximum load power the coils was changed and the load power 𝑃PLresponse response was observed. The maximum load power 𝑃 was measured to be 52 W, which gives an efficiency η of 85%. It can be observed in Figure 16 that, P was measured to be 52 W, which gives an efficiency η of 85%. It can be observed in Figure 16 L was measured to be 52 W, which gives an efficiency η of 85%. It can be observed in Figure 16 that, as the distance of between the and coils the that, the distance of misalignment between the transmitting and receiving coils increased, thetaken time as theas distance of misalignment misalignment between the transmitting transmitting and receiving receiving coils increased, increased, the time time taken by the charging station (XY properly align the possible coil increased. In taken the charging (XY to table) to properly to the nearest possible coil also increased. by theby charging stationstation (XY table) table) to properly align to toalign the nearest nearest possible coil also also increased. In the the case of a misalignment of x = 100 mm, y = 50 mm, it took almost 1.85 s to properly align the coils In the case of a misalignment of x = 100 mm, y = 50 mm, it took almost 1.85 s to properly align the case of a misalignment of x = 100 mm, y = 50 mm, it took almost 1.85 s to properly align the coils and and transfer power with designed For of 30 mm xx ==mm 50 coils andthe transfer power with efficiency. designed efficiency. For misalignments of xyy===75 = 30 transfer the powerthe with designed efficiency. For misalignments misalignments of xx == 75 75 mm, mm, 30mm, mm yand and 50 mm, y = 10 the time to align the coils was almost 1.58 s and 1.5 s, respectively. In a time ranging and x = 50 mm, y = 10 mm, the time to align the coils was almost 1.58 s and 1.5 s, respectively. In a mm, y = 10 mm, the time to align the coils was almost 1.58 s and 1.5 s, respectively. In a time ranging between 1.5 s and 1.9 s, the misalignment caused by the imperfect drone landing was eliminated and time ranging between 1.5 s and 1.9 s, the misalignment caused by the imperfect drone landing was between 1.5 s and 1.9 s, the misalignment caused by the imperfect drone landing was eliminated and the power was increased the maximum eliminated and the power transmission the maximum power. the power transmission transmission was increased to towas theincreased maximumtopower. power. In In order algorithm, obtained In order to to validate validate the the hill hill climbing climbing algorithm, algorithm, the the data data obtained obtained from from the the practical practical test test bench bench was used in MATLAB and simulations were carried out to assure the feasibility of the algorithm. simulations were carried out to assure the feasibility of the algorithm. was used in MATLAB and simulations were carried out to assure the feasibility of the algorithm. Figure Figure 17 trajectory Figure 17 shows shows the the trajectory trajectory of of the the XY XY position position for for one one transmitting transmitting coil coil in in aa three-dimensional three-dimensional space. It can can be be seen seen that, that, at at the the start, start, there there is is no no voltage voltage drop drop and and the XY table table is is moved moved to to aaa random random space. It It can be seen that, at the start, there is no voltage drop and the XY XY table is moved to random position. After two random movements, the position with the minimum possible voltage value position. After two random movements, the position with the minimum possible voltage value position. After two random movements, the position with the minimum possible voltage value is is selected and, from that position, the hill climbing algorithm is activated and it keeps finding the best from finding selected and, from that position, the hill climbing algorithm is activated and it keeps finding the best possible possible positions positions to to align align the the transmitting transmitting and and receiving receiving coils coils until until the the minimum minimum level level is is achieved. achieved. Appl. Syst. Innov. 2018, 1, 44 Appl. Syst. Innov. 2018, 2, x FOR PEER REVIEW Appl. Syst. Innov. 2018, 2, x FOR PEER REVIEW 15 of 19 15 of 19 15 of 19 possible to level alignisthe transmitting and receiving coils until the minimum levelproperly is achieved. Once thepositions minimum achieved, the transmitting and receiving coils are aligned and Once the the minimum minimum level level is is achieved, achieved, the the transmitting transmitting and and receiving receiving coils coils are are aligned aligned properly properly and and Once wireless power transmission starts with full efficiency. wireless power powertransmission transmissionstarts startswith withfull fullefficiency. efficiency. wireless Figure 16. Load power response for different misalignment cases. Figure Figure 16. 