Electric vehicle charging control method and device and electronic equipment

Abstract

The invention discloses an electric vehicle charging control method and device and electronic equipment. Relates to the technical field of electric vehicles, and the method comprises the steps: obtaining a multi-frame ORB feature image collected for a charging interface of an electric vehicle and a corresponding multi-frame point cloud feature image under the condition that the electric vehicle is detected to reach a preset charging area; performing feature fusion on the multiple frames of ORB feature images and the multiple frames of point cloud feature images to obtain multiple frames of target global feature images; determining a target pose of the charging interface based on the multiple frames of target global feature images; and the mechanical arm is controlled to plug the charging gun into the charging interface according to the target pose, the electric vehicle is charged according to a target charging strategy, and the target charging strategy is at least used for indicating the ratio of the power grid output power to the photovoltaic output power for charging the electric vehicle. The technical problems of low charging interface positioning accuracy and high charging cost in the charging process of the electric vehicle in the prior art are solved.

Description

Technical Field

[0001] This invention relates to the field of electric vehicles, and more specifically, to an electric vehicle charging control method, apparatus, and electronic device. Background Technology

[0002] In recent years, with the popularization of electric vehicles (EVs) and the development of smart grids, automated charging technology has become crucial for improving the charging experience for car owners and optimizing the allocation of power resources. Among these technologies, the automatic alignment and insertion of charging interfaces by robotic arms is a core element for achieving unattended and efficient charging. However, existing technologies face two major challenges in this process: 1) Low accuracy in charging interface positioning: Positioning solutions in related technologies often rely on single image recognition technologies or LiDAR point cloud data. However, these technologies struggle to maintain stability under complex and variable lighting conditions, especially at night or in rainy weather. Furthermore, while point cloud data can provide three-dimensional spatial positioning information, it is easily affected by non-rigid objects in the environment, leading to decreased positioning accuracy and a lower success rate for robotic arm insertion. 2) High charging costs: Current charging strategies often fail to fully utilize distributed energy sources, such as solar and wind power. Using a fixed ratio of grid power to photovoltaic power for charging without considering immediate power generation conditions and grid load not only increases charging costs but may also burden the grid and affect power supply stability. Especially during peak electricity consumption periods, blindly using grid power for charging will further exacerbate the imbalance between electricity supply and demand, hindering the promotion and application of green energy. In summary, related technologies for electric vehicle charging suffer from problems such as low accuracy in locating charging interfaces and high charging costs.

[0003] There is currently no effective solution to the above problems. Summary of the Invention

[0004] This invention provides an electric vehicle charging control method, device, and electronic device to at least solve the technical problems of low accuracy in positioning the charging interface and high charging cost in the electric vehicle charging process.

[0005] According to one aspect of the present invention, an electric vehicle charging control method is provided, comprising: when an electric vehicle is detected to have reached a preset charging area, acquiring multi-frame directional acceleration robust feature and rotational binary robust independent feature (ORB) feature images and corresponding multi-frame point cloud feature images collected from the charging interface of the electric vehicle, wherein the multi-frame ORB feature images and the multi-frame point cloud feature images correspond one-to-one; performing feature fusion on the multi-frame ORB feature images and the multi-frame point cloud feature images to obtain multi-frame target global feature images; determining the target pose of the charging interface based on the multi-frame target global feature images; controlling a robotic arm to insert a charging gun into the charging interface according to the target pose, and charging the electric vehicle according to a target charging strategy, wherein the target charging strategy is at least used to indicate the ratio of grid output power and photovoltaic output power for charging the electric vehicle.

[0006] According to another aspect of the present invention, an electric vehicle charging control device is also provided, comprising: a feature image module, configured to acquire multi-frame directional acceleration robust feature and rotational binary robust independent feature (ORB) feature images and corresponding multi-frame point cloud feature images of the charging interface of the electric vehicle when the electric vehicle is detected to have reached a preset charging area; a feature fusion module, configured to perform feature fusion on the multi-frame ORB feature images and the multi-frame point cloud feature images to obtain multi-frame target global feature images; a pose determination module, configured to determine the target pose of the charging interface based on the multi-frame target global feature images; and a charging control module, configured to control a robotic arm to insert a charging gun into the charging interface according to the target pose, and to charge the electric vehicle according to a target charging strategy, wherein the target charging strategy is at least used to indicate the ratio of grid output power and photovoltaic output power for charging the electric vehicle.

[0007] According to another aspect of the present invention, a non-volatile storage medium is also provided, which stores a plurality of instructions adapted for loading and executing any one of the instructions in an electric vehicle charging control method by a processor.

[0008] According to another aspect of the present invention, an electronic device is also provided, including one or more processors and a memory, the memory being used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors cause the one or more processors to implement any one of the electric vehicle charging control methods.

[0009] According to another aspect of the present invention, a computer program product is also provided, including a computer program that, when executed by a processor, implements the steps of any one of the electric vehicle charging control methods.

[0010] In this embodiment of the invention, when an electric vehicle is detected to have reached a preset charging area, multi-frame directional acceleration robust feature and rotational binary robust independent feature (ORB) image of the charging interface of the electric vehicle are acquired, along with corresponding multi-frame point cloud feature images. The multi-frame ORB feature images and multi-frame point cloud feature images correspond one-to-one. Feature fusion is performed on the multi-frame ORB feature images and multi-frame point cloud feature images to obtain multi-frame target global feature images. Based on the multi-frame target global feature images, the target pose of the charging interface is determined. A robotic arm is controlled to insert the charging gun into the charging interface according to the target pose, charging the electric vehicle using a target charging strategy. The target charging strategy at least indicates the ratio of grid output power and photovoltaic output power for charging the electric vehicle. This achieves the goal of determining the precise location of the charging interface by fusing multi-frame ORB image features and point cloud features, thereby controlling the robotic arm to achieve precise automatic gun insertion and dynamically adjusting the charging power ratio based on a photovoltaic-storage synergy strategy. This results in high precision, high efficiency, and intelligence in the charging process, significantly improving the electric vehicle charging experience and energy utilization efficiency. Furthermore, it solves the technical problems of low accuracy in charging interface positioning and high charging costs in related technologies during electric vehicle charging. Attached Figure Description

[0011] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0012] Figure 1 This is a flowchart of an electric vehicle charging control method according to an embodiment of the present invention;

[0013] Figure 2 This is an optional charging interface positioning flowchart according to an embodiment of the present invention;

[0014] Figure 3 This is an optional robotic arm insertion control flowchart according to an embodiment of the present invention;

[0015] Figure 4 This is an optional energy scheduling flowchart based on photovoltaic-storage coordinated power supply according to an embodiment of the present invention;

[0016] Figure 5 This is an optional fault decision-making flowchart according to an embodiment of the present invention;

[0017] Figure 6 This is a front view of an optional electric vehicle charging compartment according to an embodiment of the present invention;

[0018] Figure 7 This is a schematic diagram of a three-dimensional model of the charging compartment according to an embodiment of the present invention;

[0019] Figure 8 This is a flowchart of an optional electric vehicle charging control method according to an embodiment of the present invention;

[0020] Figure 9 This is a schematic diagram of an electric vehicle charging control device according to an embodiment of the present invention. Detailed Implementation

[0021] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0022] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0023] According to an embodiment of the present invention, a method embodiment for electric vehicle charging control is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0024] Figure 1 This is a flowchart of an electric vehicle charging control method according to an embodiment of the present invention, such as... Figure 1 As shown, the method includes the following steps:

[0025] Step S102: When the electric vehicle is detected to have reached the preset charging area, acquire multi-frame Oriented Fast and Rotated BRIEF (ORB) feature images collected from the charging interface of the electric vehicle, as well as corresponding multi-frame point cloud feature images, wherein the multi-frame ORB feature images and the multi-frame point cloud feature images correspond one-to-one.

[0026] Optionally, when an electric vehicle enters a pre-defined charging area, sensors (such as cameras, radar, and geomagnetic sensors) identify and confirm the vehicle's arrival. This ensures that the automated charging process only starts in the correct location and under the correct conditions, avoiding misoperation or unauthorized access. ORB is a highly efficient feature point detection and description algorithm that can quickly locate key feature points in an image and remain stable even under image rotation and lighting changes. Images of the electric vehicle charging interface area are continuously acquired by a camera. Each frame of the image is processed by the ORB algorithm to extract directional acceleration robust features and rotation binary robust independent features, generating multi-frame ORB feature images. Point cloud features can be acquired by 3D LiDAR or depth cameras, providing three-dimensional spatial information of the charging interface and its surrounding environment. The multi-frame point cloud feature images, each corresponding one-to-one with the ORB feature images, can comprehensively capture the precise position and orientation of the charging interface at different times, compensating for the lack of depth information in two-dimensional images.

