An aerial unmanned aerial vehicle-to-unmanned aerial vehicle wireless energy transmission method based on resonant magnetic coupling technology

Abstract

The present application relates to a kind of aerial unmanned aerial vehicle between wireless energy transmission method based on resonant magnetic coupling technology.First, energy supply unmanned aerial vehicle reaches the power receiving unmanned aerial vehicle above, using the constant voltage output characteristic of LCC-S compensation network carries out forward analysis and residual error optimization, instantaneously space mutual inductance is analyzed reversely and real-time space coupling coefficient is extracted.Subsequently, the coefficient is compared with start threshold, if not aligned, then start the positioning calibration function of multi-coil polling excitation, correct to alignment.Again after alignment, under the constraint of anti-intercoupling heating, with the maximum effective magnetic flux as the goal, solve optimal excitation complex vector, and construct high-density magnetic focusing channel at the power receiving end to transfer energy.Finally, the load voltage of the power receiving end is recursively filtered and denoised by moving average, and the effective transmission efficiency is calculated in real time, so as to monitor and confirm that it is maintained in the optimal transmission state.This method solves the problems of difficult aerial dynamic alignment, poor transmission method robustness and unstable transmission efficiency, and can realize stable energy transmission.

Description

Technical Field

[0001] This invention belongs to the field of wireless charging for unmanned aerial vehicles (UAVs), specifically relating to a method for wireless energy transfer between UAVs in the air based on resonant magnetic coupling technology. Background Technology

[0002] With the rapid development of the global low-altitude economy, in long-endurance scenarios such as drone disaster emergency rescue, the flight time of mainstream multi-rotor drones is generally short due to the current battery energy density. They need to frequently return to change batteries or perform contact charging, which limits the operating radius of drones. This makes "limited flight time" a bottleneck problem that restricts the further large-scale application of drone technology.

[0003] To address the aforementioned range anxiety issue, existing solutions primarily focus on the deployment of ground-based wireless charging stations. While patent application number "202210564456.5," titled "A Hovering Wireless Charging Device for Drones Based on Event-Triggered Control," achieves unmanned charging for drones, it is essentially still a form of "static charging." The drone must interrupt its current mission and fly to a designated fixed base station, which not only fails to meet the requirements for continuous all-weather cruising but also suffers from the limited coverage of ground base stations, making flexible deployment in complex or wide-area environments difficult.

[0004] Some existing aerial charging solutions attempt to utilize hovering technology for in-flight resupply. Application number "202311829550.X" discloses "A Magnetic Coupled Resonant Wireless Charging System for UAVs Based on Arrayed Coils," which effectively improves the system's anti-misalignment capability by using a method combining estimated mutual inductance coefficient M with arrayed coils. However, in real-world environments, it is susceptible to interference, resulting in significant fluctuations in energy transfer efficiency, making it difficult to meet the requirements for high-precision and high-stability operations.

[0005] While existing technologies have made significant explorations at the hardware system level, the underlying alignment and control methods they rely on still have significant limitations. The common problems of existing methods are as follows: First, in dynamic aerial environments, dynamic alignment is extremely difficult, and there is a lack of effective passive electrical "blind alignment" and active three-dimensional positioning mechanisms. Once a serious physical deviation occurs, it is difficult to automatically correct and recover. Second, there is a lack of dynamic feature extraction and closed-loop control strategies that can effectively cope with airflow turbulence, resulting in large fluctuations in energy transmission efficiency and insufficient robustness of control methods. Finally, most charging control methods still passively rely on ground-based fixed base stations, which cannot achieve "continuous dynamic aerial recharging," severely restricting the ability of UAVs to perform long-distance, high-stability missions. Summary of the Invention

[0006] This invention addresses the problems of alignment difficulties, poor robustness of transmission methods, and unstable transmission efficiency in existing technologies by proposing a wireless energy transmission method between unmanned aerial vehicles (UAVs) based on resonant magnetic coupling technology.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a method for wireless power transfer between unmanned aerial vehicles (UAVs) based on resonant magnetic coupling technology, comprising the following steps:

[0008] Step 1: Spatial State Analysis and Initial Coupling Assessment: The power supply UAV flies above the power receiver UAV and applies an initial excitation signal. The power receiver responds to the excitation and transmits relevant parameters back to the load end. The power supply uses the electrical response characteristics of the resonant network to perform forward analytical derivation of the theoretical voltage response, and combines it with the actual observed voltage to perform residual optimization, and inversely analyzes to obtain the optimal estimate of the instantaneous spatial mutual inductance. Finally, it performs normalization calculation to extract the real-time spatial coupling coefficient, providing a data benchmark for subsequent alignment state determination.

[0009] Step 2, Positioning calibration of multi-coil polling excitation: Set a coupling coefficient threshold based on the minimum mutual inductance value when the receiving coil has the maximum offset, and compare the extracted real-time spatial coupling coefficient with this threshold to determine the alignment status:

[0010] If aligned and the alignment conditions are met, proceed to step three.

[0011] If misalignment occurs, the power supply UAV uses the spherical launch array as an equivalent multi-port magnetic network to initiate the positioning calibration function of multi-coil polling excitation; it constructs a two-dimensional magnetic coupling correlation matrix by collecting the voltage returned from the power receiving end; combined with the finite element spatial mapping model, it accurately calculates the three-dimensional spatial offset coordinates of the power receiving end; and then converts them into flight control commands to drive the airframe to perform physical position correction until the alignment conditions are met or the maximum number of retries is reached.

[0012] Step 3: Spatial Magnetic Field Vector Synthesis and Focusing: After meeting the alignment conditions, the power supply control unit establishes a mathematical model of magnetic field superposition based on a multi-port magnetic network. Taking the maximization of the effective magnetic flux passing through the plane of the receiving coil under this model as the global optimization objective, and under the premise of strictly satisfying the self-compatible anti-mutual coupling heating constraint inside the array, the optimal excitation complex vector is solved. Finally, by executing this vector, amplitude and phase joint control is performed to complete the magnetic field vector synthesis in real physical space and construct a high-density magnetic focusing channel.

