A multi-load adaptive identification and power distribution method for a wireless charging plate

Abstract

The application discloses a kind of multi-load adaptive identification and power distribution method of wireless charging plate, it is related to wireless charging technical field, the method includes: control transmitting coil array emits electromagnetic detection pulse in turn, time domain reflection signal is synchronously collected, and time reflection feature library is constructed;Time reflection feature library is carried out time reversal processing, extracts time reflection focusing channel response, identifies load quantity and equivalent spatial position;Virtual focusing beam is structured and is formed into multi-load isolation focusing subspace set;Composite excitation waveform formed in each subspace, excite load receiving coil to generate nonlinear magnetization harmonic response, and extract energy absorption state characteristics;According to the characteristics, adjust the projection weight of each subspace, and realize the adaptive distribution of transmitting power between multiple loads by the joint modulation of the excitation amplitude and phase of each coil in transmitting coil array, can improve multi-load identification accuracy, reduce energy crosstalk between loads, improve wireless charging efficiency and safety.

Description

Technical Field

[0001] This invention relates to the field of wireless charging technology, and in particular to a method for multi-load adaptive identification and power allocation of a wireless charging pad. Background Technology

[0002] With the widespread adoption of wireless charging technology, simultaneous charging of multiple loads is becoming increasingly common. Existing wireless charging pads typically employ an array of multiple transmitting coils, transmitting energy to multiple receiving loads via electromagnetic induction. During multi-load charging, the coupling relationship between the receiving and transmitting coils is complex, and the coil positions, core characteristics, charging power requirements, and battery charging states of different loads can all change in real time. Current common multi-load identification and power allocation methods are mostly based on fixed-zone polling, impedance change detection, or single signal strength comparison. These methods struggle to accurately distinguish between multiple loads that are close to each other and cannot dynamically adjust power allocation based on the actual energy absorption state of each load. This can easily lead to problems such as energy crosstalk between loads, overcharging or undercharging of some loads, decreased charging efficiency, and reduced charging safety. Therefore, a method is needed that can adaptively identify multiple loads and achieve reasonable power allocation to improve the adaptability, charging efficiency, and safety of wireless charging pads in multi-load scenarios. Summary of the Invention

[0003] This invention overcomes the shortcomings of the prior art and provides a method for multi-load adaptive identification and power allocation of a wireless charging pad.

[0004] To achieve the above objectives, the technical solution adopted by the present invention is as follows: This invention discloses a multi-load adaptive identification and power allocation method for a wireless charging pad, comprising the following steps: S1. Control each coil in the transmitting coil array to sequentially transmit electromagnetic detection pulses with non-periodic time waveforms, and synchronously collect the time-domain reflection signals generated by near-field multi-load coupling at each coil port to construct a time-reflection feature library containing load spatial scattering information; S2. Perform time inversion processing on the time-domain reflection signals in the time-reflection feature library, extract the time-reflection focusing channel response corresponding to each load, and identify the number of loads and the equivalent spatial location of each load based on the degree of separation of peak clusters in the time-reflection focusing channel response. S3. Using the time-reflection focusing channel response, construct virtual focusing beams oriented towards each load, and spread the mutually orthogonal virtual focusing beams into a set of multi-load isolation focusing subspaces; S4. Within the set of multi-load isolated focusing subspaces, a composite excitation waveform consisting of amplitude perturbation components and phase perturbation components is superimposed on each subspace to excite the nonlinear magnetization harmonic response of each load receiving coil, and the energy absorption state characteristics corresponding to each load are extracted. S5. Based on the energy absorption state characteristics, adjust the projection weight of each subspace in the total power distribution, and by jointly modulating the excitation amplitude and phase of each coil in the transmitting coil array, make the transmitting power adaptively distributed among the loads according to the updated projection weight.

[0005] This invention addresses the technical deficiencies in the prior art and offers the following advantages: By using aperiodic probe pulses and time-reversal focusing processing, this invention can accurately distinguish multiple loads and their equivalent spatial positions under near-field coupling. Furthermore, it utilizes orthogonal virtual focusing beams to construct isolated subspaces, effectively suppressing energy crosstalk between loads. By extracting the actual energy absorption state of each load through nonlinear magnetization harmonic response, and dynamically adjusting the power distribution weights of each subspace and the excitation amplitude and phase of the transmitting coil accordingly, the transmitting power can adaptively distribute itself according to the charging needs and core state of each load. This avoids overcharging or undercharging of some loads, improving the efficiency, safety, and robustness of multi-load wireless charging. Attached Figure Description

[0006] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other embodiments can be obtained from these drawings without creative effort.

[0007] Figure 1 This is a flowchart of the first method of the present invention; Figure 2 This is a flowchart of the second method of the present invention. Detailed Implementation

[0008] To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.

[0009] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and therefore the scope of protection of the invention is not limited to the specific embodiments disclosed below.

[0010] like Figure 1 , Figure 2 As shown, this invention discloses a multi-load adaptive identification and power allocation method for a wireless charging pad, comprising the following steps: S1. Control each coil in the transmitting coil array to sequentially transmit electromagnetic detection pulses with non-periodic time waveforms, and synchronously collect the time-domain reflection signals generated by near-field multi-load coupling at each coil port to construct a time-reflection feature library containing load spatial scattering information; S2. Perform time inversion processing on the time-domain reflection signals in the time-reflection feature library, extract the time-reflection focusing channel response corresponding to each load, and identify the number of loads and the equivalent spatial location of each load based on the degree of separation of peak clusters in the time-reflection focusing channel response. S3. Using the time-reflection focusing channel response, construct virtual focusing beams oriented towards each load, and spread the mutually orthogonal virtual focusing beams into a set of multi-load isolation focusing subspaces; S4. Within the set of multi-load isolated focusing subspaces, a composite excitation waveform consisting of amplitude perturbation components and phase perturbation components is superimposed on each subspace to excite the nonlinear magnetization harmonic response of each load receiving coil, and the energy absorption state characteristics corresponding to each load are extracted. S5. Based on the energy absorption state characteristics, adjust the projection weight of each subspace in the total power distribution, and by jointly modulating the excitation amplitude and phase of each coil in the transmitting coil array, make the transmitting power adaptively distributed among the loads according to the updated projection weight.

[0011] Specifically, S1 is: The control coils in the transmitting coil array sequentially transmit electromagnetic detection pulses with non-periodic time waveforms to the charging area according to a preset polling sequence, and simultaneously collect the time-domain reflection signals generated by near-field multi-load coupling at each coil port to form a multi-channel time-domain reflection signal set; Each time-domain reflection signal in the multi-channel time-domain reflection signal set is synchronously intercepted and DC-free preprocessed to extract the scattered echo component caused by near-field multi-load coupling and generate a coupled scattered echo sequence corresponding to each transmitting coil. The coupled scattered echo sequences corresponding to each transmitting coil are combined according to the transmitting coil position index of each transmitting coil in the transmitting coil array, and a time inversion operator is applied to the combined coupled scattered echo sequences to obtain the time-inverse characteristic response corresponding to each transmitting coil; the time inversion operator is used to enable the scattering paths from different loads in the coupled scattered echo sequences to be separably reconstructed on the time axis; Based on the arrival time distribution and spatial phase relationship of the scattered paths in the time-reflection characteristic response corresponding to each transmitting coil, the time delay label of the scattered path representing the spatial location of each load and the amplitude label of the scattered path representing the strength of load coupling are extracted to form the load spatial scattering characteristic vector. The load space scattering feature vector is associated with and stored with the corresponding transmitting coil position index to construct a time-reversal feature library containing load space scattering information.

