Vehicle charging control method and system based on overhead rail type charging robot

Abstract

The invention relates to the technical field of charging robot control methods, and discloses an overhead rail type charging robot-based vehicle charging control method, which comprises the following steps of S1, importing an overhead rail layout, receiving a charging request and obtaining parking space information and charging port position information; s2, an operation point is determined according to the parking space information and the charging port position information, the real-time position of a charging robot with a charging plug is obtained, and the charging robot is controlled to move to the operation point from the real-time position along the hanger rail; s3, the charging robot is controlled to collect image information including a charging port at the operation point, and the position of the charging port is recognized and obtained by using an improved sparse representation method according to the image information; and S4, controlling the charging robot to insert the charging plug into the charging port. The beneficial effects of the invention are that the improved sparse representation method is utilized to effectively identify the charging port, and the stability and rapidity of the operation of the charging robot are further improved.

Description

Technical Field

[0001] This invention relates to the field of charging robot control methods, and in particular to a vehicle charging control method, control system, and rail-mounted charging robot based on a rail-mounted charging robot. Background Technology

[0002] With the development of electric vehicles, the demand for charging infrastructure is also showing an increasing trend. Although traditional charging piles have been widely deployed in cities, they still have some limitations, such as fixed locations that cannot be moved; high investment in the construction of charging infrastructure, and high maintenance and replacement costs; low equipment utilization, requiring manual plugging and unplugging of charging plugs, which leads to waste of charging resources if not plugged in and unplugged in time; and other issues such as insufficient number of charging piles, long charging times, and excessive distances between charging stations. These problems affect the convenience and popularity of electric vehicles. Therefore, developing more flexible and intelligent charging solutions has become an urgent task.

[0003] A charging robot is a mobile, intelligent device that automatically provides charging services. It automatically travels to an electric vehicle when charging is needed and autonomously completes the charging task. In existing technologies and patents, there are numerous cases of charging robots replacing manual labor in the automatic charging of electric vehicles. For example, invention application CN 115328149 A, entitled "Control Method for Charging Robot," discloses a control method for a charging robot, which includes: receiving a signal from a vehicle requesting charging and obtaining the position of the vehicle, the position of which is the target position for the charging robot's movement; obtaining the initial position of the charging robot and planning its movement trajectory based on the initial and target positions; controlling the charging robot to move from the initial position to the target position according to the movement trajectory; after the charging robot reaches the target position, acquiring an image of the vehicle through its camera; obtaining the charging port position of the vehicle through a convolutional neural network recognition model based on the image; and the charging robot inserting the charging gun into the charging port to charge the vehicle. It can be seen that the existing charging robot charging process mainly includes two main actions: planning the charging robot's movement trajectory and inserting the charging gun into the charging port.

[0004] However, in actual operation, the charging robot has the following problems: 1. When there are obstacles on the ground, the charging robot cannot completely cross or bypass the obstacles and cannot stably reach the target position; 2. The position and height of the charging port of existing vehicles are not uniform, and the charging robot cannot adapt to charging ports of different heights; 3. The convolutional neural network recognition model requires a large amount of training data. When the vehicle is updated, if the position of the charging port is different, it needs to be retrained. Summary of the Invention

[0005] To overcome the shortcomings of existing charging piles, such as their fixed and immobile nature, low equipment utilization, and the need for manual plugging and unplugging, and to address the problems of existing charging robots in the background art, this invention proposes a vehicle charging control method, control system, and a rail-mounted charging robot based on a rail-mounted charging robot. The detailed technical solution is as follows:

[0006] A vehicle charging control method based on a rail-mounted charging robot includes the following steps: S1, importing the rail layout, receiving a charging request, and obtaining parking space information and charging port location information; S2, determining the work point based on the parking space information and charging port location information, obtaining the real-time position of the charging robot with a charging plug, and controlling the charging robot to move along the rail from the real-time position to the work point; S3, controlling the charging robot to collect image information containing the charging port at the work point, and identifying and obtaining the location of the charging port based on the image information using an improved sparse representation method; S4, controlling the charging robot to insert the charging plug into the charging port.