[0013] Step 4, Dynamic Filtering and Optimal State Monitoring: During energy transmission, the power supply control unit performs recursive moving average filtering to denoise the load voltage fed back from the power receiver, and combines it with the real-time calculated effective transmission efficiency to confirm that it maintains the efficient and optimal transmission state in the magnetic focusing channel. Under the premise of this efficient transmission, the full charge state of the power receiver battery is determined by monitoring the duration of the load output power in the minimum threshold range, and then the power output is automatically cut off to achieve a safe closed loop for the charging task.

[0014] Furthermore, the process of forward analytically deriving the theoretical voltage response based on the electrical response characteristics of the resonant network in step one above is as follows: extract the effective value of the transmitter output voltage, and define the transmitter-side compensation inductance as the forward transfer function. When the energy receiving end has an unknown instantaneous equivalent mutual inductance At that time, its predicted theoretical one-dimensional load DC voltage output response for:

[0015] (1)

[0016] (2)

[0017] in: The theoretical load-side DC voltage response value predicted at time t; It is a forward transfer function operator determined by the underlying hardware parameters; Let t be the instantaneous theoretical equivalent mutual inductance between the active port of the transmitting array and the receiving coil at time t.

[0018] Furthermore, the process of optimizing the residuals in step one above, combined with the actual observed voltage, and inversely analyzing to obtain the optimal estimate of the instantaneous spatial mutual inductance is as follows: The objective function based on minimizing the residuals... To perform extreme value optimization, define the mutual inductance analytic operator. By performing algebraic reconstruction on the one-dimensional voltage measure, the optimal unbiased estimate of the current instantaneous spatial mutual inductance is analytically obtained. :

[0019] (3)

[0020] (4)

[0021] (5)

[0022] in: The objective function is the sum of squared residuals constructed based on the actual observed voltage and the theoretically predicted voltage. For the defined mutual inductance analytic operator; This represents the effective value of the output voltage at the transmitting end. Compensating inductor for the transmitting side; The actual load-side voltage transmitted back from the energy-receiving end at time t; This is the optimal unbiased estimate of the instantaneous spatial mutual inductance at the current moment.

[0023] Furthermore, the specific process of normalizing the above step one to extract the real-time spatial coupling coefficient is as follows: constructing a dimensionless feature projection operator. By applying the projection operator to the analytically obtained state estimate, and performing scale compression and feature normalization mapping on it, the dynamic invariant that purely characterizes the current spatial magnetic field coupling depth—the real-time spatial coupling coefficient—is extracted. :

[0024] (6)

[0025] (7)

[0026] in: The real-time spatial coupling coefficient representing the alignment depth of the real physical space is extracted at time t. It is a dimensionless feature projection operator for the self-inductance parameters of the built-in transceiver coil.

[0027] Furthermore, in step two above, the constructed two-dimensional magnetic coupling correlation matrix characterizing the spatial coupling strength distribution is as follows:

[0028] (8)

[0029] Among them, matrix elements This represents the mutual inductance between the transmitting coil and the receiving coil located in the r-th row and c-th column of the spherical array.

[0030] Furthermore, in step two above, when calculating the three-dimensional spatial offset coordinates of the energy receiving end, the energy supply end performs anomaly removal on the feature data, substitutes the elements of the correlation matrix into the finite element mapping model, and constructs the following weighted objective function based on minimizing matrix mutual inductance error. :

[0031] (9)

[0032] in: The confidence weighting coefficients assigned to valid reference magnetic ports; The three-dimensional spatial coordinates of the center of the receiving coil of the powered UAV are to be solved. This is the theoretical spatial mutual inductance value between the reference magnetic ports; This represents the actual measured instantaneous equivalent mutual inductance between the reference magnetic port in row r and column c and the receiving coil.

[0033] Furthermore, in step three above, the magnetic field superposition mathematical model is used to perform array vector superposition:

[0034] Since the spatial positions and normal vectors of each coil unit on the spherical array are different, the control unit transforms the magnetic field vectors generated by all N units to a unified coordinate system for linear superposition; the total composite magnetic field at the receiving end is expressed as:

[0035] (11)

[0036] Where: is the geometric magnetic field factor generated by the i-th unit under unit current excitation, and is the excitation weighting coefficient of the unit.

[0037] Furthermore, in step three above, the effective magnetic flux passing through the plane of the receiving coil is maximized. Represented as:

[0038] (12)

[0039] Set a hard boundary constraint on the overall transmit power: Due to the limitations of the onboard energy storage and inverter capacity of the power supply UAV, the total input power of each magnetic port must be limited within the safe operating range.

[0040] (13)

[0041] Setting array self-compatible anti-coupling constraints: Strong spatial cross-inductance interference exists during concurrent excitation of high-density magnetic ports. To avoid eddy current heating and beam distortion caused by severe coupling within the transmitting array, the total interference power received by the i-th magnetic port is set based on predictive analysis equations. The temperature must be strictly below the preset device sensitivity heating threshold. :

[0042] (14)

[0043] in: Due to total power limitations, Let be the effective value of the current in the i-th transmitting coil. The AC resistance of the coil is... The total interference power received by the i-th transmitting magnetic port from the cross-coupling superposition of other active ports at the same frequency within the array; This represents the actual output excitation power of the j-th transmitting magnetic port; This is the internal magnetic coupling transfer function extracted from the magnetic coupling correlation matrix, which couples from the j-th transmitting magnetic port to the i-th transmitting magnetic port. The maximum cross-interference heating power and electromagnetic susceptibility threshold that a single magnetic port can withstand.

[0044] Furthermore, in step four above, the process of performing recursive moving average filtering for noise reduction is as follows:

[0045] Construct a first-in-first-out buffer queue of length N, and perform a recursive moving average filter on the load voltage to obtain the smoothed and denoised load terminal voltage:

[0046] (15)

[0047] in: Let be the smoothed load-end voltage output after recursive moving average filtering at time k. The smoothed load terminal voltage history value is derived at time k-1. The original sampled value of the instantaneous DC voltage at the load terminal is obtained at the k-th sampling time; N is the sliding window length of the first-in-first-out data buffer queue. The oldest historical instantaneous voltage value that slides out of the tail of the data buffer queue.