[0027] Optionally, before executing step S102, charging preparation and coarse positioning of the electric vehicle are also required. In the charging preparation phase, vehicle entry and identity verification are performed first. The vehicle drives into the preset charging space, ensuring the charging port faces the robot's working area. The user scans the charging pile's QR code via a mobile app to initiate the charging service. The system verifies the user's identity and payment information, generating a charging task instruction. Environmental perception and safety checks are then performed. Environmental sensors on the top of the charging compartment detect the surrounding safety conditions, the roller shutter automatically opens, and the camera performs a preliminary scan of the vehicle's charging port area. The system checks the internal environment of the charging compartment to ensure there are no personnel or obstacles. Finally, equipment self-checks and initialization are performed. The robotic arm control system performs a self-check to confirm normal joint movement; the vision system is calibrated to ensure accurate camera parameters; and the photovoltaic power generation status and battery SOC status are checked by the photovoltaic energy storage system. The coarse localization stage involves the following steps: Charging port area detection: A high-definition camera (2 megapixels, such as a stereo camera) captures a wide-angle shot of the vehicle's charging port area; ORB feature point extraction algorithm is used to identify the approximate location of the charging port; combined with laser point cloud semantic segmentation, the 3D spatial information of the charging port is obtained; initial coordinates are calculated, with positioning accuracy controlled within ±10mm. Initial robotic arm localization: The robotic arm (6 degrees of freedom) moves from its standby position to near the charging port; based on the coarse localization results, the end effector of the robotic arm reaches a safe position 50mm from the charging port; the robotic arm's posture is adjusted to ensure the camera is directly facing the charging port.

[0028] Step S104: Perform feature fusion on multiple frames of ORB feature images and multiple frames of point cloud feature images to obtain multiple frames of global target feature images.

[0029] Optionally, feature fusion refers to the comprehensive analysis of multiple frames of ORB feature images and corresponding point cloud feature images. Algorithms (such as deep learning models and point cloud registration techniques) are used to combine two-dimensional image features with three-dimensional point cloud features, generating a multi-frame global feature image of the target that includes information on the charging interface's location, orientation, and environmental context. This step enhances the comprehensive understanding of the charging interface, providing a solid data foundation for subsequent precise positioning and robotic arm control.

[0030] In one optional embodiment, feature fusion is performed on multiple frames of ORB feature images and multiple frames of point cloud feature images to obtain multiple frames of target global feature images, including: performing feature point matching on multiple frames of ORB feature images and multiple frames of point cloud feature images to obtain feature point matching results; based on the feature point matching results, performing feature fusion on multiple frames of ORB feature images and corresponding point cloud feature images to obtain multiple frames of initial global feature images; and performing Kalman filtering on multiple frames of initial global feature images to obtain multiple frames of target global feature images.

[0031] Optionally, feature point matching aims to find corresponding key points in the ORB feature image and point cloud feature image, establishing a correlation between image features and spatial features, providing a matching basis for subsequent fusion. Based on the feature point matching results, multiple frames of ORB feature images are fused with the corresponding point cloud feature images to generate multiple frames of initial global feature images. The essence of fusion is to combine the orientation and rotation invariant features in the two-dimensional image with the depth and position information of the three-dimensional point cloud to construct a more comprehensive and three-dimensional target representation. This fusion not only enhances the recognition capability of the charging interface but also effectively handles the uncertainties caused by occlusion and environmental changes, improving the robustness and accuracy of localization. Performing Kalman filtering on the fused multiple frames of initial global feature images is a key step to further improve feature stability and prediction accuracy. The Kalman filter is an optimal recursive filter for dynamic system state estimation. It can predict the future state of the system based on current and past state information under noise interference and use actual measurement data for correction, thereby obtaining a more accurate estimate. In this embodiment, a Kalman filter is used to process multi-frame feature images. Through time series analysis, accidental measurement errors and environmental noise are filtered out, the feature sequence is smoothed, and finally, multi-frame global feature images of the target are obtained. These images not only contain the instantaneous position and orientation of the charging interface, but also take into account its motion trend and environmental changes, providing reliable data support for the precise control of the robotic arm.

[0032] This embodiment utilizes feature fusion and Kalman filtering techniques based on multi-frame ORB feature images and point cloud feature images to improve the accuracy of charging interface positioning. Feature point matching ensures the correspondence between the two-dimensional image and three-dimensional spatial information. Feature fusion generates an initial global feature image containing orientation, rotation, depth, and position information, while Kalman filtering further eliminates noise and improves the temporal stability of the feature image. The combined application of these techniques significantly improves the positioning accuracy and efficiency during the automated charging process of the robotic arm.

[0033] Optionally, during the precise positioning phase of the electric vehicle, multi-scale feature fusion (i.e., multi-source data fusion) is employed. The camera performs a second high-resolution capture to obtain detailed images of the charging port. Corner detection using a scale pyramid + Accelerated Segment Test (FAST) method is then employed to extract multi-scale feature points for feature matching. Feature matching can be performed, but is not limited to, through the following methods:

[0034] Where M represents the optimal matching matrix / transformation relationship; ORB_i represents the i-th ORB feature point in the ORB feature image; Cloud_j represents the j-th point cloud feature point in the point cloud feature image; depth_i represents the depth information of the image feature; and z_j represents the z-coordinate (depth) of the point cloud feature. This represents the weighting coefficient, used to balance the errors of the two items.

[0035] Optionally, during the Kalman filtering process, the state vector is set. In the position dimension, x, y, and z represent three-dimensional spatial coordinates; in the attitude dimension, These represent roll, pitch, and yaw angles (Euler angles), respectively; in the online velocity dimension, These represent the linear velocities along the x, y, and z axes of the body coordinate system, respectively; and the angular velocity dimension. These represent the angular velocities along the x, y, and z axes of the body coordinate system, respectively; in the control input dimension, u1, u2, and u3 represent the control forces / torques (usually motor thrust, etc.) along the x, y, and z axes, respectively. Set the observation equations. ,in, This represents the observation vector (measurement value) at the current time t. This represents the system state vector at time t. This represents the observation matrix (the mapping relationship between states and observations). This represents the observation noise (usually assumed to be Gaussian white noise). The prediction equation is then defined. ,in, Let F represent the state vector at time t, and let F represent the state transition matrix. B represents the state vector at the current moment; B represents the control input matrix. This represents the control input vector at the current moment.

[0036] Step S106: Determine the target pose of the charging interface based on multi-frame global feature images of the target.

[0037] Optionally, based on multi-frame global feature images of the target, a corresponding pose estimation algorithm can be used to determine the exact position and orientation of the charging interface in space. Accurate determination of the target pose is crucial for ensuring the robotic arm successfully connects the charging gun and avoids mechanical damage.

[0038] In one optional embodiment, determining the target pose of the charging interface based on multiple frames of global feature images of the target includes: performing sub-pixel level corner optimization processing on each of the multiple frames of global feature images of the target to obtain the target corner coordinates corresponding to each of the multiple frames of global feature images of the target; estimating the pose of the charging interface based on the target corner coordinates corresponding to each of the multiple frames of global feature images of the target to obtain the initial pose corresponding to each of the multiple frames of global feature images of the target; and determining the target pose based on the initial pose corresponding to each of the multiple frames of global feature images of the target.

[0039] Optionally, in determining the target pose of the charging interface based on multiple frames of global feature images, sub-pixel corner optimization processing needs to be performed on each frame of feature images first. Corner optimization is a technique used in image processing to improve the accuracy of feature point localization. After detecting preliminary feature points, the gray-level changes in the neighborhood of the feature points are further analyzed, and the sub-pixel point with the largest gray-level change is found through fitting or interpolation methods, which is then used as a more accurate corner point position. In automated electric vehicle charging scenarios, the corner coordinates of the charging interface are key information for the robotic arm's positioning and charging gun insertion. Sub-pixel accuracy can significantly improve the accuracy and reliability of the robotic arm's operation. After sub-pixel corner optimization processing, each frame of global feature images of the target obtains the precise coordinate information of the target corner points of the charging interface in that frame. These coordinates not only contain the positional information in the two-dimensional image but also combine the depth information of the point cloud features, forming a point set in a three-dimensional spatial coordinate system to characterize the accurate position and attitude of the charging interface at different time points. Based on the target corner coordinates of each of the multiple frames of global feature images of the target, the system uses a pose estimation algorithm to perform initial pose estimation. Pose estimation is the process of determining the position and orientation of a target object in three-dimensional space, which is crucial for automated robotic arm operations. By analyzing the changes in corner coordinates between consecutive frames, the motion trajectory of the charging interface in space can be calculated, obtaining the initial pose of the charging interface for each frame, including rotation angle and displacement vector. After obtaining the initial poses corresponding to each of the multiple frames of global feature images of the target, it is necessary to further integrate this pose information to determine a final, optimized target pose.

[0040] This embodiment presents a technical process for determining the target pose of the charging interface based on multi-frame global feature images of the target. This process covers the entire process from sub-pixel-level corner optimization to initial pose estimation, and finally to target pose determination. This series of processing steps ensures high-precision positioning of the charging interface by the robotic arm during automatic charging.