16. Load power response response for for different different misalignment misalignment cases. cases. The results obtained obtained from the the test bench bench shows that that the proposed proposed system is is quite efficient efficient in The The results results obtained from from the test test bench shows shows that the the proposed system system is quite quite efficient in in resolving the misalignment issue. The power transmission efficiency of 85% is reasonable for resolving issue. TheThe power transmission efficiency of 85%of is reasonable for resonant resolvingthe themisalignment misalignment issue. power transmission efficiency 85% is reasonable for resonant inductive-based wireless power transmission, while also reducing the need for the inductive-based wireless power transmission, while also reducing the need for thethe implementation resonant inductive-based wireless power transmission, while also reducing need for the implementation of complex tracking and landing algorithms for the drone. The time to align the of complex tracking and landing algorithms for the drone. The to align thetime centroid is also implementation of complex tracking and landing algorithms for time the drone. The to align the centroid is also minimized due to the use of the hill climbing algorithm. The drone is free to land minimized due minimized to the use of thetohill algorithm. Thealgorithm. drone is free land is anywhere on centroid is also due theclimbing use of the hill climbing Thetodrone free to land anywhere on charging station and, within a time of just 2 s, the centroid of the coils are aligned charging and, within a time just 2 as, time the centroid are aligned the anywherestation on charging station and,ofwithin of just 2ofs,the thecoils centroid of the properly coils are and aligned properly and the power transmission loss is mitigated. power transmission loss is mitigated. properly and the power transmission loss is mitigated. Figure Figure 17. 17. Trajectories Trajectoriesof ofhill hillclimbing climbingalgorithm. algorithm. Figure 17. Trajectories of hill climbing algorithm. Comparative ComparativeStudy Study Comparative Study A A comparative comparative study study of of the the proposed proposed and andpreviously previously presented presented solutions solutions for for drone droneimperfect imperfect A comparative study of is the proposed and previously presented solutions for drone imperfect landing (misalignment issues) presented in this section. The system is compared with three solutions landing (misalignment issues) is presented in this section. The system is compared with three landing (misalignment issues) is presented in this section. The system is compared with three presented in Reference solutions presented in [37–39]. Reference [37–39]. solutions presented in Reference [37–39]. In In Reference Reference [37], [37], the the authors authors presented presented an an approach approach based based on on the the optimal optimal design design of of the the In Reference [37], the authors presented an approach based on the optimal design of the transmitting and receiving coils with the goal of becoming less sensitive to the misalignment of coils. transmitting and receiving coils with the goal of becoming less sensitive to the misalignment of coils. transmitting and receiving of coils with thestation goal of becoming less sensitive to the misalignment of coils. The Thesystem systemwas wascomposed composed ofaacharging charging stationwith withmultiple multiplearrays arraysof ofprimary primaryor ortransmitting transmittingcoils coils The system was composed of a charging station with multiple arrays ofcoil primary or transmitting coils with a specifically designed secondary or receiving coil. The receiving was designed to perfectly with a specifically designed secondary or receiving coil. The receiving coil was designed to perfectly with a specifically designed secondary or receiving coil. The receiving coil was designed toscheme perfectly fit fitin inthe thelanding landingskid skidof ofthe thedrone. drone.The Theauthors authorsproposed proposedaatransmitting transmittingcoil coiloverlapping overlapping scheme to to fit in thecover landing