[0007] Furthermore, S3 includes the following steps:

[0008] S301: When the charging robot arrives at the work point, it collects image information of the corresponding parking space in real time.

[0009] S302, the image information is preprocessed by noise reduction, brightness and contrast adjustment to obtain preprocessed image information;

[0010] S303, using an improved sparse representation method, the preprocessed image information is represented as a linear combination of sparse bases in an overcomplete dictionary, and the key features of the charging port are extracted from the preprocessed image information;

[0011] S304, calculate the feature map based on the preprocessed image information and the learned dictionary, and obtain the charging port location based on the sparse coding model and the feature map.

[0012] Furthermore, in S301, when the charging robot with a camera arrives at the designated work point, the camera begins to capture real-time image information of the designated parking space, and the image information of the designated parking space is denoted as f(x,y).

[0013] In S302, the image information f(x,y) is first subjected to a multi-level wavelet transform (DWT) to decompose it into wavelet coefficients at multiple frequency levels, as shown in equation (1):

[0014] F(l,m)=∫∫f(x,y)·ψ lm (x,y)dxdy (1)

[0015] In equation (1): f(x,y) represents a two-dimensional image, x and y represent the horizontal and vertical pixels of the image, respectively; dx and dy represent the derivatives with respect to variables x and y, respectively; F(l,m) are the wavelet coefficients at scale l and position m; ψ lm (x,y) is the wavelet basis function corresponding to this scale and location;

[0016] For each wavelet coefficient F(l,m), a threshold T is applied to remove noise; the thresholding process is shown in equation (2):

[0017]

[0018] In equation (2): F′ (l,m) These are the wavelet coefficients after thresholding.

[0019] Wavelet coefficients F′ after thresholding (l,m) The denoised image is reconstructed by inverse wavelet transform (IDWT), as shown in equation (3):

[0020]

[0021] In equation (3): f'(x,y) is the denoised image reconstructed by inverse wavelet transform.

[0022] Furthermore, in S303, dictionary D is obtained through learning, and the learning of dictionary D is shown in equation (4):

[0023]

[0024] In equation (4): D m This represents the optimal dictionary under optimization conditions. is the reconstruction error, which measures the difference between the reconstructed image and the original image; ||s||1 is the sparsity penalty term, implemented through the L1 norm, used to promote the sparsity of the coefficient vector s; λ is the regularization parameter, used to balance reconstruction error and sparsity.

[0025] The preprocessed image information f'(x,y) is represented as a linear combination of sparse bases in an overcomplete dictionary, and the sparse representation s of f'(x,y) is shown in equation (5):

[0026]

[0027] In equation (5): S m To obtain the optimal sparse representation coefficients under optimization conditions, the most important features in the reconstructed image f'(x,y) are identified by analyzing the sparse coefficient vector s, and the dictionary basis d is extracted. i dictionary base d i Features used to indicate the charging port and charging cover.

[0028] Furthermore, in S304, based on the processed image information f'(x,y) and each dictionary base d... i Convolution to obtain feature map K i As shown in equation (6):

[0029] K i =f'(x,y)*d i (6)

[0030] In equation (6): * represents the convolution operation, K i It is the preprocessed image f'(x,y) and the dictionary base d i The feature map obtained after convolution;

[0031] Obtain feature map K i Then, sparse coding is applied to each feature map to find the most representative optimal feature position, which is the position of the charging port and the charging cover, as shown in Equation (7):

[0032]

[0033] In equation (7): K m To find the optimal feature location under optimization conditions, S i It is feature map K i The sparse representation coefficients, representing the dictionary base d i In feature map K i The weights in the equation.

[0034] Furthermore, between S3 and S4, there is also S305, which checks the activation status of the dictionary base related to the charging cover, determines whether the sparsity coefficient of the related dictionary base exceeds a preset threshold, and thus determines whether the charging cover is open. If yes, proceed to S4; otherwise, send the corresponding information to the user.