[0041] Optionally, subpixel-level optimization can be performed on corner points in the initial global feature images across multiple frames to improve localization accuracy. Subpixel offsets of corner points are calculated using a quadratic surface fitting method to obtain more precise corner coordinates. Quadratic surface fitting is performed through neighborhood gradient analysis, and extreme point localization is performed based on the obtained quadratic surface. Subpixel coordinates are then determined based on the extreme point localization results. Quadratic surface fitting can be performed, but is not limited to, in the following manner: I(x,y) = ax² + bxy + cy² + dx + ey + f, where I(x,y) represents the pixel intensity function, used to describe the brightness distribution of local areas of the image; a, b, c, d, e, and f represent the quadratic surface coefficients, which are parameters obtained through least squares fitting; x and y are local coordinates, i.e., relative coordinates centered on the initial corner point; the subpixel offset can be calculated as follows: , ,in, , These represent the sub-pixel offsets in the x-axis and y-axis directions, respectively, characterizing the correction amount from the integer pixel position 4a to the precise sub-pixel position. The discriminant det = b²-c can be set to ensure the quadratic surface has extrema. Table 1 shows the performance improvement after sub-pixel optimization.

[0042] Table 1

[0043]

[0044] In one optional embodiment, determining the target pose based on the initial poses corresponding to each of the multiple frames of global feature images of the target includes: determining the weight values ​​corresponding to each of the multiple frames of global feature images of the target; and performing a weighted average operation based on the initial poses corresponding to each of the multiple frames of global feature images of the target and their corresponding weight values ​​to obtain the target pose.

[0045] Optionally, in determining the target pose of the charging interface based on multiple frames of global feature images, the weight value of each frame is determined according to its reliability or information content. The weight value can be determined based on, but is not limited to, the following factors: Image quality: including lighting conditions, contrast, sharpness, etc. Higher quality images provide more reliable information and thus have a larger weight value. Temporal relevance: Newer images may reflect the most recent state of the charging interface, therefore they are given higher weight when fusing pose information. Environmental changes: If a frame was captured during a period of minimal environmental change, its information is more stable and should be given a higher weight. Number and quality of feature points: The more feature points and the more concentrated they are, the greater the information content of the image, and the higher the weight. Based on the initial pose and weight values ​​of each of the multiple frames of global feature images, a weighted average is performed. The initial pose of each frame is multiplied by its corresponding weight value, then summed, and finally divided by the total weight value to obtain an optimized target pose that comprehensively considers information from all frames. Weighted operations can effectively fuse pose information from different time points, and adjust the contribution of each frame in the final pose calculation according to the credibility of each frame, thereby improving the accuracy and robustness of pose estimation.

[0046] Optionally, but not limited to, the weight values ​​of the global feature image of the target in each frame can be obtained in the following ways: ,in, The weight value represents the global feature image of the target in the j-th frame, which is used as the weighting coefficient for multi-frame fusion. Represents reprojection error, used to characterize the difference between projecting 3D points onto the image plane and the measured points (unit: pixels). This represents the balance coefficient, used as a parameter to adjust the time decay and error weights; the default value is 0.2. Represents the time decay factor, where , where n is the total number of frames in the global feature image of the target across multiple frames, and j is the index of the current frame.

[0047] Optionally, to avoid feature misidentification due to insufficient lighting, illumination compensation is performed on the feature images before feature fusion of multiple ORB feature images and multiple point cloud feature images. The specific process is as follows: An ambient light sensor is used to monitor the ambient light intensity around the camera in real time, obtaining the illumination value at the time each feature image was captured. The monitored illumination values ​​are classified and evaluated to determine whether the lighting conditions meet the requirements for feature recognition. For example, a threshold is set; when the illumination intensity is below a certain threshold (e.g., below 100 lux), the current ambient light is considered insufficient, and the illumination compensation process needs to be initiated. Based on the illumination intensity, the current lighting type is identified, such as strong light, weak light, or uneven light. For different lighting conditions, an appropriate supplementary lighting algorithm is selected. For example, for strong light, dynamic range compression is used; for weak light, nonlinear enhancement supplementary lighting is used; for uneven light, illumination equalization techniques, such as the Retinex algorithm, are used to remove the influence of dark areas and suppress overly bright parts in the image, maintaining clear details. After illumination compensation, gamma correction is applied to the image to adjust its nonlinear response, making the visual effect closer to human perception, especially enriching details in dark areas. This avoids over-brightening or underexposure introduced during compensation, ensuring clear and discernible edges and texture information of feature points. After illumination compensation and gamma correction on multiple frames of ORB feature images and point cloud feature images, feature fusion is performed. Combining the illumination-optimized image with point cloud data improves the robustness and accuracy of feature point detection, maintaining good feature recognition even in complex lighting environments or low-light conditions. The compensated image is compared with the original image to verify the illumination compensation effect. If necessary, the illumination compensation algorithm and gamma correction values ​​are adjusted to ensure the feature point recognition quality reaches the expected level. Table 2 shows the comparison of the number of feature points identified and the localization error under illumination compensation.

[0048] Table 2

[0049]

[0050] Through the above-described illumination compensation process, the method in this embodiment can effectively overcome the feature recognition difficulties caused by insufficient illumination or complex illumination conditions, ensuring accurate ORB feature point detection and point cloud fusion under any illumination environment, thereby improving the overall performance and success rate of automated charging control.

[0051] Figure 2 This is an optional charging interface positioning flowchart according to an embodiment of the present invention, such as... Figure 2As shown, the method includes: acquiring images of the charger interface using a binocular camera to obtain multiple initial images; scanning the point cloud near the charging interface using a LiDAR to obtain point cloud feature images, resulting in multiple frames of point cloud feature images; performing image distortion correction on the acquired multiple initial images and multiple frames of point cloud feature images, and performing ORB feature detection on the corrected multiple initial images after correction to obtain multiple frames of ORB feature images; sequentially performing disparity map calculation and feature fusion on the corrected multiple frames of point cloud feature images and multiple frames of ORB feature images to obtain multiple initial global feature images; filtering the multiple frames of target global feature images to obtain multiple frames of target global feature images; performing 6D pose calculation and accuracy verification based on the multiple frames of target global feature images, and outputting a transformation matrix if the accuracy verification passes, thereby obtaining the target pose of the robotic arm; otherwise, repositioning the charging interface is performed.

[0052] Step S108: Control the robotic arm to plug the charging gun into the charging interface according to the target pose, and charge the electric vehicle with the target charging strategy, wherein the target charging strategy is used to indicate at least the ratio of grid output power and photovoltaic output power for charging the electric vehicle.

[0053] Optionally, after obtaining the target pose of the charging interface, the robotic arm is controlled to insert the charging gun into the electric vehicle's charging interface with a precise path and angle. This step requires high-precision motion control and path planning to ensure the safety and reliability of the automatic charging process. The target charging strategy refers to dynamically adjusting the ratio of grid output power and photovoltaic output power based on the current grid status, photovoltaic system power generation, and the electric vehicle's charging needs to charge the electric vehicle with optimal efficiency. This real-time adjustment of the strategy can effectively utilize green energy, reduce charging costs, and reduce pressure on the grid. This embodiment's method starts with detecting the vehicle's arrival at the charging area. Through the acquisition, fusion, and analysis of multimodal features (images and point clouds), it achieves high-precision positioning of the electric vehicle's charging interface, providing accurate data support for the robotic arm's automated insertion. Simultaneously, based on real-time energy status and demand, the charging strategy is dynamically adjusted, achieving efficient utilization of green energy and optimization of charging costs.

[0054] In one optional embodiment, controlling the robotic arm to insert the charging gun into the charging interface according to the target pose includes: during the process of controlling the robotic arm to insert the charging gun into the charging interface according to the target pose, acquiring the force deviation when the charging gun contacts the charging interface, wherein the force deviation represents the difference between the contact force when the charging gun contacts the charging interface and the expected contact force; determining the position adjustment amount of the robotic arm based on the force deviation; and adjusting the position of the robotic arm according to the position adjustment amount until the charging gun is successfully inserted into the charging interface.

[0055] Optionally, during the insertion process, when the robotic arm's end effector (i.e., the charging gun) contacts the charging interface, a six-axis torque sensor monitors the changes in contact force in real time. Contact force refers to the actual force applied by the charging gun to the charging interface in the Z-axis direction (i.e., the insertion direction). The desired contact force is a pre-set ideal contact force value to ensure the stability and safety of the insertion process. Force deviation is the difference between the actual contact force and the desired contact force, reflecting the deviation in the contact state between the charging gun and the charging interface, and is a key parameter for achieving compliant insertion. Once a force deviation is detected, the corresponding robotic arm position adjustment is immediately calculated based on the admittance control model. Admittance control is a compliant control strategy that converts force deviation into a position correction to ensure that the robotic arm can respond to real-time changes in contact force, adjusting the end effector position to achieve stable and precise insertion. Based on the calculated position adjustment, the robotic arm control system adjusts the robotic arm position in real time until the contact force between the charging gun and the charging interface reaches a stable state and the position deviation is less than a preset tolerance range. This adjustment process is continuous until the charging gun is fully inserted into the charging interface and the conditions for stable contact are met: the contact force is stable within the desired range (e.g., 10N~12N), the positional deviation is less than 0.5mm, and the electrical connection resistance is less than 30 milliohms. Once these conditions are met, the insertion is confirmed successful, and the electronic latch is activated to ensure mechanical locking between the charging gun and the vehicle interface, initiating the charging process. This dynamic adjustment mechanism for force-controlled insertion of the charging gun ensures precise positioning and contact force control during insertion, avoiding impacts caused by force deviations, improving the reliability and safety of the insertion, and demonstrating the system's intelligent control over the robotic arm's insertion actions.