skid of the drone. The authorsstation. proposed acalculating transmitting coil overlapping scheme to entirely entirely coverthe thecharging chargingarea areaon onthe thecharging charging station.By By calculatingthe theimpedance impedanceof ofthe themultiple multiple entirely cover the charging area on the charging station. By calculating the impedance oftransmission the multiple transmitting transmittingcoils coilsand andchoosing choosingthe thetransmitting transmittingcoil coilwith with maximum maximum impedance, impedance, power power transmission transmitting coils and choosing the transmitting coil with maximum impedance, power transmission between betweenthe thetransmitting transmittingand andreceiving receivingcoils coilswas wasachieved. achieved. between the transmitting and receiving coils was achieved. This system works well for a specific type of drone with some specific dimensions; however, if This system works well for a specific type of drone with some specific dimensions; however, if the size and dimension of the drone (physical size) changes, a whole new design of the charging the size and dimension of the drone (physical size) changes, a whole new design of the charging Appl. Syst. Innov. 2018, 1, 44 16 of 19 This system works well for a specific type of drone with some specific dimensions; however, if the size and dimension of the drone (physical size) changes, a whole new design of the charging station is required, and the transmitting coils will need to be readjusted again with some proper overlapping to cover the charging area properly. On the other hand, our proposed system has non-overlapped multiple transmitting coils which can be moved in four directions. Furthermore, the proposed system utilizes the feature of backscattering which generally happens in wireless communication in order to identify the most closely coupled receiving coil among them. Even though the closely coupled receiving coil is detected, a further fine alignment of the centroids of the transmitting coil and the receiving coil is required to obtain efficient wireless power transmission by moving multiple transmitting coils. The proposed system is autonomous and totally free from the physical dimensions of the drone; it just needs to detect a receiving coil. This coil could be placed at any part of the drone where it can come in contact with any of the multiple transmitting coils placed on the charging station. The drone can land at any part of the charging station and, within just 1 to 2 s (min to max, depending on the distance of misalignment), the coils would be properly aligned with a strong coupling factor. In Reference [33], the impedance of the transmitting coils was monitored, which requires some current and voltage sensors to provide the measured value at every instance of charging operation. It increases the complexity of the charging station and requires more electronic components. On the other hand, in our case, the voltages of the transmitting coils are enough to ensure proper alignment between coils. Regarding power transfer capabilities and efficiency, in Reference [37], the authors presented results based on the efficiency of wireless power transmission. The results showed that the efficiency of the system changed and decreased when there was misalignment between the coils. The efficiency met the minimum level requirement of 75% in all cases; however, upon misalignment, there was a reduction in efficiency from 88% (normal; no misalignment) to 83% (100-mm misalignment) and 78% (200-mm misalignment). In our system, the efficiency will always be the same (normal; no misalignment) because there is no misalignment and power transmission is done after ensuring this. In Reference [38], the authors presented a target detection technique based on image processing. After landing, the center of the coil was aligned with the transmitting coil using some specific color detection and image-processing scheme. The image-processing algorithm worked by taking the images using a drone camera and converted RGB color space to HSV color space. After applying some filters, the red color was detected and considered as a target. There is always a chance of a failure in such a scheme contingent upon the outer environment and weather conditions; research work is still being pursued to improve such image-processing techniques. In our work, the drone is independent of such outer environmental uncertainties and the system works autonomously with the use of the hill climbing algorithm. In Reference [39], the