[0035] Furthermore, the charging robot includes a camera and a robotic arm, with a charging plug fixed to the end of the robotic arm; the initial state of the robotic arm is a fully retracted state, with a position P0 = (x0, y0, z0), and the configuration of the robotic arm is represented by the angle vector Θ0 = [θ1, θ2, ..., θ n ];

[0036] In S4, the charging robot is controlled to adjust its position along the overhead rail so that the image of the identified charging port is located in the center of the camera.

[0037] The charging robot is controlled to rise and fall so that the robotic arm reaches the height of the charging port. The position of the end effector of the robotic arm after rising and falling is recorded as P. t =(x t ,y t ,zt ),but

[0038] Horizontal movement distance is calculated as follows

[0039] Joint angle Θ t The solution obtained using the gradient method is shown in equation (8):

[0040]

[0041] In equation (8): E(Θ) is the error function, which measures the difference between the current position and the target position, and α is the learning rate. It is the gradient of the error function with respect to Θ0.

[0042] Furthermore, it also includes S5, a charging robot that communicates with users, receives their vehicle usage requests, automatically disconnects and retracts the charging plug, and moves on to the next work site.

[0043] Further, S1 includes the following steps: S101, importing the overhead rail layout into the charging robot, the overhead rail layout including charging station location information, grid points, markers on the grid points, and RFID positioning devices on the grid points; S102, receiving the charging request sent by the user, and obtaining parking space information and charging port location information; the parking space information includes the location of the parking space for the charging request and the parking direction.

[0044] S2 includes the following steps: S201, obtaining the current position of the charging robot, and planning the movement path of the charging robot on the overhead rail based on the current position and the obtained charging demand parking space position; S202, controlling the charging robot to go to the work point according to the movement path.

[0045] A vehicle charging control system based on a rail-mounted charging robot is provided for implementing the aforementioned vehicle charging control method.

[0046] The beneficial effects of this invention are as follows: This invention solves the problem of difficult ground path planning for existing charging robots and their inability to effectively avoid ground obstacles by deploying overhead rails; simultaneously, it utilizes an improved sparse representation method to effectively identify charging ports, further enhancing the stability and speed of the charging robot's operation. This invention application helps the electric vehicle industry take an important step towards smarter, more convenient, and sustainable charging solutions, and is expected to improve the user experience of electric vehicles, alleviate the pressure on urban charging infrastructure, and promote the development of smart transportation. Attached Figure Description

[0047] Figure 1 This is a structural schematic diagram of the overhead rail charging robot of the present invention.

[0048] Figure 2This is a schematic diagram of the working state of the overhead rail charging robot of the present invention.

[0049] Figure 3 This is a flowchart illustrating the method of the present invention. Detailed Implementation

[0050] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. 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 are within the scope of protection of the present invention.

[0051] To overcome the problems of fixed and immobile charging piles and low utilization rate of charging equipment in existing technologies, this invention proposes a vehicle charging control method, control system, and rail-mounted charging robot based on a rail-mounted charging robot. The structure of the rail-mounted charging robot is as follows: Figure 1 As shown, the system includes a mobile mechanism 1 that provides power to move the charging robot along a suspended track; a control unit 2 located below the mobile mechanism for controlling motor rotation and robotic arm movement, and for external communication; a high-definition camera 3 located on one side of the lower surface of the control unit 2; a mobile rechargeable battery 4 located on the other side of the lower surface of the control unit 2; a multi-degree-of-freedom robotic arm 6 located on the same side of the lower surface of the control unit 2 as the high-definition camera 3; and a charging plug 7 connected to the mobile rechargeable battery 4. The end of the multi-degree-of-freedom robotic arm 6 is equipped with a flexible gripper 5, which is fixedly connected to the end of the charging plug 7 furthest from the mobile rechargeable battery 4. During operation, as... Figure 2 As shown, the vehicle to be charged is parked in the parking space, and a hanging rail is installed above the parking space. The control unit 2 receives the signal and controls the motor of the moving mechanism 1 to rotate, so that the charging robot moves along the hanging rail to the corresponding position. Then, the control unit 2 controls the robotic arm 6 to move so that it inserts the charging plug 7 into the charging port of the vehicle. After charging is completed, the control unit 2 controls the robotic arm 6 to pull out the charging plug 7 and repeats the above process.