[0056] In one optional embodiment, determining the position adjustment amount of the robotic arm based on the force deviation includes: determining the admittance coefficient, wherein the admittance coefficient represents the rate of change of displacement of the charging gun in the contact direction under a unit force; and multiplying the admittance coefficient and the force deviation to obtain the position adjustment amount.

[0057] Optionally, the admittance coefficient is a physical quantity characterizing the displacement change characteristics of the end effector of the robotic arm (in this case, the charging gun) under force, specifically reflecting the rate of displacement change of the charging gun in the contact direction (Z-axis) under a unit force. That is, if the charging gun encounters resistance or the contact force deviates from the preset desired contact force during insertion, the admittance coefficient will determine the degree of displacement adjustment of the charging gun in response to this force deviation. In this embodiment, the preferred value of the admittance coefficient is 0.05 mm / N, meaning that for every Newton of force deviation, the displacement of the charging gun along the Z-axis will be adjusted by 0.05 mm. During insertion, a six-axis torque sensor monitors the actual contact force when the charging gun contacts the charging interface in real time. When there is a deviation between the actual contact force and the desired contact force, the position adjustment amount of the robotic arm is determined by multiplying the force deviation by the admittance coefficient. This position adjustment amount guides the robotic arm to fine-tune the position of the end effector in real time to minimize the force deviation and ensure stable contact between the charging gun and the charging interface. Based on the calculated position adjustment amount, the robotic arm control system adjusts the posture and position of the robotic arm in real time. This adjustment process is continuous until the contact force between the charging gun and the charging interface reaches a stable state and the positional deviation is less than a predefined tolerance (e.g., 0.5mm). At this point, the insertion process is considered complete. Subsequent operations, such as electronic lock activation, can then be performed to ensure mechanical locking between the charging gun and the vehicle interface, preparing for charging to begin. Through this method, the embodiment achieves precise positioning and force-controlled insertion of the robotic arm during charging, significantly improving the accuracy and safety of the insertion, while reducing the demands on the operator and enhancing the user experience. The admittance control mechanism enables the system to respond quickly and gently to external interference or force deviations, effectively avoiding impacts during insertion, extending the lifespan of the robotic arm and charging gun, and ensuring the integrity of the charging interface.

[0058] Optionally, the motion equation of the robotic arm end effector, i.e., the charging gun, can be set to the following form: Where M represents the virtual mass of the charging gun, which serves as the inertial parameter in admittance control, and can be set to 0.5 kg for example; B represents the virtual damping, which serves as the damping parameter in admittance control, with a typical value of 80 N / s per meter; and K represents the virtual stiffness, which serves as the stiffness parameter in admittance control, with a typical value of 2000 N / m. , , These represent the acceleration, velocity, and positional deviation of the charging gun, respectively, as the difference between the expected and actual values; This indicates the force deviation when the charging gun contacts the charging interface. Based on this force deviation, the position adjustment amount of the robotic arm is determined. For example, the position adjustment can be simplified as follows: ,in, This indicates the position adjustment amount in the Z-axis direction. This represents the admittance coefficient (unit: mm / N), indicating the conversion ratio between force and displacement; while This refers to the Z-axis force deviation, specifically the force component in the Z-axis direction. Using this formula, we can calculate how many millimeters of adjustment the robotic arm needs in the Z-axis direction to eliminate the force deviation and ensure a smooth and precise insertion process.

[0059] Optionally, during the process of the robotic arm inserting the charging gun into the charging interface according to the target pose, a dynamic deceleration insertion method is adopted. For example, when the charging gun is in the long-distance stage, that is, far from the charging interface (e.g., >30mm), the robotic arm is controlled to move towards the charging interface at a speed of 0.2m / s; when the charging gun is in the deceleration range, that is, at a moderate distance from the charging interface (e.g., 10-30mm), the robotic arm speed is controlled to linearly decrease from 0.2m / s to 0.1m / s; when the charging gun is in the fine insertion stage, that is, at a close distance from the charging interface (e.g., <10mm), the robotic arm is controlled to move towards the charging interface at a speed of 0.1m / s, with the acceleration limited to 0.5m / s².

[0060] Optionally, to ensure the safe insertion and charging of the charging gun, appropriate safety checks are required during the insertion process. First, as the robotic arm inserts the charging gun into the charging interface according to the target pose, the insertion resistance is checked to see if it exceeds 100N. If so, the robotic arm is stopped immediately, and an alarm is issued. If the insertion resistance is less than or equal to 100N, the insertion depth is checked to see if it is less than 20mm. If so, the precise positioning and position correction of the charging gun continue. If the insertion depth reaches 20mm, the insertion is confirmed successful. The charging gun temperature is then checked to see if it exceeds a preset temperature (e.g., 50 degrees Celsius). If so, the electric vehicle is charged at reduced power; otherwise, the electric vehicle is charged normally based on a predetermined charging strategy (e.g., a target charging strategy).

[0061] Optional, Figure 3 This is an optional robotic arm insertion control flowchart according to an embodiment of the present invention, such as... Figure 3 As shown, the method includes: in the initialization phase, setting the desired contact force to 12 Newtons; defining the position tolerance as ±0.5 mm; determining the admittance coefficient as 0.05 mm / Newton; and setting the control cycle to 10 milliseconds. During the process of controlling the robotic arm to insert the charging gun into the charging interface according to the target pose, the actual contact force is read from the torque sensor, including a three-dimensional force vector; the current position and pose of the robotic arm are obtained. The contact force F_actual when the charging gun contacts the charging interface is obtained, and the force deviation when the charging gun contacts the charging interface is determined through the admittance model; the position adjustment of the charging gun in the Z-axis direction is determined based on the force deviation. The system determines whether the release conditions between the charging gun and the charging interface are met. If the actual contact force is greater than 10 Newtons and the positional deviation is within the tolerance range, stable contact is considered achieved. At this point, the electronic latch is activated, mechanically locking the charging gun to the vehicle interface and initiating charging of the electric vehicle with a high current of 500 amps. The main cycle is then exited, and the insertion process is complete. If the contact condition is not met—that is, the actual contact force is insufficient or the positional deviation exceeds the tolerance—the position of the robotic arm in the Z-axis direction is adjusted according to the calculated position adjustment amount (ΔZ) to ensure the charging gun gradually approaches the contact point. This process is repeated every 10 milliseconds until the contact condition is met. This embodiment ensures a smooth and stable insertion process by adjusting the position in real time to respond to changes in contact force, avoiding hard contact or impact. Position adjustment based on force deviation allows the robotic arm to react to real-time changes in contact force, improving the accuracy and reliability of the insertion. Once the contact force reaches the expected stable state and the positional deviation is within the set tolerance, the insertion is considered complete, and the electronic latch can be safely activated to begin charging.

[0062] It should be noted that the robot vision positioning system in this embodiment integrates sophisticated hardware configuration and highly intelligent positioning algorithms, aiming to achieve accurate identification of the charging port and precise control of the robotic arm. The system mainly consists of a high-performance robotic arm with 6 degrees of freedom, a high-definition camera with 2 megapixels, and a magnetic gripper mechanism, working together to complete a series of automated operations from vehicle identification to charging gun insertion. In the positioning process, the first stage is coarse positioning. The high-definition camera quickly scans the vehicle's charging port area, and combined with advanced image analysis technology, the system can quickly calculate the initial coordinates of the charging port. Although the positioning accuracy has an error range of ±10 mm at this stage, this step lays the foundation for subsequent precise positioning. Subsequently, the system enters the fine positioning stage. In this stage, the camera takes a second, higher-resolution photograph, and through complex image processing algorithms, the position of the charging port is accurately and dynamically corrected. Using techniques such as ORB feature point extraction, scale pyramid, and FAST corner detection, combined with laser point cloud semantic segmentation, the system's positioning accuracy is significantly improved to ±0.05 mm, approximately 20 times higher than traditional positioning techniques. Finally, based on the precise positioning results, the charging gun insertion control phase begins. During this phase, the robotic arm smoothly moves from its initial position to the charging port according to the planned path, its speed adaptively adjusting based on distance. Particularly in the 10-30 mm distance range, the robotic arm's movement speed automatically decreases by 50% to increase operational precision, further reduce positioning errors, and ensure the charging gun can be accurately and safely inserted into the charging port. Throughout the insertion process, the robotic arm maintains coordinated posture and movement, achieving seamless docking between the charging gun and the interface through sub-pixel corner optimization and dynamic Kalman filtering technologies. In summary, the robot vision positioning system in this embodiment, with its superior hardware configuration and efficient algorithm flow, ensures accurate identification and positioning of the vehicle's charging port throughout the charging process. Furthermore, through precise robotic arm control, it achieves automatic charging gun insertion, significantly improving the efficiency of the charging service and the user experience.

[0063] In an optional embodiment, before controlling the robotic arm to insert the charging gun into the charging interface according to the target pose and charging the electric vehicle according to the target charging strategy, the method further includes: determining the objective function as: Where J represents the value of the objective function, and min represents minimizing. and Here, C_grid represents the real-time electricity price, P_grid represents the grid output power for charging electric vehicles, and SOC_ref represents the reference state of charge of the electric vehicle. The real-time state of charge of the electric vehicle is determined; the constraints include: the power demand of the electric vehicle is equal to the sum of the grid output power for charging the electric vehicle and the photovoltaic output power; based on the objective function and constraints, the initial charging strategy of the electric vehicle is optimized to obtain the target charging strategy.