authors presented a positioning system using a binary distance laser sensor and ultrasonic sensors. The system was composed of a charging station comprising a single transmitting coil and a drone equipped with a single receiving coil. The system worked by detecting the position of the receiving coil and aligning the transmitting coil with it for wireless power transmission. The positioning system took almost 5 s to detect and align with the receiving coil. Using sensors, the system complexity increases; it requires some specific areas for installation of the drone and the charging station. There is the possibility of errors in installing these sensors properly at the proper position so that the alignment between coils is always perfect. In our system, there is no physical sensor that needs to be installed on the drone or charging station. The time it takes to get the best possible solution is 1 to 2 s (min to max, depending on the distance of misalignment). The algorithm used in our system is much faster and much more robust, giving accurate results. Moreover, in Reference [35], there is just a single transmitting coil used to transfer the power. In our case, the charging station is composed of multiple transmitting coils because using multiple transmitting coils gives the possibility of designing large charging stations for big drones. If the drone size is big, it will require more space for landing on the charging station. Using multiple transmitting coils decreases the time required Appl. Syst. Innov. 2018, 1, 44 17 of 19 to align the coils. Generally, the drone lands at any point on the charging station. In Reference [35], it was made sure that the drone lands in the range of the sensors to initiate the alignment process; however, it would be almost impossible to initiate this process of alignment if the drone landed on the charging station and was out of range of the sensors. For this, in our charging station, multiple arrays of transmitting coils were installed to allow the drone to land freely anywhere on the charging station without the possibility of getting out of range. 5. Conclusions In this work, an efficient wireless power transmission system for drone battery charging was developed. A charging station with multiple transmission coils was used to transfer power to the receiving side (drone) to charge the battery. This study aimed to solve the problem caused by uneven drone landing on a charging station which leads to inefficient wireless power transmission due to the poor alignment between the transmitting and receiving coils. To solve this problem, a control mechanism based on the hill climbing algorithm was proposed. The control mechanism was used to mitigate the uneven landing effect of the drone on a charging station by carefully moving the charging station to a point where wireless power transmission was maximum. A practical test bench was developed to test the system feasibility. The power transmission efficiency was 85% and the test accuracy of the system was 98.8%. It can be observed from the results that the proposed system performed well compared to previously used techniques, without any need of a physical sensor such as a position sensor. Also, the use of the hill climbing algorithm gives the system a much faster response compared to any image-based target detection scheme. It also eliminates the possibility of getting affected by environmental conditions. The drone simply needs to land on the charging station, and, within 2 s, wireless power transmission starts at its maximum designed efficiency without any misalignment. Future Work We are working on the real-time implementation of the system. For that, there are several things which need to be improved and addressed carefully. These include the battery charging issues such as charging time and the capacity of the battery. Additionally, the wireless power transmission system will be improved in the future. The efficiency of the system can be improved by looking into some design parameters. The distance between the transmitting and receiving coils will be improved and medium-range transmission will be introduced into the system. Author Contributions: Conceptualization, A.R.; Data curation, M.R.; Formal analysis, M.R.; Investigation, A.R., M.T. and S.-H.K.; Project administration, A.R.; Resources, M.T.; Software, A.R., M.T.; Supervision, S.-H.K.; Validation, A.R. and S.-H.K.; Visualization, S.-H.K.; Writing–review & editing, A.R. Funding: This research received no external funding. Conflicts of Interest: The authors declare no conflicts of interest. References 1. 