[0012] The aperiodic time waveform is composed of a periodic carrier component and an aperiodic envelope component. The aperiodic envelope component applies an aperiodic amplitude perturbation to the periodic carrier component in the time domain, so that the electromagnetic detection pulse generates an electromagnetic response containing multiple load scattering path information in the near field region.

[0013] It should be noted that the wireless charging pad includes a controller, a drive circuit, a transmitting coil array, and a signal acquisition module. Each transmitting coil in the transmitting coil array has a unique position index, which reflects the row and column coordinates or spatial arrangement order of the transmitting coil in the array plane. The controller is used to send enable signals to the drive circuit according to a preset polling sequence, causing each transmitting coil to enter the energized state sequentially; the signal acquisition module is used to synchronously acquire the time-domain reflection signals of each coil port.

[0014] Specifically, the controller enables each transmitting coil in the transmitting coil array according to a preset polling sequence. Taking the enabling of a particular transmitting coil as an example, the drive circuit outputs an aperiodic time waveform composed of a periodic carrier component and an aperiodic envelope component. The periodic carrier component provides the basic electromagnetic excitation, while the aperiodic envelope component applies an aperiodic amplitude perturbation to the periodic carrier component in the time domain, causing the envelope of the electromagnetic probe pulse actually fed into the transmitting coil to exhibit aperiodic changes on the time axis. The purpose of using an aperiodic time waveform is that when the electromagnetic probe pulse encounters multiple load receiving coils and their magnetic cores in the near-field region, the time-domain reflection signal formed at the port of the transmitting coil by the scattering paths caused by each load can carry richer spatial scattering information, avoiding the inter-path ambiguity that may be caused by the periodic waveform. Simultaneously, the signal acquisition module acquires the time-domain reflection signals generated by the near-field multi-load coupling at the port of the transmitting coil and other coil ports in the array while the transmitting coil transmits the probe pulse, forming a multi-channel time-domain reflection signal set. Here, "multi-channel" corresponds to the sampling channel of each coil port in the transmitting coil array.

[0015] For each time-domain reflection signal in the multi-channel time-domain reflection signal set, the controller first performs synchronous interception based on the probe pulse emission time to obtain a time-domain analysis window containing near-field reflection information. Then, the signal within this time-domain analysis window undergoes DC removal preprocessing, such as removing the DC bias of the sampling channel itself and the static background coupling component under no-load conditions. After DC removal preprocessing, the remaining components are mainly scattered echo components caused by near-field multi-load coupling, thereby generating a coupled scattered echo sequence corresponding to each transmitting coil. The scattered echo component can be understood as the response component of the probe pulse returning to the transmitting coil port after coupling reflection at the load receiving coil, magnetic core, and surrounding metal structure. The coupled scattered echo sequences corresponding to each transmitting coil are combined according to the position index of the transmitting coil in the transmitting coil array, so that the arrangement order of the sequences is consistent with the spatial geometric arrangement of the array. After combination, a time-reversal operator is applied to the combined coupled scattered echo sequences. The time-reversal operator reverses the time axis of each coupled scattered echo sequence to obtain the time-reversal characteristic response corresponding to that transmitting coil. Because different loads are located at different spatial positions in the near-field region, their scattered paths arrive at intertwined times in the original time-domain reflected signal. After time inversion, these scattered paths are redistributed in the time-inverse characteristic response according to their reverse time delays, thus allowing for separable reconstruction of the scattered paths from different loads along the time axis. For example, when the scattered paths of two loads partially overlap in the original signal, after time inversion, the scattered path of one load may be relatively shifted forward, while the scattered path of the other load may be relatively shifted backward, resulting in a distinguishable temporal structure in the time-inverse characteristic response.

[0016] Then, the time-reflection characteristic responses corresponding to each transmitting coil are analyzed. Based on the arrival time distribution and spatial phase relationship of the scattered paths, scattered path delay labels and scattered path amplitude labels are extracted. Specifically, in each time-reflection characteristic response, the position of the local maxima of the scattered path on the time axis constitutes the arrival time distribution. The same load will generate a set of scattered paths in the time-reflection characteristic responses corresponding to different transmitting coils. The relative arrival times of this set of scattered paths between different coils form the scattered path delay label, which is used to characterize the spatial location of the load. The amplitude or energy of this set of scattered paths forms the scattered path amplitude label, which is used to characterize the coupling strength between the load and the transmitting coil array. Taking a load located between two adjacent coils in the transmitting coil array as an example, its scattered path will form local maxima in the time-reflection characteristic responses corresponding to both coils. The relative time delay relationship of these local maxima in the two coils reflects the degree to which the load is close to the two coils, and the amplitude of the scattered path reflects the tightness of the coupling between the load and the coil. Combining the above scattered path delay labels and scattered path amplitude labels, the load spatial scattering characteristic vector of the load can be formed.

[0017] Specifically, S2 is: Retrieve the load space scattering feature vectors corresponding to the position indices of each transmitting coil from the time-reflection feature library, and rearrange the scattering path delay labels and scattering path amplitude labels in each load space scattering feature vector according to the transmitting coil position index to form a load scattering path delay-amplitude distribution matrix. Time inversion focusing processing is applied to each scattering path delay label in the load scattering path delay-amplitude distribution matrix, and the scattering path amplitude labels corresponding to each transmitting coil are superimposed along the time inversion delay direction to generate a time inversion focusing channel response characterizing the energy convergence characteristics of each load. Envelope extraction and local peak search are performed on the time-inversion focusing channel response. Local peaks that are continuous along the time delay direction and whose amplitude exceeds a preset noise floor threshold in the time inversion domain are grouped into peak clusters to form a set of peak clusters. Each peak cluster corresponds to an energy focusing region of a candidate load in the time inversion domain. Calculate the time delay interval and amplitude valley depth between adjacent peak clusters in the peak cluster set, and combine them with the time delay width of each peak cluster to obtain the peak cluster separation degree; When the peak cluster separation between adjacent peak clusters is greater than a preset separation threshold, the adjacent peak clusters are determined to belong to different loads; when the peak cluster separation between adjacent peak clusters is not greater than the preset separation threshold, the adjacent peak clusters are determined to belong to different scattering paths of the same load. The number of loads is determined based on the number of peak clusters belonging to different loads, and the equivalent spatial position of each load is inverted based on the time delay position of each peak cluster in the time-reflection focusing channel response and the spatial geometric relationship with the transmitting coil array.

[0018] In this embodiment, the load spatial scattering feature vector corresponding to the position index of each transmitting coil is retrieved from the time-reversal feature library, and the scattering path delay labels and scattering path amplitude labels contained therein are rearranged according to the transmitting coil position index to form a load scattering path delay-amplitude distribution matrix. Each row or column in this matrix corresponds to a transmitting coil, and the matrix elements are represented by pairs of delay and amplitude, used to describe the scattering path information left by the same load or different loads in the array space. Time-reversal focusing processing is applied to each scattering path delay label in the load scattering path delay-amplitude distribution matrix. It should be noted that the time-reversal operator in S1 is mainly used to make the scattering paths from different loads in the original coupled scattering echo sequence separable and reconstructable on the time axis; while the time-reversal focusing processing here is based on the already formed scattering path delay labels, aligning the scattering path delay labels formed by a certain load on different transmitting coils along the inverted delay direction, so that the scattering path amplitude labels of the load are coherently superimposed at the same delay position. For example, if a load corresponds to different scattering paths at different time delay positions on three transmitting coils, after time inversion focusing processing, these three time delay positions are mapped to the vicinity of the same focusing time delay. The three amplitudes are superimposed to form an energy concentration region, thereby generating a time inversion focusing channel response that characterizes the energy convergence characteristics of each load.