[0064] Optionally, the objective function is a multi-objective optimization function, where J represents the function value of the objective function, that is, the final result that the system attempts to pursue by optimizing the charging strategy. and , which is a weighting coefficient used to adjust the relative importance of the two optimization objectives. In cases of significant increases, greater emphasis will be placed on reducing the cost of purchasing electricity from the power grid. In larger scenarios, the focus shifts to maintaining the State of Charge (SOC) within the ideal range to ensure battery health. C_grid represents the real-time grid electricity price, i.e., the cost of obtaining electricity from the grid. P_grid represents the power output from the grid when charging the electric vehicle. SOC_ref represents the reference state of charge of the electric vehicle, the ideal battery charge level the system aims to maintain to ensure battery life and meet daily usage needs. SOC represents the real-time state of charge of the electric vehicle, i.e., the current battery charge level. Minimizing the objective function means finding the optimal charging strategy that balances charging costs and battery SOC fluctuations, ensuring that the charging needs of the electric vehicle are met while reducing the cost of purchasing electricity from the grid and bringing the battery SOC as close as possible to its reference value. The constraint ensures that the fundamental principle must be followed when optimizing the objective function: the power demand of the electric vehicle must equal the sum of the power output from the grid and the photovoltaic output power P_pv (P_demand = P_grid + P_pv). This constraint ensures energy conservation during the charging process, allowing the charging needs of the electric vehicle to be met by adjusting the ratio of grid power to photovoltaic power under any illumination conditions. Based on the aforementioned objective function and constraints, the initial charging strategy is optimized by dynamically adjusting the allocation of P_grid and P_pv. This optimization process can, but is not limited to, using methods such as genetic algorithms, particle swarm optimization, or gradient descent to search for the optimal solution. The goal of the optimization is to find a set of values ​​for P_grid and P_pv that minimizes the objective function J while satisfying all the set constraints, ensuring the economic efficiency of the electric vehicle charging process and the efficient utilization of photovoltaic energy.

[0065] Through the optimization mechanism of this embodiment, not only can the photovoltaic and energy storage system be intelligently scheduled and the power supply ratio of photovoltaic and grid be rationally allocated, but the power supply strategy can also be dynamically adjusted under constantly changing sunlight and electricity price conditions, which can significantly improve the flexibility and economic benefits of the charging system, while also better protecting the batteries of electric vehicles and extending their service life.

[0066] Optionally, in the photovoltaic-storage synergistic electric vehicle charging control of this embodiment, efficient synergy between photovoltaic power generation and energy storage batteries is achieved through ingenious control algorithms and strategies. First, the perturbation-observation method is used for photovoltaic maximum power point tracking, a dynamic adjustment strategy. By finely adjusting the voltage of the DC-DC boost circuit in real time, the system can quickly respond to changes in illumination conditions, continuously track and lock the maximum power output point of the photovoltaic panel, ensuring the full and efficient utilization of photovoltaic energy. Second, the bidirectional energy storage control mechanism employs an advanced dual-PI (Proportional Integral) controller. Through precise adjustment of the voltage outer loop and current inner loop, a PWM (Pulse Width Modulation) waveform is generated to control the switching action of the DC / DC converter, achieving precise bidirectional management of the charging and discharging of the energy storage battery. This mechanism ensures the stability and efficiency of the battery during charging and discharging. Through precise control, the battery can replenish or release energy in a timely manner according to system needs, while avoiding premature degradation of battery performance. To further optimize the operation of the photovoltaic-storage system, a charge / discharge threshold strategy was designed. When the battery's state of charge (SOC) is below 20%, the charging mode is forcibly activated to ensure the energy reserves of the energy storage system and prevent the system from being unable to respond to emergencies due to energy shortages. When the SOC exceeds 80%, the battery enters a discharge state, providing additional power support to the microgrid system, thereby achieving dynamic energy scheduling and balance, improving grid stability and the photovoltaic energy absorption rate. Finally, the three-phase AC-DC converter adopts feedforward decoupling control technology in the dq coordinate system. This is an advanced power electronic control strategy that achieves zero reactive power flow (i.e., q-axis current equals 0) during charging through real-time monitoring and precise control. This design not only improves the power quality during charging and reduces the reactive power burden on the grid, but also ensures the efficient operation of the charging system, reduces energy loss, and improves the overall system's energy utilization efficiency. In summary, this embodiment achieves intelligent management and optimized scheduling of photovoltaic energy and energy storage batteries through photovoltaic maximum power point tracking, bidirectional energy storage control, charge and discharge threshold strategies, and feedforward decoupling control of the three-phase AC / DC converter. This ensures the stable and efficient operation of the charging system under various environments, while reducing energy consumption and costs, providing users with a more intelligent and environmentally friendly charging solution.

[0067] Figure 4 This is an optional energy dispatch flowchart based on photovoltaic-storage coordinated power supply according to an embodiment of the present invention, such as... Figure 4As shown, the initialization constraints may include, but are not limited to, the maximum power output constraint of the energy storage battery (e.g., 20kW), the time interval constraint of the control cycle (e.g., 0.1 seconds), and the total capacity constraint of the electric vehicle battery (e.g., 100 kWh). Next, the current photovoltaic power generation P_pv is compared with the electric vehicle's charging demand power P_demand, handling power allocation under two scenarios: sufficient photovoltaic power and insufficient photovoltaic power. If P_pv is greater than or equal to P_demand, it indicates that the current photovoltaic power generation is sufficient to meet the charging demand. The grid power P_grid is set to 0, meaning no electricity is purchased from the grid. The battery supply power P_batt is determined to be the smaller value between the actual power demand of the electric vehicle P_demand and the maximum power output of the energy storage battery. The electric vehicle battery SOC (State of Charge) is updated: SOC_update = SOC_d - (P_batt) The formula is: P_batt = (dt) / C_batt, where SOC_d represents the current SOC of the battery, P_batt represents the power output or absorption by the energy storage battery to the system (such as an electric vehicle charging station) during the current control cycle, dt represents the time interval of the control cycle, and C_batt represents the total capacity of the battery. This means consuming battery energy to meet charging demand, while considering both the control cycle and the total battery capacity. If P_pv is less than P_demand, there is a power deficit, which needs to be made up by additional power. The system checks if the current SOC of the battery is greater than the energy storage protection threshold of 0.3 to prevent over-discharge: if the SOC is greater than 30%, the battery can participate in power supply, and P_batt is set to the smaller of the power deficit and the battery's maximum power. If the SOC is less than or equal to 30%, the battery does not participate in power supply, and P_batt is 0. The grid power P_grid will make up the remaining power deficit: P_grid = deficit - P_batt. The battery SOC is updated to SOC_update = SOC_d - (P_batt) / C_batt. The function `dt) / C_batt` ensures that even when the battery is involved in power supply, the correct energy consumption pattern is followed. To protect the battery and prevent the State of Charge (SOC) from being too high or too low, SOC boundary constraints are further applied: if `SOC_update` is less than 0.2, its value is set to 0.2 to prevent the battery SOC from being too low; if `SOC_update` is greater than 0.9, its value is set to 0.9 to prevent the battery SOC from being too high. Finally, the function returns the power obtained from the grid, the power output from the battery, and the updated battery SOC for the current scheduling cycle.

[0068] As an optional embodiment, Figure 5This is an optional fault decision flowchart according to an embodiment of the present invention. It can monitor the status monitoring parameters shown in Table 3 in real time and take corresponding protective actions when each parameter reaches the corresponding value range. Simultaneously, a multi-level fault handling mechanism is set up to take corresponding countermeasures for different fault levels. For example, at the first-level fault handling level, the design focuses on automatic recovery capability. Once a positioning deviation exceeding a preset range is detected, a visual recalibration process will be immediately initiated. The image of the charging interface will be captured again by the camera, and the position data will be recalculated to correct the positioning error of the robotic arm and ensure the accurate insertion process. Furthermore, in the face of instantaneous overcurrent, the system will automatically implement PWM (Pulse Width Modulation) current limiting protection. By adjusting the PWM signal, power delivery is controlled to avoid current peaks damaging the system or battery, thereby ensuring the safety and stability of system operation without affecting charging efficiency. Secondly, for second-level faults, further intervention measures are taken. If the contact resistance exceeds the normal range, i.e., the contact resistance exceeds the standard, the system will automatically activate the ultrasonic cleaning device to remove dust and impurities from the surface of the charging interface through high-frequency vibration, restoring the conductivity of the charging interface and ensuring the reliability of the electrical connection. Meanwhile, to address abnormal heat dissipation, the system is equipped with a dual-mode cooling system—a composite cooling method combining air cooling and liquid cooling. This system automatically activates to rapidly reduce the temperature of the charging interface, preventing overheating and potential safety hazards, while ensuring smooth charging. Finally, in the most severe Level 3 fault scenario, manual intervention is required to handle more complex issues. In the event of mechanical jamming, a safety emergency stop mechanism is immediately activated to prevent further movement of the robotic arm and avoid greater physical damage, while awaiting on-site troubleshooting and repair by technicians. In the event of insulation failure, the system automatically implements power-off protection, cutting off the power supply to prevent leakage or short circuits, and simultaneously sends an alarm to the operator, prompting for professional electrical inspection and repair to ensure a safe restart and re-operation of the system. This multi-level fault handling mechanism not only effectively addresses various faults that may occur during charging but also ensures that in most cases, the system can recover automatically or resume normal operation through system-level intervention, minimizing the impact of faults on charging services and fully protecting the safety of personnel and equipment in the event of serious problems.