2. 3. 4. Wang, G.; Liu, W.; Sivaprakasam, M.; Humayun, M.; Weiland, J. Power supply topologies for biphasic stimulation in inductively powered implants. In Proceedings of the IEEE International Symposium on Circuits and Systems (ISCAS), Kobe, Japan, 23–26 May 2005; Volume 3, pp. 2743–2746. Hoffmann, G.M.; Rajnarayan, D.G.; Waslander, S.L.; Dostal, D.; Jang, J.S.; Tomlin, C.J. The Stanford Testbed of Autonomous Rotorcraft for Multi-Agent Control (STARMAC). In Proceedings of the 23rd Digital Avionics System Conference, Salt Lake City, UT, USA, 28 October 2004. Available online: https://www.icao.int/Meetings/UAS/Documents/Circular%20328_en.pdf (accessed on 22 May 2018). Stafford, J. How a Quadcopter Works | Clay Allen; University of Alaska: Fairbanks, AK, USA, 2014. Appl. Syst. Innov. 2018, 1, 44 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 18 of 19 Luukkonen, T. Modelling and Control of Quadcopter; Independent Research Project in Applied Mathematics; Aalto University: Helsinki, Finnland, 2011. Lange, S.; Sunderhauf, N.; Protzel, P. A vision based onboard approach for landing and position control of an autonomous multirotor UAV in GPS-denied environments. In Proceedings of the International Conference on Advanced Robotics, Munich, Germany, 22–26 June 2009; pp. 1–6. Bhaskaranand, M.; Gibson, J.D. Low-complexity video encoding for UAV reconnaissance and surveillance. In Proceedings of the IEEE Military Communications Conference (MILCOM), Baltimore, MD, USA, 7–10 November 2011; pp. 1633–1638. Doherty, P.; Rudol, P. A UAV search and rescue scenario with human body detection and geolocalization. In Advances in Artificial Intelligence; Springer: Berlin/Heidelberg, Germany, 2007; pp. 1–13. Tomic, T.; Schmid, K.; Lutz, P.; Domel, A.; Kassecker, M.; Mair, E.; Grixa, I.L.; Ruess, F.; Suppa, M.; Burschka, D. Toward a fully autonomous UAV: Research platform for indoor and outdoor urban search and rescue. IEEE Robot. Autom. Mag. 2012, 19, 46–56. [CrossRef] DIJ. Available online: https://www.dji.com/ (accessed on 15 June 2018). Mostafa, T.M.; Muharam, A.; Hattori, R. Wireless battery charging system for drones via capacitive power transfer. In Proceedings of the IEEE PELS Workshop on Emerging Technologies: Wireless Power Transfer, Chongqing, China, 20–22 May 2017; pp. 1–6. Campi, T.; Cruciani, S.; Feliziani, M.; Maradei, F. High efficiency and lightweight wireless charging system for drone batteries. In Proceedings of the AEIT International Annual Conference, Cagliari, Italy, 20–22 September 2017; pp. 1–6. Shoki, H. Issues and Initiatives for Practical Deployment of Wireless Power Transfer Technologies in Japan. Proc. IEEE 2013, 101, 1312–1320. [CrossRef] Finkenzeller, K. RFID Handbook: Fundamentals and Applications in Contactless Smart Cards and Identification; Wiley: Hoboken, NJ, USA, 2003. Jiang, B.; Smith, J.R.; Philipose, M.; Roy, S.; Rajan, K.S.; Mamishev, A.V. Energy scavenging for inductively coupled passive rfid systems. IEEE Trans. Instrum. Meas. 2007, 56, 118–125. [CrossRef] Li, P.; Principe, J.; Bashirullah, R. A wireless power interface for rechargeable battery operated neural recording implants. In Proceedings of the 28th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBS), New York, NY, USA, 30 August–3 September 2006; pp. 6253–6256. Aldhaher, S.; Yates, D.C.; Mitcheson, P.D. Design and development of a Class EF2 inverter and rectifier for multi megahertz wireless power transfer systems. IEEE Trans. Power Electron. 2016, 31, 8138–8150. [CrossRef] Sekitani, T.; Takamiya, M.; Noguchi, Y.; Nakano, S.; Kato, Y.; Hizu, K.; Kawaguchi, H.; Sakurai, T.; Someya, T. A large-area flexible wireless power transmission sheet using printed plastic mems switches and organic field-effect transistors. In Proceedings of the International Electron Devices Meeting (IEDM), San Francisco, CA, USA, 11–13 December 2006; pp. 1–4. Aldhaher, S.; Mitcheson, P.D.; Yates, D.C. Load-independent Class EF inverters for inductive wireless power transfer. In Proceedings of the IEEE Wireless Power Transfer Conference (WPTC), Aveiro, Portugal, 5–6 May 2016; pp. 1–4. Aldhaher, S.; Luk, P.C.-K.; Whidborne, J.F. Tuning Class E inverters applied in inductive links using saturable reactors. IEEE Trans. Power Electron. 2014, 29, 2969–2978. [CrossRef] Choi, J.; Tsukiyama, D.; Tsuruda, Y.; Rivas, J. 13.56 MHz 1.3 kW resonant converter with GaN FET for wireless power transfer. In Proceedings of the IEEE Wireless Power Transfer Conference (WPTC), Boulder, CO, USA, 13–15 May 2015; pp. 1–4. Pinuela, M.; Yates, D.C.; Lucyszyn, S.; Mitcheson, P.D. Maximizing dc-to-load efficiency for inductive power transfer. IEEE Trans. Power Electron. 