[0048] Furthermore, in step four above, the effective energy transfer efficiency of the current control unit at the k-th sampling time is calculated in real time. :

[0049] (16)

[0050] (17)

[0051] in: This represents the total power of the transmitting end; This represents the effective value of the output voltage at the transmitting end. This represents the effective value of the output current at the transmitting end. This represents the load impedance at the receiving end.

[0052] Compared with the prior art, the beneficial effects of the present invention are:

[0053] (1) To address the lack of sensorless electrical alignment mechanisms in existing technologies, this method uses the characteristics of resonant networks as a physical premise and establishes a mathematical analytical relationship between voltage and mutual inductance based on known circuit response characteristics. The power supply control unit only needs the underlying one-dimensional voltage parameters to algebraically reconstruct the real-time spatial coupling coefficient, thus eliminating the dependence on external spatial sensors and providing an objective data benchmark for subsequent positioning and correction.

[0054] (2) To address the challenges of initial alignment difficulties and recovery after misalignment in dynamic aerial environments, this method initiates a positioning calibration function using multi-coil polling excitation when misalignment is detected. By constructing a two-dimensional magnetic coupling correlation matrix combined with a finite element mapping model, the three-dimensional spatial offset coordinates of the receiving end are accurately calculated, and the aircraft is actively guided to perform translational correction, thus endowing the transmission process with active optimization and dynamic correction capabilities.

[0055] (3) To address the issues of limited anti-offset capability of array coils and the tendency for internal cross-interference and heating during concurrent operation, this method is based on a magnetic field superposition mathematical model. Under the boundary constraints such as array self-compatibility and anti-mutual coupling heating, the optimal excitation complex vector is solved for amplitude and phase control. Even if the UAV tilts or translates, the power supply control unit can still reshape the magnetic focusing channel and maintain the stability of directional energy transmission.

[0056] (4) To address the problem of efficiency evaluation distortion caused by jitter in real-world environments, which in turn leads to control oscillations, this method utilizes a first-in-first-out buffer queue and recursive moving average filtering technology to eliminate data spikes. Based on the smoothed and denoised voltage characteristics, the system accurately calculates the real-time effective transmission efficiency, realizing closed-loop monitoring of the dynamic magnetic focusing link, and precisely verifying and ensuring that the system always maintains an optimal transmission state with high efficiency and stability.

[0057] (5) This invention deeply integrates magnetic port networks, spatial positioning algorithms, and magnetic field vector control technology through a complete technical logic of "alignment perception - positioning calibration - magnetic field focusing - dynamic monitoring". The synergistic effect of multiple methods is as follows: reusing the same magnetic port network for both spatial perception and energy focusing (hardware synergy); directly converting positioning coordinates into phase compensation for vector control (data synergy); and using the concurrent heating threshold as a forced boundary for vector optimization (safety synergy). The three construct a closed-loop power transfer architecture that resists dynamic offset, objectively solving the pain points of UAVs in dynamic and limited aerial environments, such as difficulty in alignment, weak resistance to pose offset, and fluctuations in transmission efficiency, and effectively improving the overall stability of aerial wireless power transfer methods. Attached Figure Description

[0058] Figure 1 The overall flowchart of aerial UAV alignment and transmission based on resonant magnetic coupling;

[0059] Figure 2 This is a schematic diagram of a spherical transmitting coil array structure;

[0060] Figure 3 A schematic diagram of positioning calibration using multi-coil polling; Detailed Implementation

[0061] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0062] See Figure 1The basic idea of ​​this invention is as follows: First, the powered UAV hovers and applies initial excitation. The theoretical voltage response is derived analytically using the characteristics of the resonant network, and residual optimization is performed by combining the actual observed voltage. Inverse analysis is then used to determine the instantaneous spatial mutual inductance, and the real-time spatial coupling coefficient is extracted through normalization to complete the initial evaluation. Next, this coefficient is compared with the start-up threshold. If misalignment is found, a multi-coil time-sequential polling excitation is initiated to construct a two-dimensional magnetic coupling correlation matrix. The three-dimensional offset coordinates are calculated, and the UAV position is translated and corrected until alignment is achieved. Then, after alignment is achieved, a multi-port array magnetic field superposition mathematical model is established. Constraints are set with maximizing the effective magnetic flux as the objective, and the optimal excitation complex vector is solved. A high-density magnetic focusing channel is constructed through amplitude and phase modulation. Finally, during charging, a recursive moving average filtering process is performed on the load voltage. The effective energy transfer efficiency is calculated in real time, and the full-charge shutdown state is determined, achieving dynamic closed-loop monitoring throughout the entire process.

[0063] Based on the above basic idea, this invention provides a method for wireless power transfer between aerial unmanned aerial vehicles (UAVs) based on resonant magnetic coupling technology, the specific steps of which are as follows:

[0064] Step 1: Spatial State Resolution and Initial Coupling Assessment

[0065] The power supply drone flies above the power receiving drone and applies an initial excitation signal. The power receiving end responds to the excitation and transmits relevant parameters back to the load end. The power supply end uses the electrical response characteristics of the resonant network to perform forward analytical derivation of the theoretical voltage response, and combines it with the actual observed voltage to perform residual optimization, and inversely analyzes to obtain the optimal estimate of the instantaneous spatial mutual inductance. Finally, it performs normalization calculation to extract the real-time spatial coupling coefficient, providing a data benchmark for subsequent alignment state determination. Specifically, the process includes the following sub-steps:

[0066] 1.1 The power supply drone flies to the location, hovers, and applies initial excitation.