[0052] A vehicle charging control method based on a rail-mounted charging robot includes the following steps: S1, importing the rail layout, receiving a charging request, and obtaining parking space information and charging port location information; S2, determining the work point based on the parking space information and charging port location information, obtaining the real-time position of the charging robot with a charging plug, and controlling the charging robot to move along the rail from the real-time position to the work point; S3, controlling the charging robot to collect image information containing the charging port at the work point, and identifying and obtaining the location of the charging port based on the image information using an improved sparse representation method; S4, controlling the charging robot to insert the charging plug into the charging port.

[0053] S1 specifically includes S101, importing the overhead rail layout into the charging robot, the overhead rail layout including charging pile location information of the charging robot, grid points, markers on the grid points, and RFID positioning devices on the grid points; S102, receiving the charging request sent by the user, and obtaining parking space information and charging port location information; the parking space information includes the location of the parking space for the charging request and the parking direction;

[0054] First, the overhead rail layout is imported into the system. This layout includes the charging robot's charging station location information, grid points, markers on the grid points, and RFID positioning devices on the grid points. Users need to park their vehicles in the parking spaces below the overhead rails and open the corresponding charging port cover to expose the charging port. They can then send a charging request to the system by scanning a QR code or logging into the app. Simultaneously, the system collects the parking space information and charging port location information for the vehicle. Parking space information includes the location of the parking space requiring charging and the parking direction. The charging port location information is primarily one of four: front left, rear left, front right, or rear right. Since different vehicles have different charging port designs, and even the same vehicle parked in different directions will have different charging port locations, manual selection allows for easy identification of the charging port's position on the vehicle.

[0055] S2 specifically includes S201, obtaining the current position of the charging robot, and planning the movement path of the charging robot on the overhead rail based on the current position and the obtained charging demand parking space position; S202, controlling the charging robot to move to the work point according to the movement path. The system uses the current position of the charging robot as the starting point and the charging demand parking space position as the ending point, and plans the movement path of the charging robot according to the overhead rail layout.

[0056] In S2, unlike traditional ground-based path planning, the overhead rail-mounted charging robot can only move along the overhead rail. When its speed and direction are constant, only the robot's running time needs to be controlled to reach the corresponding location, without considering ground obstacles, significantly reducing the system's computational load. The system sends signals to the charging robot's control unit 2, which drives the motor to rotate the synchronous pulley, belt, and drive wheel, thus enabling the charging robot to move.

[0057] In S3, when the charging robot arrives at the work site, it captures real-time footage using camera 3 and uses an improved sparse representation method to identify the location of the charging port of the vehicle to be charged, laying the foundation for subsequent automatic charging services. The identification process is as follows:

[0058] 1) Image Acquisition

[0059] When the charging robot arrives at the designated work point, the camera begins to capture real-time footage of the designated parking space. This footage will serve as the direct data source f(x,y) for identifying the opening and closing status of the charging port and its cover.