[0019] In the time-inversion focusing channel response, its amplitude envelope is obtained through envelope extraction. Envelope extraction can be achieved by modulo operation or moving root mean square calculation of the time-inversion focusing channel response. Then, a local peak search is performed, i.e., the envelope amplitude is compared point by point along the time delay direction. If the amplitude of a point is simultaneously greater than the amplitudes of its two adjacent points and exceeds a preset noise floor threshold, then that point is recorded as a local peak. The preset noise floor threshold can be determined based on the noise floor envelope root mean square value collected under a no-load static environment, for example, three times the noise floor envelope root mean square value. A group of local peaks continuously distributed along the time delay direction in the time inversion domain with amplitudes exceeding the preset noise floor threshold are grouped into a peak cluster, resulting in a peak cluster set. Each peak cluster corresponds to the energy focusing region of a candidate load in the time inversion domain. For example, a peak cluster may contain multiple closely spaced local peaks within the same continuous time delay interval, corresponding to the energy concentration performance of the same candidate load after multipath focusing.

[0020] In a set of peak clusters, the time delay interval and amplitude valley depth between adjacent peak clusters are calculated. The time delay interval can be taken as the time delay difference between the peak positions of two adjacent peak clusters; the amplitude valley depth can be taken as the ratio of the amplitude of the lowest point of the envelope between two peak clusters to the amplitude of the lower peak in the two peak clusters; the peak cluster time delay width can be taken as the time delay span corresponding to a certain proportion when the amplitude of the peak cluster envelope decreases from the peak to the peak. Combining the time delay interval, amplitude valley depth, and peak cluster time delay width, the peak cluster separation degree can be obtained. In one exemplary method, the peak cluster separation degree can be obtained by correcting the amplitude valley depth for the ratio of the time delay interval to the average of the time delay widths of two peak clusters. A preset separation degree threshold is, for example, 0.7. When the peak cluster separation degree between adjacent peak clusters is greater than 0.7, it indicates that the two peak clusters are distinguishable in time delay and have deep valleys, and it is determined that the adjacent peak clusters belong to different loads; when the peak cluster separation degree is not greater than 0.7, it indicates that the two peak clusters are continuous in time delay or have shallow valleys, and it is determined that the adjacent peak clusters belong to different scattering paths of the same load.

[0021] The number of loads is determined based on the number of peak clusters belonging to different loads. The equivalent spatial position is retrieved by utilizing the time delay position of each peak cluster in the time-inverse focusing channel response and the spatial geometric relationship with the transmit coil array. Since the position indices of each coil in the transmit coil array are known, the time delay difference of the scattering path formed by the same load on different transmit coils is related to the spatial distance difference from the load to each transmit coil. Therefore, by comparing the position offsets of the peak clusters corresponding to the same load in the corresponding time delay directions of multiple transmit coils, the spatial coordinates of the load relative to the transmit coil array can be determined, thus obtaining the equivalent spatial position of each load. The equivalent spatial position serves as the spatial reference for subsequently constructing the virtual focusing beam and the multi-load isolated focusing subspace.

[0022] Specifically, S3 is: Based on each peak cluster in the time-reverse focusing channel response, the time delay center position and peak cluster energy envelope corresponding to each peak cluster are extracted, and the time-reverse focusing channel response segment corresponding to each peak cluster is used as the initial focusing weight vector of the load corresponding to the peak cluster. Based on the spatial distance relationship between the equivalent spatial position of each load and the position index of each transmitting coil in the transmitting coil array, spatial phase compensation is applied to each initial focusing weight vector to form a spatial focusing beam that matches the spatial position of each load. Calculate the overlap component between any two spatial focusing beams, and sequentially subtract the overlap component with other spatial focusing beams from each spatial focusing beam until the remaining overlap component is lower than a preset overlap component threshold, to obtain mutually orthogonal virtual focusing beams. Using the mutually orthogonal virtual focusing beams as basis vectors, a set of multi-load isolated focusing subspaces corresponding to the number of loads is formed; wherein, each virtual focusing beam corresponds to an isolated focusing subspace, and the isolated focusing subspaces are orthogonally isolated from each other; Calculate the subspace isolation degree between any two isolated focusing subspaces in the multi-load isolated focusing subspace set. If the isolation degree of any subspace is lower than the preset isolation degree threshold, apply orthogonal compensation correction to the virtual focusing beams corresponding to the two isolated focusing subspaces and update the multi-load isolated focusing subspace set until the isolation degree of all subspaces meets the preset isolation degree threshold.

[0023] Specifically, for each peak cluster in the time-reversal focusing channel response, the time delay center position of the peak cluster is first determined. The time delay center position can be obtained by energy-weighted averaging of the time delay values ​​corresponding to the local peaks within the peak cluster, or by directly taking the time delay position corresponding to the local peak with the largest amplitude in the peak cluster. The peak cluster energy envelope can be determined by extracting the amplitude envelope curve of the peak cluster in the time delay direction and intercepting the time delay interval corresponding to the portion where the envelope amplitude drops to a certain percentage of the peak value (e.g., 80% of the peak value) centered on the time delay center position. The segment in the time-reversal focusing channel response corresponding to the peak cluster is used as the initial focusing weight vector for the load corresponding to that peak cluster. The initial focusing weight vector can be understood as a set of complex weighting coefficients applied to each channel of the transmit coil array, used to form energy focusing at the location of the load.

[0024] Since the equivalent spatial positions of each load have been obtained in step S2, and the position indices of each transmitting coil in the transmitting coil array are known, the spatial distance difference from each load to each transmitting coil can be calculated. This spatial distance difference causes phase delay differences in electromagnetic waves or near-field coupling fields. Based on this spatial distance relationship, spatial phase compensation is applied to the elements corresponding to each transmitting coil in the initial focusing weight vector for each load. The phase value of this spatial phase compensation is related to the spatial distance difference from the load to each transmitting coil. After compensation, the excitation signals of each transmitting coil can be superimposed in phase at the equivalent spatial position of the load, thereby forming a spatially focused beam that matches the spatial position of each load. For example, if the distance difference from a load to two adjacent transmitting coils causes phase lag, a positive phase offset is added to the coil element that is closer in the initial focusing weight vector corresponding to that load, and a negative phase offset is added to the coil element that is farther away, so that the two are in phase at the load position.

[0025] Then, the overlap component between any two spatial focusing beams is calculated. The overlap component can be measured by the inner product or correlation of the two spatial focusing beams in the excitation space of the transmitting coil array. The overlap component with other spatial focusing beams is sequentially subtracted from each spatial focusing beam, i.e., the projection of the current beam onto other beam directions is subtracted from the current beam, until the residual overlap component is below a preset overlap component threshold. The preset overlap component threshold is, for example, 10% of the energy percentage of the overlap component before subtraction. After processing, the resulting beams are mutually orthogonal and are called virtual focusing beams. Each virtual focusing beam corresponds to a load, and its directional isolation is significantly improved compared to the initial spatial focusing beam.