[0067] After the charging command is issued, the power supply drone automatically flies to the airspace above the receiving drone and hovers, adjusting its attitude. During this stage, the vertical distance remains within the effective coupling radius, allowing for a certain degree of positional deviation (i.e., coarse alignment). Subsequently, the power supply drone deploys its wireless charging pod, completes internal self-checks, and establishes a communication link; after confirming that the status is normal, it controls the inverter to apply a preset low-power initial excitation signal. At this time, the spherical transmitting array synthesizes an initial guiding magnetic field in space, pointing towards the approximate direction of the receiving end, thereby avoiding excessive electromagnetic radiation or heat loss in the case of inaccurate alignment.

[0068] 1.2 After receiving the initial excitation from step 1.1, the receiving end collects relevant parameters from the load end and transmits them back to the supply end. The supply end performs a forward analysis of the load voltage and spatial mutual inductance to derive the theoretical voltage response:

[0069] Based on this, the power supply end first extracts the effective value of the transmitter's output voltage. Transmitter-side compensation inductor Define it as a positive transfer function. When the energy receiving end has an unknown instantaneous equivalent mutual inductance. At that time, its predicted theoretical one-dimensional load DC voltage output response It can be obtained from the state variables through the mapping of this operator:

[0070] (1)

[0071] (2)

[0072] in: The theoretical load-side DC voltage response value predicted at time t; It is a positive transfer function operator; Let t be the instantaneous theoretical equivalent mutual inductance between the active port of the transmitting array and the receiving coil at time t.

[0073] 1.3 Combining the theoretical voltage response and the actual observed voltage from step 1.2, residual optimization is performed to analytically obtain the optimal estimate of the instantaneous spatial mutual inductance:

[0074] In actual operation, the power supply end acquires the actual discrete load end voltage observation sequence transmitted back by the power receiving end in real time through a low-latency wireless link at a fixed sampling period. To accurately isolate spatial mutual inductance characteristics, the control unit compares the theoretical voltage response derived in step 1.2 with the actual observed voltage using residuals, and internally constructs an estimation objective function based on minimizing the residuals. :

[0075] (3)

[0076] in: The objective function is the sum of squared residuals constructed based on the actual observed voltage and the theoretically predicted voltage. Let be the instantaneous spatial mutual inductance estimation variable to be solved at time t; Let t be the actual load-side voltage transmitted back from the receiving end.

[0077] By performing extremum optimization on the above objective function (i.e., setting...) Define the reverse state observation operator. This observation operator directly reconstructs the algebraic space of the one-dimensional voltage measure, analytically outputting the optimal unbiased estimate of the instantaneous spatial mutual inductance at the current moment. :

[0078] (4)

[0079] (5)

[0080] in: To define the reverse state observation operator; This represents the effective value of the output voltage at the transmitting end. Compensating inductor for the transmitting side; This is the optimal unbiased estimate of the instantaneous spatial mutual inductance at the current moment.

[0081] This method enables the reverse analysis from a single voltage signal to the physical quantity of spatial coupling strength.

[0082] 1.4 Normalize the optimal estimate of the instantaneous spatial mutual inductance obtained in step 1.3 and extract the real-time spatial coupling coefficient:

[0083] To eliminate the absolute dimensional shift introduced by differences in the geometric and physical dimensions of the coils carried by different UAV models and to ensure the cross-platform universality of the decision logic, a dimensionless feature projection operator is constructed in the power supply control unit. By applying this operator to the inverted state estimate By normalizing the self-inductance parameters of the power supply and receiving coils, the dynamic invariant that purely characterizes the current spatial magnetic field coupling degree at both ends of the power supply and receiving coils—the real-time spatial coupling coefficient—is extracted. :

[0084] (6)

[0085] (7)

[0086] in: The real-time spatial coupling coefficient, representing the alignment of the real physical space, is extracted at time t. A dimensionless feature projection operator for the self-inductance parameters of the built-in transceiver coil; This is the optimal unbiased estimate of the instantaneous spatial mutual inductance calculated in the preceding steps; This is the actual load-side voltage transmitted back from the receiving end at time t. This coefficient serves as the core decision data for determining the alignment state, completing the initial evaluation closed loop in step one.

[0087] Step 2: Positioning calibration of multi-coil polling excitation:

[0088] The coupling coefficient threshold is set based on the minimum mutual inductance value when the receiving coil is at its maximum offset, and the extracted real-time spatial coupling coefficient is compared with this threshold to determine the alignment status.

[0089] If aligned and the alignment conditions are met, proceed to step three.

[0090] If misalignment occurs, the power supply UAV uses the spherical launch array as an equivalent multi-port magnetic network to initiate the positioning calibration function of multi-coil polling excitation; it constructs a two-dimensional magnetic coupling correlation matrix by collecting the voltage returned from the power receiving end; combined with the finite element spatial mapping model, it accurately calculates the three-dimensional spatial offset coordinates of the power receiving end; and then converts them into flight control commands to drive the airframe to perform physical position correction until the alignment conditions are met or the maximum number of retries is reached.

[0091] See Figure 3 , Figure 3 "Voltage response feedback" refers to the process by which the receiving end transmits the voltage value it measures back to the supplying end. Whether this feedback process continues depends on whether the current operating phase is an alignment phase or an misalignment phase.

[0092] This step is the misalignment "polling calibration" phase, which is intermittent. The receiving end will only send back voltage data once when the power supply tests each coil individually, in order to calculate the positional deviation.

[0093] The misaligned "polling calibration" phase specifically includes the following sub-steps:

[0094] 2.1 The power supply control unit compares the real-time coupling coefficient with the safe start-up threshold, and autonomously decides whether to perform pose fine-tuning or full-power magnetic field focusing based on the judgment result:

[0095] Obtain the coupling coefficient between the current drones Then, based on the minimum coupling coefficient when the receiving coil is in the maximum offset state, the minimum start-up threshold is set. Subsequently, the power supply control unit compares the current actual coupling coefficient with the threshold. A comparison is performed to determine whether the charging terminals of the power-supplying drone and the power-receiving drone are aligned:

[0096] like (Alignment valid) indicates that the coarse alignment error is within the electronic focusing compensation range of the spherical array. The power supply control unit generates a "coupling valid" flag and directly starts step three.