2013, 28, 2437–2447. [CrossRef] Low, Z.N.; Casanova, J.J.; Maier, P.H.; Lin, J. Method of load/fault detection for loosely coupled planar wireless power system with power delivery tracking. IEEE Trans. Ind. Electron. 2010, 57, 1478–1486. Zhong, W.X.; Hui, S.Y.R. Maximum energy efficiency tracking for wireless power transfer systems. IEEE Trans. Power Electron. 2015, 30, 4025–4034. [CrossRef] Fu, M.; Yin, H.; Zhu, X.; Ma, C. Analysis and tracking of optimal load in wireless power transfer systems. IEEE Trans. Power Electron. 2015, 30, 3952–3963. [CrossRef] Appl. Syst. Innov. 2018, 1, 44 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 19 of 19 Yin, J.; Lin, D.; Lee, C.K.; Hui, S.Y.R. A systematic approach for load monitoring and power control in wireless power transfer systems without any direct output measurement. IEEE Trans. Power Electron. 2015, 30, 1657–1667. [CrossRef] Jang, B.J.; Lee, S.; Yoon, H. HF-band wireless power transfer system: Concept, issues, and design. Prog. Electromagn. Res. 2012, 124, 211–231. [CrossRef] Si, P.; Hu, A.P.; Hsu, J.W.; Chiang, M.; Wang, Y.; Malpas, S.; Budgett, D. Wireless power supply for implantable biomedical device based on primary input voltage regulation. In Proceedings of the 2nd IEEE Conference on Industrial Electronics and Applications, Harbin, China, 23–25 May 2007; pp. 235–239. Low, Z.N.; Chinga, R.A.; Tseng, R.; Lin, J. Design and test of a high-power high-efficiency loosely coupled planar wireless power transfer system. IEEE Trans. Ind. Electron. 2008, 56, 1801–1812. Available online: http://lasermotive.com/category/uavs/VideoofLaser-PoweredQuadrocopterEnduranceFlight (accessed on 20 July 2018). Available online: http://www.koreaittimes.com/story/6601/unmanned-aerialvehiclesdeveloped-solelykoreantechnologies (accessed on 20 July 2018). Uri, K.; Stern, H.; Edan, Y.; Feied, C.; Handler, J.; Smith, M.; Gillam, M. Vision-based autonomous robot self-docking and recharging. In Proceedings of the IEEE World Automation Congress, Budapest, Hungary, 24–26 July 2006; pp. 1–8. Fetisov, V.; Dmitriyev, O.; Neugodnikova, L.; Bersenyov, S.; Sakayev, I. Continuous monitoring of terrestrial objects by means of duty group of multicopters. In Proceedings of the XX IMEKO World Congress, Busan, Korea, 9–10 September 2012. Heeseo, C.; Park, J.; Song, H.; Kim, Y.; Jeong, H. The IoT based automate landing system of a drone for the round-the-clock surveillance solution. In Proceedings of the IEEE International Conference on Advanced Intelligent Mechatronics (AIM), Busan, Korea, 7–11 July 2015; pp. 1575–1580. Campi, T.; Cruciani, S.; Feliziani, M. Wireless Power Transfer Technology Applied to an Autonomous Electric UAV with a Small Secondary Coil. Energies 2018, 11, 352. [CrossRef] Junaid, A.B.; Konoiko, A.; Zweiri, Y.; Sahinkaya, M.N.; Seneviratne, L. Autonomous Wireless Self-Charging for Multi-Rotor Unmanned Aerial Vehicles. Energies 2017, 10, 803. [CrossRef] Choi, C.H.; Jang, H.J.; Lim, S.G.; Lim, H.C.; Cho, S.H.; Gaponov, I. Automatic wireless drone charging station creating essential environment for continuous drone operation. In Proceedings of the 2016 International Conference on Control, Automation and Information Sciences (ICCAIS), Ansan, Korea, 27–29 October 2016; pp. 132–136. Arakawa, T.; Goguri, S.; Krogmeier, J.V.; Kruger, A.; Love, D.J.; Mudumbai, R.; Swabey, M.A. Optimizing Wireless Power Transfer from Multi Transmit Coils. IEEE Access 2018. [CrossRef] Casanova, J.J.; Low, Z.N.; Lin, J. Design and optimization of a Class-E amplifier for a loosely coupled planar wireless power system. IEEE Trans. Circuits Syst. II Express Briefs 2009, 56, 830–834. [CrossRef] Bolaji, A.L.; Femi, B.A.; Shola, P.B. Hill Climbing Algorithm for Solving Patient Admission Scheduling Problem. Knowl.-Based Syst. 2018. [CrossRef] Merino, L.; Caballero, F.; Martínez-de Dios, J.R.; Ferruz, J.; Ollero, A. A cooperative perception system for multiple UAVs: Application to automatic detection of forest fires. J. Field Robot. 2006, 23, 165–184. [CrossRef] Choudhary, V.; Singh, S.P.; Kumar, V.; Prashar, D. Wireless Power Transmission: An Innovative Idea. Int. J. Educ. Plan. Adm. 2011, 1, 203–210. Bao, D.; Gu, J.; Di, Z.; Zhang, T. Optimization of Airport Shuttle Bus Routes Based on Travel Time Reliability. Math. Prob. Eng. 2018. [CrossRef] © 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).