[0060] 2) Image preprocessing

[0061] The input image undergoes preprocessing steps, including noise reduction, brightness and contrast adjustment to reduce background interference and highlight the features of the charging port. Considering the underground parking garage scenario where this invention may be implemented, where lighting conditions are relatively dark, image noise reduction based on wavelet basis transformation is chosen to ensure more accurate identification of the charging port in subsequent processing stages. First, a multi-level wavelet transform (DWT) is performed on the image f(x,y) to decompose it into wavelet coefficients at multiple frequency levels, as shown in equation (1):

[0062] F(l,m)=∫∫f(x,y)·ψ lm (x,y)dxdy (1)

[0063] In equation (1): F(l,m) are the wavelet coefficients at scale l and position m, ψ lm (x,y) is the wavelet basis function corresponding to this scale and location. Next, for the wavelet coefficients F(l,m) of each layer, a threshold T is applied to remove noise. The thresholding process is shown in equation (2):

[0064]

[0065] In equation (2): F′ (l,m) These are the wavelet coefficients after thresholding. Using the thresholded wavelet coefficients F′ (l,m) The denoised image is reconstructed using inverse wavelet transform (IDWT), as shown in equation (3):

[0066] f'(x,y)=∑ l,m F′ lm ·ψ lm (x,y) (3)

[0067] In equation (3), f'(x,y) is the denoised image reconstructed through inverse wavelet transform. Denoising effectively removes noise from the image while preserving its main features, providing a valid input image for subsequent charging port identification.

[0068] 3) Feature extraction

[0069] An improved sparse representation method is used to extract key features of the charging port from the preprocessed image. This involves representing the image as a linear combination of sparse bases in an overcomplete dictionary. The bases in the dictionary are obtained through previous learning and optimization, which can effectively capture the features of the charging port and its cover (such as edges, shape, texture, etc.). The dictionary learning process is shown in Equation (4):

[0070]

[0071] In equation (4): D m This represents the optimal dictionary under optimization conditions. is the reconstruction error, which measures the difference between the reconstructed image and the original image; ||s||1 is the sparsity penalty term, implemented through the L1 norm, used to promote the sparsity of the coefficient vector s; λ is the regularization parameter, used to balance reconstruction error and sparsity.

[0072] The preprocessed image information f'(x,y) is represented as a linear combination of sparse bases in an overcomplete dictionary, and the sparse representation s of f'(x,y) is shown in equation (5):

[0073]

[0074] In equation (5): S m To obtain the optimal sparse representation coefficients under optimization conditions, the most important features in the reconstructed image f'(x,y) are identified by analyzing the sparse coefficient vector s, and the dictionary basis d is extracted. i dictionary base d i Features used to indicate the charging port and charging cover.

[0075] 4) Charging port identification and positioning

[0076] By using a sparse coding model, combined with the learned dictionary D = [d1, d2, ..., d...], k ] and the preprocessed image f'(x,y), where each dictionary base d i This represents a feature (such as edge, shape, texture, etc.) of the charging port and its cover. The identification and localization of the charging port can be achieved by calculating the image f'(x,y) and each dictionary basis d. i The convolution of the sparse representation is then solved to achieve the result shown in equation (6):

[0077] K i =f'(x,y)*d i (6)

[0078] In equation (5): * represents the convolution operation, K i It is the preprocessed image f'(x,y) and the dictionary base d i The feature map obtained after convolution. Feature map K is obtained. iThen, sparse coding is applied to each feature map to find the most representative feature location, i.e., the location of the charging port, as shown in Equation (7):

[0079]

[0080] In equation (7): K m To find the optimal feature location under optimization conditions, S i It is feature map K i The sparse representation coefficients, representing the dictionary base d i In feature map K i The weights in the image are determined. This process uses sparsity constraints to find the location in the image that best matches the charging port features. By analyzing specific features around the charging port features (e.g., the edges and shape of the charging port cover) and checking the activation of specific dictionary bases associated with the charging port cover, i.e., analyzing whether the sparsity coefficients of these bases exceed a preset threshold system, the open / closed state of the charging port cover can be further determined.