[0097] like (Misalignment) indicates that the physical offset is too large, exceeding the limit of electronic focusing. The multi-coil polling positioning calibration function is activated, and a "position correction request" is sent to the flight control module.

[0098] 2.2 Multi-coil timing polling activation:

[0099] After determining misalignment, the power supply control unit defines the spherical transmitting array equivalently as a multi-port magnetic network. (See also...) Figure 2 The spherical emission array uses a radius of A hemispherical shell structure, with its surface evenly distributed Each is a separate transmitting coil. Centered on the sphere... Establish a local three-dimensional coordinate system with the origin as the reference point, and define the centers of each transmitting coil. All are fixed to the sphere, and the normal vector of the central axis Strictly pointing towards the center of the sphere. Based on the spatial topology of the multi-port network, the central region of the array and its adjacent N transmitting coils are dynamically selected as positioning reference magnetic ports. The power supply end sequentially applies low-power high-frequency detection signals to each of the N positioning reference magnetic ports according to a set timing period, performing polling excitation of each reference magnetic port to obtain the voltage response characteristics of the power receiving end.

[0100] 2.3 Constructing a two-dimensional magnetic coupling correlation matrix:

[0101] During the time slot when each reference magnetic port is independently activated, the receiving coil of the powered UAV induces a corresponding voltage signal. N sets of corresponding load-side DC voltage data are then acquired sequentially. , , , The instantaneous equivalent mutual inductance between the converted receiving coil and each activated magnetic port Subsequently, based on the physical topological arrangement of the spherical transmitting array, the power supply control unit performs a two-dimensional mapping of the discrete mutual inductance values ​​according to the relative topological positions of the transmitting coils in real space, constructing a two-dimensional magnetic coupling correlation matrix characterizing the spatial coupling strength distribution. :

[0102] (8)

[0103] Among them, matrix elements This represents the mutual inductance between the transmitting coil and the receiving coil located in the r-th row and c-th column of the spherical array (the element is padded with zeros if there is no active coil at this position). This matrix reflects the "spatial distribution characteristics of magnetic field coupling" formed above the power supply array at the receiving end. Simultaneously, the power supply control unit extracts the absolute spatial geometric center coordinates of each reference magnetic port preset in the local memory to construct the corresponding three-dimensional reference coordinate matrix. .

[0104] 2.4 The power supply end performs anomaly removal on the feature data, substitutes the elements of the correlation matrix into the finite element mapping model, and iteratively optimizes the objective function using the least squares algorithm to solve for the actual three-dimensional spatial offset coordinates of the center of the receiving coil of the powered UAV:

[0105] To overcome single-point data distortion caused by transient electromagnetic noise in dynamic environments, the power supply first modulates the two-dimensional magnetic coupling correlation matrix. Spatial continuity smoothing and outlier removal are performed, and confidence weights are assigned to valid ports. Subsequently, the control unit substitutes the elements of the actual inverted instantaneous mutual inductance matrix into the "spatial coupling-position mapping model" to construct the following weighted objective function based on minimizing the matrix mutual inductance error. :

[0106] (9)

[0107] in: The confidence weighting coefficient is assigned to the valid reference magnetic port located in row r and column c, which is used to reduce the interference of sudden environmental noise on the solution accuracy after outlier removal;

[0108] Based on a pre-defined finite element "spatial coupling-position mapping model", when the center coordinates of the receiving coil are... At that time, the theoretical spatial mutual inductance value between it and the reference magnetic port in the r-th row and c-th column;

[0109] The instantaneous equivalent mutual inductance between the reference magnetic port in the r-th row and c-th column and the receiving coil is the actual measured value obtained by inverting the voltage at the load end of the powered UAV.

[0110] To apply a nonlinear least squares algorithm to the objective function After iterative optimization, the global optimal solution that minimizes the weighted sum of squared residuals is obtained, which is the actual three-dimensional spatial offset coordinate of the center of the receiving coil of the powered UAV.

[0111] Power supply control unit combined with coordinate matrix The objective function is obtained by applying the least squares algorithm. Perform iterative optimization to find the global optimal solution that minimizes the weighted sum of squared residuals. This solution mechanism deeply integrates the topology of the spatial array with prior data from finite element simulation, improving the accuracy and stability of dynamic three-dimensional positioning in complex electromagnetic environments.

[0112] 2.5 Translational correction of the machine body position:

[0113] The power supply drone receives the horizontal offset coordinates calculated above. This is then converted into flight control commands to drive the aircraft to make physical position corrections. Flight control commands refer to flight attitude adjustment commands, such as roll angle. With pitch angle The fine-tuning amount. The powered drone executes this displacement command, translating and correcting its hovering position to reduce the offset.

[0114] Once the correction is complete, the process automatically returns to step 2.1 and restarts the coupling evaluation:

[0115] If the coupling coefficient If the location is successful, proceed to step three;

[0116] If the conditions are still not met, steps 2.2-2.5 are repeated iteratively until the threshold is met or the maximum number of retries is reached, at which point the process jumps to step three. In this embodiment, the maximum number of retries is 5 to prevent the system from getting stuck in the positioning bottleneck due to continuous retries, thereby excessively consuming the drone's power resources.

[0117] Step 3: Spatial magnetic field vector synthesis and focusing:

[0118] After the alignment conditions are met, the control unit uses a mathematical model for spatial magnetic field vector synthesis based on a multi-port magnetic network to maximize the effective magnetic flux passing through the plane of the receiving coil under this model as the global optimization objective. Under the premise of strictly satisfying the self-compatible anti-mutual coupling heating constraint inside the array, the optimal excitation complex vector is solved. Finally, by executing this vector, amplitude and phase joint control is performed to complete the magnetic field vector synthesis in real physical space, and a high-density magnetic focusing channel is constructed at the energy receiving end to achieve efficient and stable energy transmission.