[0069] Table 3

[0070]

[0071] It should be noted that, Figure 6 This is a front view of an optional electric vehicle charging compartment according to an embodiment of the present invention. Figure 7 This is a schematic diagram of a three-dimensional model of a charging compartment according to an embodiment of the present invention. The method of this embodiment is applied to, for example... Figure 6 and Figure 7The pre-installed integrated structure shown is a comprehensive solution combining solar panels, energy storage batteries, charging piles, intelligent roller shutters, and convenient ramps, aiming to provide an environmentally friendly, efficient, and user-friendly charging environment (such as...). Figure 6 and Figure 7 (As shown). This structural design achieves seamless integration of all key components, greatly improving the overall performance and flexibility of the system. In terms of operating modes, it provides two intelligent switching power supply schemes to cope with different charging needs and environmental conditions. In grid-connected mode, the system prioritizes using the clean energy generated by solar panels to provide charging services for electric vehicles. Any remaining electricity is efficiently stored in the energy storage battery for later use. This mode not only maximizes the utilization of photovoltaic energy but also ensures the stability and continuity of the charging process through the dynamic scheduling of the energy storage battery. In off-grid mode, the system can operate independently from the grid, with power directly provided by the energy storage battery to support charging needs in emergency situations. This mode is particularly suitable for charging scenarios in areas with unstable grids or remote locations, ensuring the normal operation of charging services in the event of power outages or other emergencies, demonstrating the system's excellent self-sufficiency and emergency response capabilities. Through this pre-installed integrated structure, this embodiment not only efficiently utilizes solar energy under normal power supply conditions, reducing dependence on the traditional power grid, but also ensures the continuity of charging services through the internal energy storage battery during grid failures, thus constructing an environmentally friendly and stable microgrid system, providing users with a new, reliable, and convenient charging solution.

[0072] Through steps S102 to S108, the precise location of the charging interface can be determined by fusing features from multiple frames of ORB images and point cloud features. This allows for precise automatic insertion of the charging gun by controlling the robotic arm, and dynamic adjustment of the charging power ratio based on a photovoltaic-storage synergy strategy. This achieves high precision, high efficiency, and intelligence in the charging process, significantly improving the electric vehicle charging experience and energy utilization efficiency. It also solves the technical problems of low accuracy in charging interface positioning and high charging costs in related technologies for electric vehicles. Specifically, this embodiment utilizes the ORB algorithm and laser point cloud information from visual positioning technology to fuse multi-source data, resulting in more stable and accurate charging interface location information. Through intelligent control of the robotic arm, the insertion of the charging gun is precisely controlled based on the fused global feature image, avoiding the inconvenience and errors of manual operation. Simultaneously, the system intelligently analyzes the real-time power supply capabilities of photovoltaics and the power grid, dynamically adjusting the power ratio to optimize the charging strategy, ensuring a stable and efficient charging process while fully utilizing renewable energy and reducing charging costs.

[0073] Based on the above embodiments and optional embodiments, the present invention proposes an optional implementation method for electric vehicle charging control. This method can be applied to the following system architecture, the core components of which include:

[0074] The solar photovoltaic panel (marked as 2) located on top of the charging compartment is installed at a 30° angle to optimize sunlight reception. It plays a crucial role in converting solar energy into DC power. With a peak power of 5kW, the photovoltaic panel provides the foundation for the system's green energy while ensuring sufficient power supply. The MPPT (Maximum Power Point Tracking) controller is the intelligent engine in the photovoltaic energy conversion process. Employing an advanced algorithm using perturbation and observation, it can quickly track and lock the maximum power point of the photovoltaic panel within 10ms, greatly improving the efficiency and stability of photovoltaic power generation and ensuring the system's efficient capture and utilization of solar energy. The energy storage battery module, a high-performance lithium-ion battery pack with a total capacity of 50kWh, not only stores ample electrical energy but also demonstrates intelligent SOC (State of Charge) management. When the SOC falls below 20%, the system automatically initiates a forced charging program; when the SOC exceeds 80%, the battery enters a discharging state to balance system energy and support the needs of the grid or the charging process. The bidirectional AC / DC converter, employing a dq coordinate system feedforward decoupling control method, enables bidirectional energy flow between the DC bus and the grid. This ensures the system's flexible scheduling capability for different power sources, allowing for a smooth transition whether powered by photovoltaics or drawn from the grid, meeting charging needs while maintaining stable system operation. The central control center, acting as the brain, manages and coordinates various functions of the photovoltaic-storage system, including real-time calculation of the ratio of photovoltaic to energy storage battery power and automatic switching to off-grid mode (when SOC is above 20%). This allows the system to continue providing stable and reliable charging services under various grid conditions or emergency situations. The robotic system integrates a vision positioning module and a magnetic gripper mechanism. The vision positioning module first quickly identifies the charging port location through coarse positioning (error ±10mm), and then uses the SIFT feature matching algorithm for precise positioning (error ±0.05mm), ensuring the accurate operation of the 6-DOF articulated robotic arm. The magnetic gripper mechanism utilizes electromagnetic control and an adaptive insertion force design to ensure the stability and safety of the charging gun during gripping and insertion. In particular, the 50% speed reduction motion control strategy for the robotic arm within a distance of 10 to 30 mm further enhances the accuracy and reliability of the charging gun insertion action, avoiding positioning errors and collision risks that may arise from high-speed operation. The collaborative work of these components constitutes the unique and powerful pre-installed integrated structure of this embodiment, which not only improves the automation level and energy efficiency of the charging station but also ensures the stability of charging services and the convenience of user operation. This reflects the innovative achievements of deep integration in the fields of industrial automation and new energy technologies, paving a new path for the development of future charging infrastructure.

[0075] Figure 8 This is a flowchart of an optional electric vehicle charging control method according to an embodiment of the present invention, such as... Figure 8As shown, the method includes:

[0076] S1, the charging preparation phase, includes:

[0077] S11, Vehicle entry and identity verification: The vehicle drives into the preset charging space, keeping the charging interface facing the robot's working area; the user scans the charging pile's QR code through the mobile application APP to start the charging service; the system verifies the user's identity and payment information, and generates a charging task instruction.

[0078] S12, Environmental Perception and Safety Check: Environmental sensors on the top of the charging compartment detect the surrounding safety conditions; the roller shutter door opens automatically, and the camera performs a preliminary scan of the vehicle's charging port area; the system checks the internal environment of the charging compartment to ensure there are no people or obstacles.

[0079] S13, Equipment self-test and initialization: The robotic arm control system performs a self-test to confirm that the movement of each joint is normal; the vision system is calibrated to ensure that the camera parameters are accurate; the photovoltaic energy storage system checks the photovoltaic power generation status and the battery SOC status.

[0080] S2, visual precision positioning, specifically includes two stages: coarse positioning and precision positioning.

[0081] Perform the following operations during the coarse positioning stage:

[0082] S211, charging port area detection: A high-definition camera (2 megapixels, such as a binocular camera) takes a wide-angle shot of the vehicle's charging port area; the ORB feature point extraction algorithm is used to identify the approximate location of the charging port; combined with laser point cloud semantic segmentation, the three-dimensional spatial information of the charging port is obtained; the initial coordinates are calculated, and the positioning accuracy is controlled within ±10mm.

[0083] S212, the robotic arm is initially positioned, and the robotic arm (6 degrees of freedom) moves from the standby position to near the charging port; based on the coarse positioning results, the end of the robotic arm reaches a safe position 50mm away from the charging port; adjust the posture of the robotic arm to ensure that the camera is facing the charging interface.

[0084] Perform the following operations during the fine positioning phase:

[0085] S221, Multi-scale feature fusion (i.e., multi-source data fusion): The camera performs a second high-resolution capture to obtain detailed images of the charging port. Multi-scale feature points are extracted using a scale pyramid + accelerated segment test (FAST) corner detection method. The specific implementation process of feature matching is the same as in the aforementioned embodiments and will not be repeated here.

[0086] S222, Dynamic Kalman Filter, real-time correction of pose estimation and suppression of sensor noise. The specific implementation process is the same as the aforementioned embodiment, and will not be repeated here.

[0087] S223, pixel-level precision optimization, calculates the sub-pixel offset of corner points using a quadratic surface fitting method to achieve sub-pixel edge detection, thereby obtaining more accurate corner coordinates. The specific implementation process is the same as in the aforementioned embodiments, and will not be repeated here.

[0088] S224, multi-frame pose fusion, uses pose filters to improve stability, such as weighted averaging of poses from 5 consecutive frames.

[0089] The S225 features adaptive ambient light compensation, automatically selecting overexposure, underexposure, or uneven lighting processing strategies based on average brightness to ensure stable operation within an illuminance range of 10-5000 lx.