[0026] Using each virtual focusing beam as a basis vector, a set of multi-load isolated focusing subspaces corresponding to the number of loads is formed. Specifically, each virtual focusing beam is considered a basis vector, and all basis vectors together span a set of subspaces, where each basis vector direction corresponds to an isolated focusing subspace. Since the virtual focusing beams are essentially orthogonal, the isolated focusing subspaces can maintain orthogonal isolation. Orthogonal isolation means that when excitation is applied in any isolated focusing subspace, the energy leakage to other isolated focusing subspaces is small. To ensure that the isolation degree between the isolated focusing subspaces meets the requirements for subsequent load energy absorption state extraction, the subspace isolation degree between any two isolated focusing subspaces in the multi-load isolated focusing subspace set is calculated. The subspace isolation degree can be measured by the orthogonality between the corresponding basis vectors of the two subspaces, for example, using the square of the product modulus in the two basis vectors as a leakage index; the higher the isolation degree, the smaller the leakage. The preset isolation degree threshold is, for example, 0.9 (or equivalently represented by 20dB isolation degree). If the isolation degree of any subspace is lower than the preset isolation degree threshold, orthogonal compensation correction is applied to the virtual focusing beams corresponding to these two isolated focusing subspaces. Orthogonal compensation correction can be achieved by applying orthogonal constraints to the two virtual focusing beams again on the basis of the original orthogonalization process. For example, the residual overlap components between the two beams can be subtracted again, and the corrected virtual focusing beams can be normalized. After updating the multi-load isolation focusing subspace set, the isolation degree is recalculated until the isolation degree of all subspaces reaches the preset isolation degree threshold.

[0027] Specifically, S4 is: For each isolation focusing subspace in the set of multi-load isolation focusing subspaces, a composite excitation waveform formed by the superposition of amplitude perturbation components and phase perturbation components is generated; Each composite excitation waveform is applied to the corresponding isolated focusing subspace, so that each load receiving coil generates nonlinear magnetization harmonic components under the action of the focusing magnetic field after superimposed perturbation; the nonlinear magnetization harmonic components carry the nonlinear magnetization characteristics of the magnetic core of the load receiving coil. The nonlinear magnetization harmonic components of each load receiving coil are synchronously acquired at the transmitting end by the transmitting coil array, and the nonlinear magnetization harmonic response corresponding to each isolated focusing subspace is separated from the reflected coupling signal. The nonlinear magnetization harmonic responses between different isolated focusing subspaces do not crosstalk each other due to orthogonal isolation. Harmonic order decomposition is performed on each of the nonlinear magnetization harmonic responses to extract the relative amplitude distribution and phase shift characteristics of each harmonic component, and the energy absorption state characteristics of the corresponding load are constructed based on the relative amplitude distribution and phase shift characteristics. The energy absorption state characteristics of each load are associated with its isolated focusing subspace to form a set of load energy absorption state characteristics. If the similarity between any two energy absorption state features exceeds a preset similarity threshold, then a phase shift adjustment is applied to the phase perturbation component of the corresponding isolated focusing subspace and the load energy absorption state feature set is updated.

[0028] The amplitude perturbation component has a smaller perturbation bandwidth than the phase perturbation component, in order to maintain the orthogonal isolation between the isolation focusing subspaces and to induce the load receiving coil to enter a nonlinear magnetization state.

[0029] It should be noted that, for each isolated focusing subspace, a composite excitation waveform is first generated, formed by the superposition of amplitude and phase perturbation components. The amplitude perturbation component is used to time-vary modulate the amplitude of the basic excitation, and the phase perturbation component is used to time-vary modulate the phase of the basic excitation. The perturbation bandwidth of the amplitude perturbation component is set smaller than that of the phase perturbation component, resulting in relatively smooth amplitude changes and relatively rapid phase changes. The perturbation bandwidth can be understood as a measure of the main energy distribution range or perturbation rate of the perturbation component in the frequency domain. For example, the amplitude perturbation component can use an envelope signal with a low rate of change, and the phase perturbation component can use a phase modulation signal with a high rate of change; their superposition forms the composite excitation waveform. The purpose of this setting is to maintain the orthogonal isolation between the isolated focusing subspaces while enabling the magnetic core of the load receiving coil to be excited with a more significant nonlinear magnetization effect.

[0030] The composite excitation waveform corresponding to each isolated focusing subspace is loaded into that subspace. Specifically, each isolated focusing subspace uses the virtual focusing beam obtained in step S3 as its basis vector, and the basis vector corresponds to a set of complex weighting coefficients applied to the transmitting coil array. The composite excitation waveform is used as the excitation signal for that isolated focusing subspace and weighted with the virtual focusing beam to obtain the actual driving signal fed into each transmitting coil. In this way, at the spatial location of the corresponding load, the magnetic fields generated by each transmitting coil are superimposed to form a focusing magnetic field carrying composite disturbances. Under the action of this focusing magnetic field, the magnetic core of the load receiving coil enters a nonlinear magnetization state, generating nonlinear magnetization harmonic components other than the fundamental wave; these nonlinear magnetization harmonic components carry the nonlinear magnetization characteristics of the magnetic core of the load receiving coil, such as the degree of core saturation and the strength of nonlinear permeability.

[0031] Subsequently, the reflected coupling signals induced at the transmitting end by the nonlinear magnetization harmonic components of each load receiving coil are synchronously acquired through the transmitting coil array. Since the isolated focusing subspaces maintain orthogonal isolation, the nonlinear magnetization harmonic responses corresponding to different loads are superimposed but separable in the reflected coupling signals. During separation, the virtual focusing beams corresponding to each isolated focusing subspace are used as references, and the multi-channel reflected coupling signals are projected onto the directions of each virtual focusing beam to obtain the nonlinear magnetization harmonic responses corresponding to each isolated focusing subspace. For example, by performing a correlation operation between the reflected coupling signal and a certain virtual focusing beam, the response component corresponding to that subspace can be extracted, while the components of other orthogonal subspaces are suppressed due to low correlation.

[0032] Harmonic order decomposition is performed on each nonlinear magnetization harmonic response. Specifically, the amplitude and phase of each harmonic component, such as the fundamental component, second harmonic component, and third harmonic component, can be extracted using Fourier transform or phase-locked loop detection. Using the fundamental component as a reference, the relative amplitude distribution of each harmonic component is calculated, such as the ratio of the second harmonic amplitude to the fundamental amplitude, the ratio of the third harmonic amplitude to the fundamental amplitude, etc. Simultaneously, the phase shift characteristics of each harmonic component relative to the fundamental component are extracted. The relative amplitude distribution and phase shift characteristics are combined to construct the energy absorption state characteristics of the load. These energy absorption state characteristics reflect the load's absorption capacity of the current focusing magnetic field and whether the magnetic core is close to saturation. The energy absorption state characteristics corresponding to each load are associated with its respective isolated focusing subspace to form a set of load energy absorption state characteristics. If the similarity between any two energy absorption state characteristics exceeds a preset similarity threshold, it indicates that the two loads may have produced similar nonlinear magnetization behaviors under the current excitation conditions, which could easily lead to feature confusion during subsequent power allocation. In this case, a phase shift adjustment is applied to the phase perturbation component of the corresponding isolated focusing subspace. For example, a preset similarity threshold of 0.9 can be used; the degree of similarity can be measured by the normalized correlation coefficient between two energy absorption state features. Phase offset adjustment can be performed by adding different fixed phase offsets to the two isolated focusing subspaces, so that the phase perturbation components of the two differ in time. Then, the energy absorption state features are re-acquired and extracted, and the load energy absorption state feature set is updated until there is sufficient distinguishability between each energy absorption state feature.