[0119] 3.1 Establishing a mathematical model for magnetic field superposition based on a multi-port magnetic network:

[0120] To achieve precise magnetic field focusing, the control unit first establishes a magnetic field superposition mathematical model, which is a vector superposition analytical model used to describe the linear mapping relationship between the excitation current of each magnetic port in the spherical transmitting array and the total magnetic induction intensity vector at the target field point of the receiving end. This model provides a physical topological basis for subsequent optimal excitation optimization by extending the magnetic field distribution generated by a single coil to the vector superposition of a three-dimensional spherical array in a unified coordinate system.

[0121] Single-coil model: For any i-th transmitting coil unit in the spherical array, let its excitation current be... According to the Biot-Savart law, the magnetic field strength vector generated by this unit at any point in space (i.e., the center of the receiving coil of the powered UAV) is... for:

[0122] (10)

[0123] in, The permeability of free space, For coil integration line elements, This is the distance vector from the source point to the field point.

[0124] Array vector superposition: Due to the different spatial positions and normal vectors of each coil unit on the spherical array, the control unit transforms the magnetic field vectors generated by all N units into a unified coordinate system for linear superposition. The total resultant magnetic field at the receiving end... Represented as:

[0125] (11)

[0126] in: Let be the geometric magnetic field factor generated by the i-th element under unit current excitation. This is the excitation weighting coefficient for this unit.

[0127] 3.2 Set constraints with the goal of maximizing effective magnetic flux:

[0128] Based on the mathematical model of magnetic field superposition, the control unit defines a global decision variable vector containing the amplitude and phase of all transmitting coils to solve a constrained optimization problem with the goal of maximizing the effective magnetic flux, and introduces total power limit and array self-compatible anti-coupling constraint.

[0129] To maximize transmission efficiency, the optimization objective of this control unit is not to pursue the scalar peak value of the field strength at a single point in space, but to maximize the effective magnetic flux passing through the plane of the receiving coil. The control unit, based on the magnetic field superposition model established in the previous steps, defines a global decision variable vector containing the amplitude and phase of all transmitting coils. The following multi-parameter nonlinear constraint optimization is constructed:

[0130] (1) Define the global extremum objective function:

[0131] The core control objective of this transmission method is not to maximize the scalar field strength at a single point in space, but to maximize the effective magnetic flux passing through the plane of the receiving coil.

[0132] (12)

[0133] in: For the effective area of ​​the receiving coil, The normal vector of the receiving coil. Let be the magnetic field strength generated by the i-th unit under current excitation.

[0134] (2) Set hard boundary constraints for global transmit power:

[0135] Due to the limitations of the onboard energy storage and inverter capacity of the power supply drone, the total input power of each magnetic port must be limited to a safe operating range:

[0136] (13)

[0137] in: Due to total power limitations, Let be the effective value of the current in the i-th transmitting coil. is the AC resistance of the coil.

[0138] (3) Set array self-compatible anti-coupling constraints:

[0139] When high-density magnetic ports are concurrently excited, strong spatial cross-inductance interference exists. To avoid eddy current heating and beam distortion caused by severe coupling inside the transmitter array, the total interference power received by the i-th magnetic port is set based on the predictive analysis equation. The temperature must be strictly below the preset device sensitivity heating threshold. :

[0140] (14)

[0141] in: The total interference power received by the i-th transmitting magnetic port from the cross-coupling superposition of other active ports at the same frequency within the array; This represents the actual output excitation power of the j-th transmitting magnetic port; This is the internal magnetic coupling transfer function extracted from the magnetic coupling correlation matrix, which couples from the j-th transmitting magnetic port to the i-th transmitting magnetic port. The maximum cross-interference heating power and electromagnetic susceptibility threshold that a single magnetic port can withstand.

[0142] 3.3 Optimization of the optimal excitation complex vector:

[0143] By solving this equation, the control unit calculates the optimal current excitation vector required for each coil unit. Physical meaning: In this process, the physical essence of adjusting the current phase is not to compensate for the wave propagation delay (unlike far-field radar), but to cancel the difference in the directional components of the magnetic field vector generated by coils at different spatial locations at the receiving point.

[0144] 3.4 Amplitude and phase modulation to construct a high-density magnetic focusing channel:

[0145] The control unit converts the solved optimal excitation complex vector into hardware instructions, driving the underlying inverter to execute them. Phase rotation ensures that all... In the normal direction of the receiving coil The projection components on the surface are of the same sign (i.e., they are superimposed in the same direction), thereby forming a high-density magnetic focusing channel in the near field region, ensuring that the maximum induced voltage can be obtained even when the receiving coil is tilted.

[0146] Step 4: Dynamic Filtering and Optimal State Monitoring

[0147] During energy transfer, the power supply control unit performs recursive moving average filtering and noise reduction on the load voltage fed back from the power receiver, and combines this with real-time calculated effective transfer efficiency to ensure it remains in the optimal high-efficiency transfer state within the magnetic focusing channel. Under this premise of efficient transfer, the full charge status of the power receiver battery is determined by monitoring the duration of the load output power within the minimum threshold range, and then the power output is automatically cut off, achieving a safe closed-loop charging process. This step specifically includes the following sub-steps:

[0148] 4.1 Perform recursive moving average filtering on the load voltage:

[0149] A first-in-first-out buffer queue of length N is constructed, and the load voltage is filtered and denoised using a recursive moving average. The effective transmission efficiency is calculated in real time to confirm that it is in the optimal transmission state.

[0150] The power supply drone receives the load-side voltage from the power receiving drone in real time via a wireless communication link. To prevent data glitches from the underlying hardware from penetrating to the control layer and causing malfunctions, a first-in-first-out (FIFO) data buffer queue of length N is established in the data preprocessing layer to perform recursive moving average filtering on the raw voltage signal. See also... Figure 3 During the "focused charging" phase after alignment, "voltage response feedback" is ongoing. The receiving end continuously transmits voltage data back so that the supply end can understand and monitor the current charging status.