[0090] S3, photovoltaic-storage collaborative power supply, includes:

[0091] S31, real-time energy status monitoring, including:

[0092] S311, photovoltaic power generation monitoring, the MPPT controller uses the perturbation observation method to track the maximum power point in real time; monitors the output power P_pv of the photovoltaic panel with a response time of <10ms; calculates the available photovoltaic power generation and prioritizes it for charging needs.

[0093] S312, energy storage system status assessment, real-time monitoring of lithium battery pack SOC status (0-100%); checking battery health status and temperature parameters; setting SOC operating range: 20%-90% to avoid overcharging and over-discharging.

[0094] S313, charging demand analysis, calculates the required power P_demand based on the vehicle battery type and current power level; considers the real-time electricity price C_grid to optimize electricity costs.

[0095] S32, Dynamic Power Allocation, includes:

[0096] S321 employs a photovoltaic-storage collaborative scheduling algorithm to formulate targeted charging strategies for different light intensity scenarios, determining the ratio of grid output power and photovoltaic output power for electric vehicle charging. The specific implementation process is the same as in the aforementioned embodiments and will not be repeated here.

[0097] S322, Optimize the achievement of the objective. In the power allocation process based on optical-storage coordinated scheduling, the objective function is set as follows: With the goal of minimizing the function value of the objective function, the power grid consumption and battery charging and discharging strategies are adjusted in real time; the photovoltaic absorption rate is ≥95%, and the system energy consumption is reduced by 37%.

[0098] In power conversion control, S32 employs bidirectional AC / DC converter control and dq coordinate system feedforward decoupling control. It utilizes a dual PI controller with an outer voltage loop and an inner current loop to achieve zero reactive power flow (q-axis current = 0). Regarding operating mode switching, it prioritizes photovoltaic power supply in grid-connected mode, storing surplus electricity and supplementing the grid; in off-grid mode, it provides independent power supply through energy storage batteries, supporting emergency charging; and implements automatic mode switching to ensure continuous power supply.

[0099] S4, the robotic arm insertion control process, specifically includes:

[0100] S41, Gun Grabbing and Pre-positioning. A magnetic gun gripping mechanism is used. After the robotic arm moves to the charging gun storage location, it is activated by an electromagnetic gripping mechanism, applying a constant 50N suction force to ensure a firm grip on the charging gun, ready for movement. In terms of path planning and motion, the optimal motion trajectory is planned based on visual positioning results, controlling the robotic arm's 6 degrees of freedom to coordinate movements and avoid obstacles; the end effector's orientation towards the charging port is optimized.

[0101] S42, Compliant Interlock Control, specifically includes:

[0102] S421 employs an impedance control algorithm, using the admittance control core algorithm as the control core, to obtain the force deviation when the charging gun contacts the charging interface, and determines the position adjustment amount of the robotic arm based on the force deviation; the position of the robotic arm is adjusted according to the position adjustment amount, and the specific implementation process is the same as the aforementioned embodiment, and will not be repeated here.

[0103] S422: During the process of the robotic arm inserting the charging gun into the charging interface according to the target pose, a dynamic deceleration insertion method is adopted. For example, when the charging gun is in the long-distance stage, that is, far from the charging interface (e.g., >30mm), the robotic arm is controlled to move towards the charging interface at a speed of 0.2m / s; when the charging gun is in the deceleration range, that is, at a moderate distance from the charging interface (e.g., 10-30mm), the robotic arm speed is controlled to linearly decrease from 0.2m / s to 0.1m / s; when the charging gun is in the fine insertion stage, that is, at a close distance from the charging interface (e.g., <10mm), the robotic arm is controlled to move towards the charging interface at a speed of 0.1m / s, with the acceleration limited to 0.5m / s².

[0104] S423, real-time force feedback adjustment. It monitors contact force in real time via a six-axis torque sensor; and dynamically adjusts the robotic arm position based on force deviation to ensure a smooth insertion process and avoid impact.

[0105] S43, Connection Confirmation and Charging Start, includes contact status verification, electronic lock activation, and charging process initiation.

[0106] During the contact state verification phase, the contact force is controlled to be stable within the range of 10-12N, the position deviation is less than 0.5mm, and the electrical connection resistance is less than 30 milliohms.

[0107] During the electronic lock activation phase, the electronic lock is activated after stable contact conditions are met, controlling the mechanical locking between the charging gun and the vehicle interface, preparing for high-current charging.

[0108] During the initial charging phase, the system sends a charging start command, sets the initial charging current to 500A, and monitors the charging status in real time.

[0109] S5, real-time status monitoring and fault handling, specifically includes:

[0110] The S51 provides real-time monitoring of multiple parameters. The parameters to be monitored and their corresponding thresholds are as follows: Contact resistance: Normal <30 milliohms, Dangerous >50 milliohms; Interface temperature: Normal 30-60℃, Dangerous >80℃; Charging current: Normal 0-500A, Dangerous >550A; Positioning deviation: Normal <0.2mm, Dangerous >0.5mm. During parameter acquisition, different data acquisition frequencies can be set according to the data characteristics. For example, the force sensor acquisition frequency can be set to 100Hz (every 10ms), the temperature sensor acquisition frequency to 10Hz (every 100ms), the electrical parameter acquisition frequency to 50Hz (every 20ms), and the vision system acquisition frequency to 30Hz (every 33ms).

[0111] S52, a three-level fault handling mechanism is set up, including:

[0112] Level 1 fault (automatic recovery): Visual recalibration is automatically triggered when positioning deviation is too large; PWM current limiting protection automatically recovers during transient overcurrent. The corresponding handling method is: the system automatically diagnoses and recovers, requiring no manual intervention.

[0113] Level 2 fault (system intervention): The ultrasonic cleaning device is activated when contact resistance exceeds the standard; the dual-mode cooling system (air cooling + liquid cooling) is automatically activated when heat dissipation is abnormal. The corresponding handling method is: system-level automated processing and fault log recording.

[0114] Level 3 fault (manual intervention): In the event of mechanical jamming, an immediate safety emergency stop will be initiated, awaiting technical personnel; in the event of insulation failure, automatic power-off protection will be activated, and a maintenance request will be issued. The corresponding handling method is: on-site handling by professional technicians is required.

[0115] The S53 features optimized charging processes, including dynamic power adjustment and updated user interaction.

[0116] Dynamic power adjustment can optimize the charging curve in real time based on battery temperature and charging efficiency; it automatically reduces power when the temperature is >50℃ to ensure charging safety and battery life.

[0117] User interaction updates can display charging progress and estimated completion time in real time through the app, promptly push notifications to users of abnormal statuses, and provide charging data statistics and analysis.

[0118] S6, charging complete and device returned to its place, specifically includes:

[0119] S61, during the charging completion processing phase, firstly, a charging completion judgment is made. When the vehicle battery reaches the preset charging target (100% or a user-defined value) and the charging current naturally decreases to the termination threshold, the system confirms that the charging task is complete. At the same time, safety power-off measures are implemented, gradually reducing the charging current to zero. After confirming that the electrical connection is safely disconnected, the electronic locking mechanism is released.

[0120] S62, Equipment return operation, sequentially performing robotic arm return, system status reset, and system status reset.

[0121] During the robotic arm's return phase, the robotic arm smoothly pulls out the charging gun, returns to the standby position along the planned path, and accurately places the charging gun back into the storage rack.

[0122] During the system status reset phase, the vision system switches to standby mode, the optical storage system adjusts to the default working state, and the roller shutter door closes automatically.

[0123] During the charging record generation phase, the system generates a detailed report of the charging session, including data such as charging amount, time taken, and energy consumption, and updates user account information and equipment maintenance records.

[0124] S7, Performance Indicators and Quality Control: Collect key performance indicators to verify the charging quality of electric vehicles. This may include, but is not limited to, collecting and testing the following indicators:

[0125] Positioning accuracy indicators include: coarse positioning accuracy: ±10mm; fine positioning accuracy: ±0.05mm (200 times better than traditional methods); positioning success rate: >99.6%.

[0126] Time efficiency metrics include: total process time: <5 minutes (285s); connection process: average 1.2 seconds; system response: <30ms.

[0127] Energy efficiency indicators include: photovoltaic absorption rate: ≥95%; system energy consumption: 37% lower than traditional solutions; charging efficiency: >89%.

[0128] Reliability metrics include mean time between failures (MTBF): >2000 hours; environmental adaptability: 10-5000 lx illuminance range; operating temperature range: -20℃ to 50℃.

[0129] This complete implementation plan achieves intelligent, precise, and safe charging throughout the entire process through a four-layer closed-loop control system of visual positioning, energy scheduling, force-controlled plugging, and status diagnosis, providing users with an efficient and convenient charging service experience.