[0033] Specifically, S5 is: The energy absorption state features corresponding to each load are retrieved from the set of load energy absorption state features. The relative amplitude distribution and phase shift features in each energy absorption state feature are analyzed to extract the saturation margin feature reflecting the load's absorption capacity of the current focusing magnetic field and the coupling efficiency feature reflecting the load's utilization efficiency of the current focusing magnetic field. The saturation margin feature and the coupling efficiency feature are then fused into the load power demand weight. The power demand weights of each load are mapped to the corresponding isolation and focusing subspaces in the set of multi-load isolation and focusing subspaces to form the projection weights of each isolation and focusing subspace. Normalize each projection weight to obtain a subspace power allocation vector, where each dimension of the subspace power allocation vector corresponds to a projection weight of an isolated focusing subspace. Based on the subspace power allocation vector, and combined with the spatial correspondence between the position index of each transmitting coil in the transmitting coil array and each isolated focusing subspace, the projection weight of each isolated focusing subspace is decomposed to each transmitting coil in the transmitting coil array to form a transmitting coil excitation amplitude allocation vector. Phase constraints are applied to each element in the excitation amplitude allocation vector of the transmitting coil to keep the excitation phases of each transmitting coil corresponding to the same isolation focusing subspace synchronized, and a phase isolation bias is introduced between the excitation phases corresponding to different isolation focusing subspaces to obtain the joint excitation vector of the transmitting coils. The excitation amplitude and excitation phase of each coil in the transmitting coil array are jointly modulated using the joint excitation vector of the transmitting coils, and the energy absorption state characteristics of each load are sampled back in real time during the modulation process. If the energy absorption state characteristics obtained after re-sampling show a decrease in saturation margin characteristics compared with those before modulation, the projection weights of the corresponding isolated focusing subspace are rolled back by a preset step size and the subspace power allocation vector is updated until the energy absorption state characteristics of each load are restored to a preset stable state, so that the transmit power is adaptively allocated among the loads according to the updated projection weights.

[0034] It should be noted that the energy absorption state characteristics include at least the relative amplitude distribution of each harmonic component relative to the fundamental wave, as well as the phase shift characteristics of each harmonic component relative to the fundamental wave. When analyzing the energy absorption state characteristics of each load, the relative amplitudes of the second and third harmonics are the primary analytical objects. The relative amplitude of the third harmonic is particularly sensitive to the saturation level of the load's receiving coil core; when the core approaches saturation, the relative amplitude of the third harmonic increases significantly. The relative amplitude of the second harmonic is related to the nonlinear magnetization of the core. Therefore, the saturation margin characteristic reflecting the load's absorption capacity of the current focusing magnetic field can be extracted using the trend or ratio of the relative amplitude of the third harmonic to that of the second harmonic. For example, when the ratio of the relative amplitude of the third harmonic to that of the second harmonic exceeds a certain reference range, the core of the load can be considered to be close to saturation, with a small saturation margin. Simultaneously, the coupling efficiency characteristic reflecting the load's utilization efficiency of the current focusing magnetic field can be extracted using the ratio of the fundamental wave reflected coupling signal amplitude or the total energy of each harmonic component to the excitation energy of the subspace. A higher coupling efficiency indicates that the load can absorb energy more effectively under the current focusing magnetic field.

[0035] The saturation margin feature and coupling efficiency feature are fused into load power demand weights. The fusion method can be a weighted superposition or a product. For example, the saturation margin feature can be used as an upper limit constraint factor, and the coupling efficiency feature as a demand intensity factor; multiplying the two yields the load power demand weights. A larger weight indicates that the corresponding load needs more transmit power in the current state; a smaller weight indicates that the corresponding load is either close to saturation or has low coupling efficiency, and power allocation should be reduced. After obtaining the power demand weights for each load, they are mapped to the corresponding isolated focusing subspaces in the multi-load isolated focusing subspace set. Since each isolated focusing subspace corresponds one-to-one with a load, the power demand weight of each load can be directly assigned to the corresponding isolated focusing subspace as its projection weight. Subsequently, each projection weight is normalized so that the sum of all projection weights equals 1, resulting in a subspace power allocation vector. Each dimension of this vector corresponds to the projection weight of an isolated focusing subspace, reflecting the proportion of that subspace in the total transmit power.

[0036] Next, the subspace power allocation vector needs to be decomposed into the transmit coils in the transmit coil array. Since each isolated focusing subspace is spanned by a corresponding virtual focusing beam, and this virtual focusing beam contains complex weighted coefficients for that subspace on each transmit coil, the magnitudes of these coefficients reflect the contribution of each transmit coil to the energy focusing of that subspace. Therefore, the projection weight of each dimension in the subspace power allocation vector can be multiplied by the amplitude coefficient of the corresponding virtual focusing beam on each transmit coil to obtain the amplitude component of that subspace allocated to each transmit coil. Summing the amplitude components of all subspaces on the same transmit coil forms the transmit coil excitation amplitude allocation vector. Each element in this vector corresponds to the basic excitation amplitude of a transmit coil.

[0037] Phase constraints are applied based on amplitude allocation. For each transmitting coil corresponding to the same isolated focusing subspace, its excitation phase is kept consistent with the phase specified by the virtual focusing beam of that subspace, ensuring that the magnetic fields of each transmitting coil are superimposed in phase at the corresponding load equivalent spatial position, thus guaranteeing focusing effect. For the excitation phases corresponding to different isolated focusing subspaces, a phase isolation bias is introduced. The phase isolation bias can apply different fixed phase shifts to different subspaces, for example, adding 0 phase shift and 90-degree phase shift to two subspaces respectively. In this way, even if there is residual spatial overlap, the magnetic fields of different subspaces are distinguishable in time phase, which helps to further reduce energy crosstalk. The combination of amplitude allocation and phase constraints forms a joint excitation vector for the transmitting coils.

[0038] The excitation amplitude and phase of each coil in the transmitter coil array are jointly modulated using a joint excitation vector of the transmitter coils. During modulation, the energy absorption state characteristics corresponding to each load are sampled in real time. If the saturation margin characteristic of a certain load decreases compared to before modulation, it indicates that the power currently allocated to that load may exceed the linear range that its magnetic core can withstand, and further increasing the power will reduce the energy absorption efficiency. At this time, the projection weight of the isolation focusing subspace corresponding to that load is rolled back by a preset step size. The preset step size is, for example, 10% of the current projection weight of the isolation focusing subspace, and the subspace power allocation vector is updated. After the update, the joint excitation vector of the transmitter coils is recalculated and modulation continues until the sampled energy absorption state characteristics of each load return to a preset stable state. The preset stable state can be defined as the saturation margin characteristic change amplitude not exceeding 5%, or the saturation margin characteristic stabilizing within a preset reference range. In this way, the transmitter power can be adaptively allocated among the loads according to the actual energy absorption state of each load.