[0081] S4, Charging Plug and Charging Port Docking: After identifying the charging port and obtaining its position information, the robot first determines whether the charging port cover is closed. If the charging port cover is open, the robotic arm will proceed directly to the next docking operation; if the cover is closed, the user will be notified. After ensuring the charging port cover is open, the robot first adjusts its position horizontally using a moving mechanism, placing the identified charging port image at the center of the camera. The robotic arm starts in a fully retracted state, with a position of P0 = (x0, y0, z0). The configuration of the robotic arm is represented by the angle vector Θ0 = [θ1, θ2, ..., θ...]. n The robotic arm is raised to the height of the charging port through a lifting operation. Let P be the position of the end effector of the robotic arm after the lifting operation. t =(x t ,y t ,z t The horizontal movement distance is calculated as follows: Solve the joint angle Θ required to achieve the target position using inverse kinematics. t The gradient method is solved as shown in equation (7):

[0082]

[0083] In equation (7): E(Θ) is the error function, which measures the difference between the current position and the target position, and α is the learning rate. It is the gradient of the error function with respect to Θ0. The robot's motion path is calculated using the above principle to achieve precise docking. After docking, the robot uses built-in sensors to confirm the stability of the connection and sends a charging start notification to the user via software.

[0084] More specifically, it also includes S5, remote communication: a communication unit is built into the charging robot to communicate with users and collect user parking information, including parking space number, charging port opening / closing status, and charging port location range. Users can also automatically disconnect and retract the charging plug when they need to end the charging service early and move to the next work point, effectively improving the utilization rate of the device and avoiding resource waste.

[0085] This invention aims to provide a vehicle charging control method based on a rail-mounted charging robot. By utilizing the accompanying rail and parking space, the method simplifies the path planning for the charging robot to reach the work location, effectively reducing the impact of ground obstacles. Simultaneously, the improved sparse representation method in this invention effectively extracts the features of the charging port, facilitating the charging robot's identification and location of the charging port. This allows the charging robot to quickly and accurately insert the charging plug into the charging port to complete the charging operation.

[0086] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A vehicle charging control method based on a rail-mounted charging robot, characterized in that, Includes the following steps: S1, import the hanging rail layout, receive charging requests and obtain parking space information and charging port location information; S2, determine the work point based on parking space information and charging port location information, obtain the real-time position of the charging robot with charging plug, and control the charging robot to move from the real-time position to the work point along the overhead rail; S3, control the charging robot to collect image information containing the charging port at the work point, and use the improved sparse representation method to identify and obtain the location of the charging port based on the image information; S4, control the charging robot to insert the charging plug into the charging port.

2. The vehicle charging control method based on a rail-mounted charging robot according to claim 1, characterized in that, S3 includes the following steps: S301: When the charging robot arrives at the work point, it collects image information of the corresponding parking space in real time. S302, the image information is preprocessed by noise reduction, brightness and contrast adjustment to obtain preprocessed image information; S303, using an improved sparse representation method, the preprocessed image information is represented as a linear combination of sparse bases in an overcomplete dictionary, and the key features of the charging port are extracted from the preprocessed image information; S304, calculate the feature map based on the preprocessed image information and the learned dictionary, and obtain the charging port location based on the sparse coding model and the feature map.

3. The vehicle charging control method based on a rail-mounted charging robot according to claim 2, characterized in that, In S301, when the charging robot with a camera arrives at the designated work point, the camera begins to capture real-time image information of the designated parking space, and the image information of the designated parking space is denoted as f(x,y). In S302, the image information f(x,y) is first subjected to a multi-level wavelet transform (DWT) to decompose it into wavelet coefficients at multiple frequency levels, as shown in equation (1): F(l,m)=∫∫f(x,y)·ψ lm (x,y)dxdy (1) In equation (1): f(x,y) represents a two-dimensional image, x and y represent the horizontal and vertical pixels of the image, respectively; dx and dy represent the derivatives with respect to variables x and y, respectively; F(l,m) are the wavelet coefficients at scale l and position m; ψ lm (x,y) is the wavelet basis function corresponding to this scale and location; For each wavelet coefficient F(l,m), a threshold T is applied to remove noise; the thresholding process is shown in equation (2): In equation (2): F ( ' l,m) These are the wavelet coefficients after thresholding. Wavelet coefficients F after thresholding ( ' l,m) The denoised image is reconstructed by inverse wavelet transform (IDWT), as shown in equation (3): In equation (3): f'(x,y) is the denoised image reconstructed by inverse wavelet transform.