[0151] A first-in, first-out (FIFO) data buffer queue of length N is allocated in memory. The window length N is a constant preset based on the sampling rate. Data update mechanism: Whenever a new instantaneous voltage is acquired, the new data is pushed to the head of the queue, and historical data overflowing from the tail is removed. The control unit uses the following recursive equation to obtain the smoothed and denoised load-side voltage:

[0152] (15)

[0153] in: This is the smoothed load-side voltage output after recursive moving average filtering at the current k-th sampling time; This is the smoothed historical value of the load terminal voltage derived at time k-1 of the previous sampling period; The original sampled value of the instantaneous DC voltage at the load terminal is obtained at the current k-th sampling time; N is the sliding window length preset according to the sampling rate and environmental noise characteristics; It is the oldest historical instantaneous voltage value that has just slipped out of the tail of the data buffer queue.

[0154] 4.2 Calculation of real-time effective transmission efficiency and determination of full charge shutdown:

[0155] Based on the above filtering process, the output is a smoothed load terminal voltage. By combining the load impedance R_L at the energy receiving end, the effective energy transfer efficiency at the k-th sampling time can be calculated in real time. :

[0156] (16)

[0157] (17)

[0158] in: This represents the total power of the transmitting end; This represents the effective value of the output voltage at the transmitting end. This represents the effective value of the output current at the transmitting end. This represents the load impedance at the receiving end.

[0159] The logic for determining whether charging has stopped is as follows:

[0160] The power supply control unit passes through While confirming that the magnetic focusing channel is in optimal transmission condition, monitor Dynamic change characteristics.

[0161] High-efficiency transmission monitoring: When Continuously maintain within the preset optimal efficiency range (e.g.) When this is done, confirm that the current magnetic focusing channel is properly aligned and in a controlled energy replenishment state;

[0162] Full charge status recognition: Under the premise of meeting the above-mentioned high-efficiency transmission, if the load output power is detected to show a trend of decreasing and remains within the preset minimum power threshold range for a set time period, it is determined that the power receiving battery has reached a full charge state.

[0163] Shutdown execution: After determining that the battery is fully charged, the control unit immediately deregisters the optimal excitation complex vector, cuts off the power output, issues a "charging complete" command, and enters the reset state.

[0164] It should be noted that this invention relies on the unique "load-independent constant voltage output" characteristic of the LCC-S compensation network: under fixed input excitation, the voltage at the receiving end is analytically mapped only to the spatial equivalent mutual inductance, and is unaffected by dynamic interference from the receiver impedance. This characteristic is the physical prerequisite for achieving "sensorless electrical alignment": it not only enables the calculation of the coupling coefficient and the three-dimensional offset coordinates of the receiving end in steps one and two based solely on the receiver voltage parameters, but also ensures that the feedback data in steps three and four during dynamic optimization and monitoring is not distorted, providing stable underlying data support for the entire process control closed loop of the system.

[0165] The above description is a specific illustration of the present invention, and not a limitation thereof. Those skilled in the art can make various equivalent technical solutions without departing from the scope of the present invention; therefore, all equivalent technical solutions should be included within the protection scope of the present invention.

Claims

1. A method for wireless power transfer between aerial unmanned aerial vehicles based on resonant magnetic coupling technology, characterized in that: Includes the following steps: Step 1: Spatial State Analysis and Initial Coupling Assessment: The power supply UAV flies above the power receiver UAV and applies an initial excitation signal. The power receiver responds to the excitation and transmits relevant parameters back to the load end. The power supply end uses the electrical response characteristics of the resonant network to perform forward analytical derivation of the theoretical voltage response, and combines the actual observed voltage to perform residual optimization, and reversely analyzes to obtain the optimal estimate of the instantaneous spatial mutual inductance; finally, it performs normalization calculation to extract the real-time spatial coupling coefficient, providing a data benchmark for subsequent determination of alignment status; Step 2, Positioning calibration of multi-coil polling excitation: Set a coupling coefficient threshold based on the minimum mutual inductance value when the receiving coil has the maximum offset, and compare the extracted real-time spatial coupling coefficient with this threshold to determine the alignment status: If aligned and the alignment conditions are met, proceed to step three. If misalignment occurs, the power supply UAV uses the spherical launch array as an equivalent multi-port magnetic network to initiate the positioning calibration function of multi-coil polling excitation; it constructs a two-dimensional magnetic coupling correlation matrix by collecting the voltage returned from the power receiving end; combined with the finite element spatial mapping model, it accurately calculates the three-dimensional spatial offset coordinates of the power receiving end; and then converts them into flight control commands to drive the airframe to perform physical position correction until the alignment conditions are met or the maximum number of retries is reached. Step 3: Spatial Magnetic Field Vector Synthesis and Focusing: After meeting the alignment conditions, the power supply control unit establishes a mathematical model of magnetic field superposition based on a multi-port magnetic network. Taking the maximization of the effective magnetic flux passing through the plane of the receiving coil under this model as the global optimization objective, and under the premise of strictly satisfying the self-compatible anti-mutual coupling heating constraint inside the array, the optimal excitation complex vector is solved. Finally, by executing this vector, amplitude and phase joint control is performed to complete the magnetic field vector synthesis in real physical space and construct a high-density magnetic focusing channel. Step 4, Dynamic Filtering and Optimal State Monitoring: During energy transmission, the power supply control unit performs recursive moving average filtering to denoise the load voltage fed back from the power receiver, and combines it with the real-time calculated effective transmission efficiency to confirm that it maintains the efficient and optimal transmission state in the magnetic focusing channel. Under the premise of this efficient transmission, the full charge state of the power receiver battery is determined by monitoring the duration of the load output power in the minimum threshold range, and then the power output is automatically cut off to achieve a safe closed loop for the charging task.

2. The method for wireless power transfer between aerial unmanned aerial vehicles based on resonant magnetic coupling technology according to claim 1, characterized in that: The first step, which involves analytically deriving the theoretical voltage response based on the electrical response characteristics of the resonant network, involves: extracting the effective value of the transmitter output voltage and defining the transmitter-side compensation inductance as a forward transfer function. When the energy receiving end has an unknown instantaneous equivalent mutual inductance At that time, its predicted theoretical one-dimensional load DC voltage output response for: （1） （2） in: The theoretical load-side DC voltage response value predicted at time t; For a positive transfer function operator determined by parameters; Let t be the instantaneous theoretical equivalent mutual inductance between the active port of the transmitting array and the receiving coil.