[0039] In this embodiment, it also includes: The excitation amplitude of each coil in the transmitting coil array is encoded as an amplitude sub-vector of the search individual position vector, and the excitation phase of each coil is encoded as a phase sub-vector. The two vectors are then concatenated to form the search individual position vector. Initialize a population of search individuals consisting of multiple search individual position vectors, and map each search individual position vector to the corresponding isolation focusing subspace in the set of multiple load isolation focusing subspaces to form an initial population of joint excitation vectors for the transmitting coils; For each search individual's position vector in the initial population, the total transmission efficiency feature, subspace isolation feature, and load absorption balance feature are calculated based on the load energy absorption state feature set. The three are then weighted and fused into the search individual's fitness value. When the subspace isolation feature is lower than the preset isolation threshold, an orthogonal isolation constraint penalty term is added to the search individual's fitness value. The search individuals are sorted according to their fitness values, and the current global best search individual position vector is recorded. For each search individual position vector, an adaptive weight coefficient is determined based on its fitness value, and the amplitude sub-vector is updated using this adaptive weight coefficient. At the same time, the phase isolation bias correction is calculated based on the difference between the phase sub-vectors corresponding to adjacent isolation focusing subspaces, and the phase sub-vector is updated to obtain the updated search individual position vector population. The position vector of each search individual in the updated population is projected onto the feasible region that satisfies the orthogonal isolation constraint according to the subspace isolation direction corresponding to the phase isolation bias correction, to obtain the corrected search individual position vector population; the fitness value of each corrected search individual position vector is recalculated, and the orthogonal isolation constraint penalty term is reapplied to update the current global optimal search individual position vector; Compare the change in fitness value of the global best search individual position vector before and after the update. If it is less than the preset convergence threshold, the current global best search individual position vector is output as the joint excitation vector of the transmitting coil. Otherwise, apply saturation margin backoff adjustment to the amplitude subvector in the updated population according to the adaptive weight coefficient of each search individual position vector, and return to execute the search individual fitness sorting and update.

[0040] It should be noted that when the number of loads is large or the scale of the transmitting coil array is large, if the distribution is directly decomposed to each transmitting coil based solely on the projection weights of each isolated focusing subspace, the combination space of the excitation amplitude and phase of each coil is large. Furthermore, there is mutual coupling between the isolation degree of different subspaces and the load energy absorption state, which can easily lead to a decrease in the isolation degree of some subspaces or premature saturation of individual loads. Therefore, this embodiment uses the load energy absorption state feature set and subspace isolation degree as optimization criteria, and solves for the joint excitation vector of the transmitting coils within the feasible region through a population search method. Specifically, the excitation amplitude of each coil in the transmitting coil array is encoded as an amplitude subvector, and the excitation phase of each coil is encoded as a phase subvector. The two are concatenated end-to-end to form the search individual position vector. Multiple search individual position vectors are randomly generated to form a search individual population, and each search individual position vector is mapped to the corresponding isolated focusing subspace in the multi-load isolated focusing subspace set to obtain the initial population of the joint excitation vector of the transmitting coils. For each search individual's position vector in the initial population, three features are calculated based on the load energy absorption state feature set: the total transmission efficiency feature is the ratio of the sum of the coupling efficiency features corresponding to all loads to the total excitation energy corresponding to the search individual; the subspace isolation feature is the minimum or average value of the subspace isolation between adjacent isolated focusing subspaces; and the load absorption balance feature is the degree of difference between each load saturation margin feature, for example, expressed as the standard deviation or range of each saturation margin feature. These three features are then weighted and fused according to preset weights to obtain the search individual's fitness value. When the subspace isolation feature is lower than a preset isolation threshold, an orthogonal isolation constraint penalty term positively correlated with the degree of exceeding the threshold is added to the search individual's fitness value to reduce its fitness value. Then, the search individual population is sorted according to the fitness values ​​of each search individual, and the current globally optimal search individual's position vector is recorded. For each search individual's position vector, an adaptive weighting coefficient is determined based on its fitness value. Search individuals with higher fitness values ​​retain more information during amplitude updates. This adaptive weighting coefficient is used to weight and update the amplitude sub-vectors, ensuring that better solutions retain more amplitude information. Simultaneously, the differences in phase sub-vectors corresponding to adjacent isolated focusing subspaces are compared, and a phase isolation bias correction is calculated: when the difference between the phase sub-vectors of two adjacent subspaces is too small, leading to a decrease in subspace isolation, the phase isolation bias between them is increased; when the difference is too large, affecting load absorption balance, the phase isolation bias is decreased, and the phase sub-vectors are updated accordingly, resulting in an updated population of search individual position vectors.Next, the position vector of each search individual in the updated population is projected onto the feasible region satisfying the orthogonal isolation constraint according to the subspace isolation direction corresponding to the phase isolation bias correction. That is, the component that violates the isolation constraint is subtracted from the vector to obtain the corrected search individual position vector population. The search individual fitness value of each corrected search individual position vector is recalculated and the orthogonal isolation constraint penalty term is reapplied to update the current global optimal search individual position vector. The change in the search individual fitness value of the global optimal search individual position vector before and after the update is compared. If it is less than the preset convergence threshold, the current global optimal search individual position vector is output as the joint excitation vector of the transmitting coil. Otherwise, the saturation margin backoff adjustment is applied to the amplitude subvector in the updated population according to the adaptive weight coefficient of each search individual position vector. That is, the coil amplitude that causes the saturation margin to decrease is reduced by a step size, and the search individual fitness sorting and subsequent update process are returned to be executed.

[0041] This preferred embodiment, through the combined processing of population search and orthogonal isolation constraints, can further optimize the amplitude and phase configuration of the transmitting coils based on the multi-load isolation focusing subspace set. This ensures that the final output joint excitation vector of the transmitting coils guarantees the subspace isolation while taking into account the overall transmission efficiency and load absorption balance, thereby improving the global performance of multi-load power distribution.

[0042] In this embodiment, it also includes: The positions of each coil port, the equivalent spatial positions of each load, and the energy absorption state characteristics corresponding to each load in the transmitting coil array are mapped to node feature vectors respectively. The equivalent spatial position nodes, energy absorption state feature nodes, and transmitting coil nodes within the near-field coupling range of the same load are associated as hyperedges to construct a high-order spatial coupling hypergraph containing multiple hyperedges. Each hyperedge connects multiple nodes to characterize the common near-field coupling mode between multiple loads and transmitting coils. For each node in the high-order spatially coupled hypergraph, the features of neighboring nodes are aggregated along each hyperedge to which the node belongs, and the node’s own representation is updated according to the relationship between the node’s feature vector and the hyperedge, thus obtaining the spatially coupled embedding vector of each node. The spatial coupling embedding vectors of each node are weighted and pooled according to the set of transmitting coil nodes and the set of load nodes respectively to obtain the global coupling state vector of the transmitting coil array and the local coupling state vector of the load side. The global coupling state vector of the transmitting coil array and the local coupling state vector of each load are aligned according to the load number and then concatenated to form a high-order spatial coupling relationship representation of multiple loads. Using the high-order spatial coupling relationship of multiple loads at the current and historical times as input, the temporal relationship is transferred according to the correlation strength between nodes to obtain the load movement trend prediction vector corresponding to each load. When the magnitude of the load movement trend prediction vector corresponding to any load exceeds the preset movement trend threshold, the number of the isolation focus subspace to be reconstructed is determined according to the load movement trend prediction vector of the load, and the orthogonal basis vector of the corresponding isolation focus subspace in the multi-load isolation focus subspace set is pre-reconstructed. Simultaneously, the updated spatial coupling embedding vector is used to correct the load spatial scattering feature vector of the load in the time-reversal feature library, so that the transmitting coil array switches in advance to the isolated focusing subspace that matches the predicted equivalent spatial position of the load.

[0043] This embodiment primarily addresses scenarios where the load may move or undergo slight positional changes during charging. In multi-load wireless charging, the equivalent spatial position of the load is typically not completely fixed. Factors such as user device placement / removal, table vibration, or misalignment of the receiver's built-in magnetic alignment can all cause slow or sudden changes in the load's position. Once the load position changes, a spatial mismatch will occur between the equivalent spatial position determined in step S2 and the multi-load isolation focusing subspace set constructed in step S3. This causes the focus of the virtual focusing beam to deviate from the actual load position, resulting in decreased subspace isolation, reduced power distribution efficiency, and potentially even causing some loads to exit the effective charging area. Therefore, it is necessary to predict the load's movement trend in advance and pre-reconstruct the isolation focusing subspace before the load position change causes significant power loss.