[0130] This embodiment also provides an electric vehicle charging control device for implementing the above embodiments and preferred embodiments; details already described will not be repeated. As used below, the terms "module" and "device" can refer to a combination of software and / or hardware that performs a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0131] According to an embodiment of the present invention, an apparatus embodiment for implementing the above-described electric vehicle charging control method is also provided. Figure 9 This is a schematic diagram of the structure of an electric vehicle charging control device according to an embodiment of the present invention, as shown below. Figure 9 As shown, the above-mentioned electric vehicle charging control device includes: a feature image module 200, a feature fusion module 202, a pose determination module 204, and a charging control module 206, wherein:

[0132] The feature image module 200 is used to acquire multi-frame directional acceleration robust feature and rotation binary robust independent feature (ORB) feature images collected from the charging interface of the electric vehicle when the electric vehicle is detected to have reached the preset charging area, as well as the corresponding multi-frame point cloud feature images. The multi-frame ORB feature images and the multi-frame point cloud feature images correspond one-to-one.

[0133] The feature fusion module 202 is connected to the feature image module 200 and is used to perform feature fusion on multiple frames of ORB feature images and multiple frames of point cloud feature images to obtain multiple frames of target global feature images.

[0134] The pose determination module 204 is connected to the feature fusion module 202 and is used to determine the target pose of the charging interface based on multi-frame global feature images of the target.

[0135] The charging control module 206 is connected to the pose determination module 204 and is used to control the robotic arm to insert the charging gun into the charging interface according to the target pose, so as to charge the electric vehicle with the target charging strategy. The target charging strategy is used to indicate at least the ratio of grid output power and photovoltaic output power for charging the electric vehicle.

[0136] It should be noted that the above modules can be implemented by software or hardware. For example, for the latter, it can be implemented in the following ways: the above modules can be located in the same processor; or the above modules can be located in different processors in any combination.

[0137] It should be noted that the aforementioned feature image module 200, feature fusion module 202, pose determination module 204, and charging control module 206 correspond to steps S102 to S108 in the embodiments. The instances and application scenarios implemented by these modules and their corresponding steps are the same, but they are not limited to the content disclosed in the above embodiments. It should be noted that the aforementioned modules, as part of the device, can run on a computer terminal.

[0138] It should be noted that the optional or preferred implementation methods of this embodiment can be found in the relevant descriptions in the embodiments, and will not be repeated here.

[0139] The electric vehicle charging control device described above may also include a processor and a memory. The feature image module 200, feature fusion module 202, pose determination module 204, charging control module 206, etc., are all stored in the memory as program modules. The processor executes the program modules stored in the memory to realize the corresponding functions.

[0140] The processor contains a core that retrieves the corresponding program modules from memory. One or more cores may be configured. Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory includes at least one memory chip.

[0141] According to an embodiment of this application, an embodiment of a non-volatile storage medium is also provided. Optionally, in this embodiment, the non-volatile storage medium includes a stored program, wherein, when the program is running, it controls the device containing the non-volatile storage medium to execute any of the electric vehicle charging control methods described above.

[0142] Optionally, in this embodiment, the non-volatile storage medium may be located in any computer terminal in a group of computer terminals in a computer network, or in any mobile terminal in a group of mobile terminals, and the non-volatile storage medium includes stored programs.

[0143] Optionally, a program that controls the device containing the non-volatile storage medium to execute any of the above-described electric vehicle charging control method steps during program execution.

[0144] According to an embodiment of this application, an embodiment of a processor is also provided. Optionally, in this embodiment, the processor is used to run a program, wherein the program executes any of the above-described electric vehicle charging control methods.

[0145] According to an embodiment of this application, an embodiment of a computer program product is also provided, which, when executed on a data processing device, is adapted to execute a program that initializes the electric vehicle charging control method steps described above.

[0146] This invention provides an electronic device, which includes a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps of any of the above-described electric vehicle charging control methods.

[0147] The order of the above embodiments of the present invention is merely for description and does not represent the superiority or inferiority of the embodiments.

[0148] In the above embodiments of the present invention, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0149] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of modules described above can be a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, or indirect coupling or communication connection between modules, and may be electrical or other forms.

[0150] The modules described above as separate components may or may not be physically separate. Similarly, the components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple modules. Some or all of the modules can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0151] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The integrated modules described above can be implemented in hardware or as software functional modules.

[0152] If the aforementioned integrated modules are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable non-volatile storage medium. Based on this understanding, the technical solution of this invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a non-volatile storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned non-volatile storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0153] The above are merely preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for controlling the charging of an electric vehicle, characterized in that, include: When an electric vehicle is detected to have reached a preset charging area, a multi-frame image of directional acceleration robust features and rotational binary robust independent features (ORB features) collected from the charging interface of the electric vehicle is acquired, along with a corresponding multi-frame point cloud feature image. The multi-frame ORB feature image and the multi-frame point cloud feature image correspond one-to-one. The multi-frame ORB feature image and the multi-frame point cloud feature image are fused to obtain a multi-frame global feature image of the target. Based on the multi-frame global feature images of the target, the target pose of the charging interface is determined; The robotic arm is controlled to insert the charging gun into the charging interface according to the target pose, and to charge the electric vehicle according to the target charging strategy, wherein the target charging strategy is used to indicate at least the ratio of grid output power and photovoltaic output power for charging the electric vehicle.

2. The method according to claim 1, characterized in that, The step of fusing the multi-frame ORB feature images and the multi-frame point cloud feature images to obtain multi-frame global feature images of the target includes: Feature point matching is performed on the multi-frame ORB feature image and the multi-frame point cloud feature image to obtain the feature point matching result; Based on the feature point matching results, feature fusion is performed on the multi-frame ORB feature images and the corresponding point cloud feature images to obtain multi-frame initial global feature images. Kalman filtering is performed on the multi-frame initial global feature images to obtain the multi-frame target global feature images.

3. The method according to claim 1, characterized in that, Determining the target pose of the charging interface based on the multi-frame global feature image of the target includes: Subpixel-level corner optimization processing is performed on the multi-frame global feature images of the target to obtain the target corner coordinates corresponding to each of the multi-frame global feature images of the target; Based on the target corner coordinates corresponding to each of the multi-frame target global feature images, the pose of the charging interface is estimated to obtain the initial pose corresponding to each of the multi-frame target global feature images. The target pose is determined based on the initial pose corresponding to each of the multi-frame global feature images of the target.

4. The method according to claim 3, characterized in that, Determining the target pose based on the initial poses corresponding to each of the multi-frame global feature images of the target includes: Determine the weight values ​​corresponding to each of the multi-frame target global feature images; Based on the initial poses corresponding to each of the multi-frame global feature images of the target, and their respective weight values, a weighted average operation is performed to obtain the target pose.

5. The method according to claim 1, characterized in that, The controlled robotic arm connects the charging gun to the charging interface according to the target pose, including: During the process of controlling the robotic arm to insert the charging gun into the charging interface according to the target pose, the force deviation when the charging gun contacts the charging interface is obtained, wherein the force deviation represents the difference between the contact force when the charging gun contacts the charging interface and the expected contact force; Based on the force deviation, the position adjustment amount of the robotic arm is determined; The position of the robotic arm is adjusted according to the stated adjustment amount until the charging gun is successfully plugged into the charging interface.

6. The method according to claim 5, characterized in that, Determining the position adjustment amount of the robotic arm based on the force deviation includes: Determine the admittance coefficient, wherein the admittance coefficient represents the rate of change of displacement of the charging gun in the contact direction under a unit force; The position adjustment amount is obtained by multiplying the admittance coefficient and the force deviation.

7. The method according to any one of claims 1 to 6, characterized in that, Before the controlled robotic arm inserts the charging gun into the charging interface according to the target pose to charge the electric vehicle according to the target charging strategy, the method further includes: The objective function is determined as follows: Where J represents the value of the objective function, and min represents minimization. and Here, C_grid represents the real-time electricity price, P_grid represents the grid output power for charging the electric vehicle, and SOC_ref represents the reference state of charge of the electric vehicle. This refers to the real-time state of charge of the electric vehicle. The constraints include: the power demand of the electric vehicle is equal to the sum of the grid output power and the photovoltaic output power for charging the electric vehicle; Based on the objective function and the constraints, the initial charging strategy of the electric vehicle is optimized to obtain the target charging strategy.

8. An electric vehicle charging control device, characterized in that, include: The feature image module is used to acquire, when the electric vehicle is detected to have reached the preset charging area, a multi-frame directional acceleration robust feature and rotation binary robust independent feature (ORB) feature image collected from the charging interface of the electric vehicle, as well as a corresponding multi-frame point cloud feature image, wherein the multi-frame ORB feature image and the multi-frame point cloud feature image correspond one-to-one. The feature fusion module is used to fuse the multi-frame ORB feature images and the multi-frame point cloud feature images to obtain multi-frame global feature images of the target. The pose determination module is used to determine the target pose of the charging interface based on the multi-frame target global feature image; A charging control module is used to control the robotic arm to insert the charging gun into the charging interface according to the target posture, so as to charge the electric vehicle with a target charging strategy, wherein the target charging strategy is used to indicate at least the ratio of grid output power and photovoltaic output power for charging the electric vehicle.

9. A non-volatile storage medium, characterized in that, The non-volatile storage medium stores multiple instructions, which are adapted to be loaded by a processor and executed by the electric vehicle charging control method according to any one of claims 1 to 7.

10. An electronic device, characterized in that, It includes one or more processors and a memory, the memory being used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors cause the one or more processors to implement the electric vehicle charging control method according to any one of claims 1 to 7.

Images