3. The method for wireless power transfer between unmanned aerial vehicles based on resonant magnetic coupling technology according to claim 1, characterized in that: Estimation objective function based on residual minimization To perform extreme value optimization, define the mutual inductance analytic operator. By performing algebraic reconstruction on the one-dimensional voltage measure, the optimal unbiased estimate of the current instantaneous spatial mutual inductance is analytically obtained. : （3） （4） （5） in: The objective function is the sum of squared residuals constructed based on the actual observed voltage and the theoretically predicted voltage. For the defined mutual inductance analytic operator; This represents the effective value of the output voltage at the transmitting end. Compensating inductor for the transmitting side; The actual load-side voltage transmitted back from the energy-receiving end at time t; This is the optimal unbiased estimate of the instantaneous spatial mutual inductance at the current moment.

4. The method for wireless power transfer between aerial unmanned aerial vehicles based on resonant magnetic coupling technology according to claim 1, characterized in that: The specific process of normalizing the calculation and extracting the real-time spatial coupling coefficient in step one is as follows: constructing a dimensionless feature projection operator. By applying the projection operator to the analytically obtained state estimate, and performing scale compression and feature normalization mapping on it, the dynamic invariant that purely characterizes the current spatial magnetic field coupling depth—the real-time spatial coupling coefficient—is extracted. : （6） （7） in: The real-time spatial coupling coefficient representing the alignment depth of the real physical space is extracted at time t. It is a dimensionless feature projection operator for the self-inductance parameters of the built-in transceiver coil.

5. A method for wireless power transfer between unmanned aerial vehicles based on resonant magnetic coupling technology according to claim 1, characterized in that: In step two, the constructed two-dimensional magnetic coupling correlation matrix characterizing the spatial coupling strength distribution is as follows: （8） Among them, matrix elements This represents the mutual inductance between the transmitting coil and the receiving coil located in the r-th row and c-th column of the spherical array.

6. The method for wireless power transfer between aerial unmanned aerial vehicles based on resonant magnetic coupling technology according to claim 1, characterized in that: In step two, when calculating the three-dimensional spatial offset coordinates of the receiving end, the supply end performs anomaly removal on the feature data and substitutes the elements of the correlation matrix into the finite element mapping model to construct the following weighted objective function based on minimizing the matrix mutual inductance error. : （9） in: The confidence weighting coefficients assigned to valid reference magnetic ports; The three-dimensional spatial coordinates of the center of the receiving coil of the powered UAV are to be solved. This is the theoretical spatial mutual inductance value between the reference magnetic ports; This represents the actual measured instantaneous equivalent mutual inductance between the reference magnetic port in row r and column c and the receiving coil.

7. A method for wireless power transfer between aerial unmanned aerial vehicles based on resonant magnetic coupling technology according to claim 1, characterized in that: In step three, the magnetic field superposition mathematical model is used to perform array vector superposition: Since the spatial positions and normal vectors of each coil unit on the spherical array are different, the control unit transforms the magnetic field vectors generated by all N units to a unified coordinate system for linear superposition; the total composite magnetic field at the receiving end is expressed as: （11） Where: is the geometric magnetic field factor generated by the i-th unit under unit current excitation, and is the excitation weighting coefficient of the unit.

8. A method for wireless power transfer between unmanned aerial vehicles based on resonant magnetic coupling technology according to claim 1, characterized in that: In step three, the effective magnetic flux passing through the plane of the receiving coil is maximized. Represented as: （12） Set a hard boundary constraint on the overall transmit power: Due to the limitations of the onboard energy storage and inverter capacity of the power supply UAV, the total input power of each magnetic port must be limited within the safe operating range. （13） Setting array self-compatible anti-coupling constraints: Strong spatial cross-inductance interference exists during concurrent excitation of high-density magnetic ports. To avoid eddy current heating and beam distortion caused by severe coupling within the transmitting array, the total interference power received by the i-th magnetic port is set based on predictive analysis equations. The temperature must be strictly below the preset device sensitivity heating threshold. : （14） in: Due to total power limitations, Let be the effective value of the current in the i-th transmitting coil. The AC resistance of the coil is... The total interference power received by the i-th transmitting magnetic port from the cross-coupling superposition of other active ports at the same frequency within the array; This represents the actual output excitation power of the j-th transmitting magnetic port; This is the internal magnetic coupling transfer function extracted from the magnetic coupling correlation matrix, which couples from the j-th transmitting magnetic port to the i-th transmitting magnetic port. The maximum cross-interference heating power and electromagnetic susceptibility threshold that a single magnetic port can withstand.

9. A method for wireless power transfer between unmanned aerial vehicles based on resonant magnetic coupling technology according to claim 1, characterized in that: In step four, the process of performing recursive moving average filtering for noise reduction is as follows: Construct a first-in-first-out buffer queue of length N, and perform a recursive moving average filter on the load voltage to obtain the smoothed and denoised load terminal voltage: （15） in: Let be the smoothed load-end voltage output after recursive moving average filtering at time k. The smoothed load terminal voltage history value is derived at time k-1. The original sampled value of the instantaneous DC voltage at the load terminal is obtained at the k-th sampling time; N is the sliding window length of the first-in-first-out data buffer queue. The oldest historical instantaneous voltage value that slides out of the tail of the data buffer queue.

10. A method for wireless power transfer between unmanned aerial vehicles based on resonant magnetic coupling technology according to claim 1, characterized in that: In step four, the effective energy transfer efficiency of the current control unit at the k-th sampling time is calculated in real time. : （16） （17） in: This represents the total power of the transmitting end; This represents the effective value of the output voltage at the transmitting end. This represents the effective value of the output current at the transmitting end. This represents the load impedance at the receiving end.

Images