[0044] Specifically, the port positions of each coil in the transmitting coil array, the equivalent spatial positions of each load, and the energy absorption state characteristics corresponding to each load are mapped to node feature vectors. The transmitting coil node feature vector contains at least the port position coordinates of the transmitting coil; the load node feature vector contains at least the equivalent spatial position coordinates and energy absorption state characteristics of the load. The equivalent spatial position nodes, energy absorption state characteristic nodes, and multiple transmitting coil nodes within the near-field coupling range of the same load are associated with a hyperedge, thus constructing a high-order spatial coupling hypergraph. A hyperedge connecting three or more nodes can express the common pattern of a load simultaneously engaging with multiple transmitting coils in the near field, rather than just pairwise relationships. Each hyperedge corresponds to a local coupling structure of a load and its associated transmitting coil.

[0045] For each node in the higher-order spatial coupling hypergraph, neighboring node features are aggregated along each hyperedge to which the node belongs. For example, a transmit coil node may belong to the hyperedges of multiple loads simultaneously. In this case, the features of the adjacent load nodes within different hyperedges and other transmit coil nodes within the same hyperedge are weighted and aggregated. The weights can be determined based on the coupling strength or spatial distance between nodes. After aggregation, the node's own representation is updated according to the association between the node's current feature vector and the hyperedge, resulting in the spatial coupling embedding vector of each node. The spatial coupling embedding vector can be understood as the implicit feature representation of a node in the higher-order spatial coupling structure, integrating its own position, energy absorption state, and coupling relationship information with neighboring nodes.

[0046] Then, the spatial coupling embedding vectors of each node are weighted and pooled separately for the sets of transmitting coil nodes and load nodes to obtain the global coupling state vector of the transmitting coil array and the local coupling state vector of the load side. Weighted pooling can be implemented by weighted summation based on the node's energy absorption state or coupling strength. The global coupling state vector of the transmitting coil array and the local coupling state vector of each load are aligned by load number and concatenated to form a high-order spatial coupling relationship representation of multiple loads. This representation simultaneously includes the global coupling characteristics of the entire transmitting coil array and the local coupling characteristics of each load. Using the high-order spatial coupling relationship representations of multiple loads at the current and historical moments as input, temporal relationship transfer is performed according to the correlation strength between nodes to obtain the load movement trend prediction vector corresponding to each load. Specifically, the load node corresponding parts in the high-order spatial coupling relationship representations of multiple consecutive moments can be arranged in chronological order, and the representations at historical moments can be weighted and transferred along the time direction using the correlation strength between nodes as the transfer weight. For example, if the correlation strength between a load and several transmitting coils has been continuously increasing in recent moments, while the correlation strength with other transmitting coils has weakened, it indicates that the load is moving towards the transmitting coil with the increased correlation. In this case, the load movement trend prediction vector will show a component pointing in that direction; the magnitude of this vector reflects the strength of the load movement trend.

[0047] When the magnitude of the load movement trend prediction vector corresponding to any load exceeds a preset movement trend threshold, it indicates that the position change of the load has reached a level requiring early intervention. At this point, the number of the isolation focusing subspace to be reconstructed is determined based on the load movement trend prediction vector, and pre-reconstruction of the orthogonal basis vectors of the corresponding isolation focusing subspace in the multi-load isolation focusing subspace set is triggered. During pre-reconstruction, the spatial phase compensation of each transmitting coil in the corresponding virtual focusing beam can be adjusted in advance according to the direction and amplitude indicated by the load movement trend prediction vector, causing the focus of the virtual focusing beam to shift towards the predicted position. Orthogonalization processing is then re-executed to ensure that the updated isolation focusing subspace remains orthogonally isolated from other subspaces.

[0048] Simultaneously, the updated spatial coupling embedding vector is used to correct the load spatial scattering feature vector of this load in the time-reversal feature library. Since the updated spatial coupling embedding vector contains the new spatial relationship between the load node and the transmitting coil node, the reflected scattering path delay and coupling amplitude changes can be mapped to updated scattering path delay and amplitude labels, replacing the original load spatial scattering feature vector in the time-reversal feature library. After correction, the transmitting coil array can switch to an isolated focusing subspace matching the predicted equivalent spatial position of the load in advance, without waiting for the load to actually move and then re-execute the complete detection and identification process.

[0049] In summary, this embodiment models the common coupling mode between the load and the transmitting coil using a high-order spatial coupling hypergraph, and predicts the load movement trend based on the time-series relationship. It can complete the pre-reconstruction of the isolation focusing subspace and the correction of the time-reversal feature library before the load position changes, thereby reducing the energy focusing offset and the decrease in subspace isolation caused by load movement, and improving the response speed and power distribution stability of the multi-load wireless charging system in dynamic usage scenarios.

[0050] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A multi-load adaptive identification and power allocation method for a wireless charging pad, characterized in that, Includes the following steps: S1. Control each coil in the transmitting coil array to sequentially transmit electromagnetic detection pulses with non-periodic time waveforms, and synchronously collect the time-domain reflection signals generated by near-field multi-load coupling at each coil port to construct a time-reflection feature library containing load spatial scattering information; S2. Perform time inversion processing on the time-domain reflection signals in the time-reflection feature library, extract the time-reflection focusing channel response corresponding to each load, and identify the number of loads and the equivalent spatial location of each load based on the degree of separation of peak clusters in the time-reflection focusing channel response. S3. Using the time-reflection focusing channel response, construct virtual focusing beams oriented towards each load, and spread the mutually orthogonal virtual focusing beams into a set of multi-load isolation focusing subspaces; S4. Within the set of multi-load isolated focusing subspaces, a composite excitation waveform consisting of amplitude perturbation components and phase perturbation components is superimposed on each subspace to excite the nonlinear magnetization harmonic response of each load receiving coil, and the energy absorption state characteristics corresponding to each load are extracted. S5. Based on the energy absorption state characteristics, adjust the projection weight of each subspace in the total power distribution, and by jointly modulating the excitation amplitude and phase of each coil in the transmitting coil array, make the transmitting power adaptively distributed among the loads according to the updated projection weight.

2. The method according to claim 1, characterized in that, Specifically, S1 is: The control coils in the transmitting coil array sequentially transmit electromagnetic detection pulses with non-periodic time waveforms to the charging area according to a preset polling sequence, and simultaneously collect the time-domain reflection signals generated by near-field multi-load coupling at each coil port to form a multi-channel time-domain reflection signal set; Each time-domain reflection signal in the multi-channel time-domain reflection signal set is synchronously intercepted and DC-free preprocessed to extract the scattered echo component caused by near-field multi-load coupling and generate a coupled scattered echo sequence corresponding to each transmitting coil. The coupled scattered echo sequences corresponding to each transmitting coil are combined according to the transmitting coil position index of each transmitting coil in the transmitting coil array, and a time inversion operator is applied to the combined coupled scattered echo sequences to obtain the time inversion characteristic response corresponding to each transmitting coil. The time inversion operator is used to enable the scattering paths from different loads in the coupled scattering echo sequence to be separably reconstructed on the time axis; Based on the arrival time distribution and spatial phase relationship of the scattered paths in the time-reflection characteristic response corresponding to each transmitting coil, the time delay label of the scattered path representing the spatial location of each load and the amplitude label of the scattered path representing the strength of load coupling are extracted to form the load spatial scattering characteristic vector. The load space scattering feature vector is associated with and stored with the corresponding transmitting coil position index to construct a time-reversal feature library containing load space scattering information.

3. The method according to claim 2, characterized in that: The aperiodic time waveform is composed of a periodic carrier component and an aperiodic envelope component. The aperiodic envelope component applies an aperiodic amplitude perturbation to the periodic carrier component in the time domain, so that the electromagnetic detection pulse generates an electromagnetic response containing multiple load scattering path information in the near field region.

4. The method according to claim 1, characterized in that, Specifically, S2 is: Retrieve the load space scattering feature vectors corresponding to the position indices of each transmitting coil from the time-reflection feature library, and rearrange the scattering path delay labels and scattering path amplitude labels in each load space scattering feature vector according to the transmitting coil position index to form a load scattering path delay-amplitude distribution matrix. Time inversion focusing processing is applied to each scattering path delay label in the load scattering path delay-amplitude distribution matrix, and the scattering path amplitude labels corresponding to each transmitting coil are superimposed along the time inversion delay direction to generate a time inversion focusing channel response characterizing the energy convergence characteristics of each load. Envelope extraction and local peak search are performed on the time-inversion focusing channel response. Local peaks that are continuous along the time delay direction and whose amplitude exceeds a preset noise floor threshold in the time inversion domain are grouped into peak clusters to form a set of peak clusters. Each peak cluster corresponds to an energy focusing region of a candidate load in the time inversion domain. Calculate the time delay interval and amplitude valley depth between adjacent peak clusters in the peak cluster set, and combine them with the time delay width of each peak cluster to obtain the peak cluster separation degree; When the peak cluster separation between adjacent peak clusters is greater than a preset separation threshold, the adjacent peak clusters are determined to belong to different loads; when the peak cluster separation between adjacent peak clusters is not greater than the preset separation threshold, the adjacent peak clusters are determined to belong to different scattering paths of the same load. The number of loads is determined based on the number of peak clusters belonging to different loads, and the equivalent spatial position of each load is inverted based on the time delay position of each peak cluster in the time-reflection focusing channel response and the spatial geometric relationship with the transmitting coil array.

5. The method according to claim 1, characterized in that, Specifically, S3 is: Based on each peak cluster in the time-reverse focusing channel response, the time delay center position and peak cluster energy envelope corresponding to each peak cluster are extracted, and the time-reverse focusing channel response segment corresponding to each peak cluster is used as the initial focusing weight vector of the load corresponding to the peak cluster. Based on the spatial distance relationship between the equivalent spatial position of each load and the position index of each transmitting coil in the transmitting coil array, spatial phase compensation is applied to each initial focusing weight vector to form a spatial focusing beam that matches the spatial position of each load. Calculate the overlap component between any two spatial focusing beams, and sequentially subtract the overlap component with other spatial focusing beams from each spatial focusing beam until the remaining overlap component is lower than a preset overlap component threshold, to obtain mutually orthogonal virtual focusing beams. Using the mutually orthogonal virtual focusing beams as basis vectors, a set of multi-load isolated focusing subspaces corresponding to the number of loads is formed; wherein, each virtual focusing beam corresponds to an isolated focusing subspace, and the isolated focusing subspaces are orthogonally isolated from each other; Calculate the subspace isolation degree between any two isolated focusing subspaces in the multi-load isolated focusing subspace set. If the isolation degree of any subspace is lower than the preset isolation degree threshold, apply orthogonal compensation correction to the virtual focusing beams corresponding to the two isolated focusing subspaces and update the multi-load isolated focusing subspace set until the isolation degree of all subspaces meets the preset isolation degree threshold.

6. The method according to claim 1, characterized in that, Specifically, S4 is: For each isolation focusing subspace in the set of multi-load isolation focusing subspaces, a composite excitation waveform formed by the superposition of amplitude perturbation components and phase perturbation components is generated; Each composite excitation waveform is applied to the corresponding isolated focusing subspace, so that each load receiving coil generates nonlinear magnetization harmonic components under the action of the focusing magnetic field after superimposed perturbation; the nonlinear magnetization harmonic components carry the nonlinear magnetization characteristics of the magnetic core of the load receiving coil. The nonlinear magnetization harmonic components of each load receiving coil are synchronously acquired at the transmitting end by the transmitting coil array, and the nonlinear magnetization harmonic response corresponding to each isolated focusing subspace is separated from the reflected coupling signal. The nonlinear magnetization harmonic responses between different isolated focusing subspaces do not crosstalk each other due to orthogonal isolation. Harmonic order decomposition is performed on each of the nonlinear magnetization harmonic responses to extract the relative amplitude distribution and phase shift characteristics of each harmonic component, and the energy absorption state characteristics of the corresponding load are constructed based on the relative amplitude distribution and phase shift characteristics. The energy absorption state characteristics of each load are associated with its isolated focusing subspace to form a set of load energy absorption state characteristics. If the similarity between any two energy absorption state features exceeds a preset similarity threshold, then a phase shift adjustment is applied to the phase perturbation component of the corresponding isolated focusing subspace and the load energy absorption state feature set is updated.

7. The method according to claim 6, characterized in that: The perturbation bandwidth of the amplitude perturbation component is smaller than that of the phase perturbation component, so as to maintain the orthogonal isolation between the isolation focusing subspaces and cause the load receiving coil to enter a nonlinear magnetization state.

8. The method according to claim 1, characterized in that, Specifically, S5 is: The energy absorption state features corresponding to each load are retrieved from the set of load energy absorption state features. The relative amplitude distribution and phase shift features in each energy absorption state feature are analyzed to extract the saturation margin feature reflecting the load's absorption capacity of the current focusing magnetic field and the coupling efficiency feature reflecting the load's utilization efficiency of the current focusing magnetic field. The saturation margin feature and the coupling efficiency feature are then fused into the load power demand weight. The power demand weights of each load are mapped to the corresponding isolation and focusing subspaces in the set of multi-load isolation and focusing subspaces to form the projection weights of each isolation and focusing subspace. The projection weights are normalized to obtain the subspace power allocation vector, where each dimension of the subspace power allocation vector corresponds to a projection weight of an isolated focusing subspace. Based on the subspace power allocation vector, and combined with the spatial correspondence between the position index of each transmitting coil in the transmitting coil array and each isolated focusing subspace, the projection weight of each isolated focusing subspace is decomposed to each transmitting coil in the transmitting coil array to form the transmitting coil excitation amplitude allocation vector. Phase constraints are applied to each element in the excitation amplitude allocation vector of the transmitting coil to keep the excitation phases of each transmitting coil corresponding to the same isolation focusing subspace synchronized, and a phase isolation bias is introduced between the excitation phases corresponding to different isolation focusing subspaces to obtain the joint excitation vector of the transmitting coils. The excitation amplitude and phase of each coil in the transmitting coil array are jointly modulated using the joint excitation vector of the transmitting coils, and the energy absorption state characteristics of each load are sampled back in real time during the modulation process. If the energy absorption state characteristics obtained after re-sampling show a decrease in saturation margin characteristics compared with those before modulation, the projection weights of the corresponding isolated focusing subspace are rolled back by a preset step size and the subspace power allocation vector is updated until the energy absorption state characteristics of each load are restored to the preset stable state, so that the transmit power is adaptively allocated among the loads according to the updated projection weights.

Images