4. The vehicle charging control method based on a rail-mounted charging robot according to claim 3, characterized in that, In S303, dictionary D is obtained through learning, and the learning of dictionary D is as shown in equation (4): In equation (4): D m This represents the optimal dictionary under optimization conditions. is the reconstruction error, which measures the difference between the reconstructed image and the original image; ||s||1 is the sparsity penalty term, implemented through the L1 norm, used to promote the sparsity of the coefficient vector s; λ is a regularization parameter used to balance reconstruction error and sparsity; The preprocessed image information f'(x,y) is represented as a linear combination of sparse bases in an overcomplete dictionary, and the sparse representation s of f'(x,y) is shown in equation (5): In equation (5): S m To obtain the optimal sparse representation coefficients under optimization conditions, the most important features in the reconstructed image f'(x,y) are identified by analyzing the sparse coefficient vector s, and the dictionary basis d is extracted. i dictionary base d i Features used to indicate the charging port and charging cover.

5. The vehicle charging control method based on a rail-mounted charging robot according to claim 4, characterized in that, In S304, the processed image information f'(x,y) is compared with each dictionary base d. i Convolution to obtain feature map K i As shown in equation (6): K i =f'(x,y)*d i (6) In equation (6): * represents the convolution operation, K i It is the preprocessed image f'(x,y) and the dictionary base d i The feature map obtained after convolution; Obtain feature map K i Then, sparse coding is applied to each feature map to find the most representative optimal feature position, which is the position of the charging port and the charging cover, as shown in Equation (7): In equation (7): K m To find the optimal feature location under optimization conditions, S i It is feature map K i The sparse representation coefficients, representing the dictionary base d i In feature map K i The weights in the equation.

6. The vehicle charging control method based on a rail-mounted charging robot according to claim 5, characterized in that, Between S3 and S4, there is also S305, which checks the activation status of the dictionary base related to the charging cover, determines whether the sparsity coefficient of the related dictionary base exceeds a preset threshold, and determines whether the charging cover is open. If yes, proceed to S4; otherwise, send the corresponding information to the user.

7. The vehicle charging control method based on a rail-mounted charging robot according to claim 5, characterized in that, The charging robot includes a camera and a robotic arm, with a charging plug fixed to the end of the robotic arm. The initial state of the robotic arm is a fully retracted state, with a position of P0 = (x0, y0, z0). The configuration of the robotic arm is represented by the angle vector Θ0 = [θ1, θ2, ..., θ]. n ]; In S4, the charging robot is controlled to adjust its position along the overhead rail so that the image of the identified charging port is located in the center of the camera. The charging robot is controlled to rise and fall so that the robotic arm reaches the height of the charging port. The position of the end effector of the robotic arm after rising and falling is recorded as P. t =(x t ,y t ,z t ),but Horizontal movement distance is calculated as follows Joint angle Θ t The solution obtained using the gradient method is shown in equation (8): In equation (8): E(Θ) is the error function, which measures the difference between the current position and the target position, and α is the learning rate. It is the gradient of the error function with respect to Θ0.

8. The vehicle charging control method based on a rail-mounted charging robot according to claim 7, characterized in that, It also includes the S5, a charging robot that communicates with users, receives their vehicle usage requests, automatically disconnects and retracts the charging plug, and moves on to the next work site.

9. The vehicle charging control method based on a rail-mounted charging robot according to claim 1, characterized in that, S1 includes the following steps: S101, Import the overhead rail layout into the charging robot. The overhead rail layout includes charging station location information of the charging robot, grid points, markers on the grid points, and RFID positioning devices on the grid points. S102, Receive the charging request sent by the user, and obtain parking space information and charging port location information; the parking space information includes the location of the parking space for the charging request and the parking direction; S2 includes the following steps: S201, Obtain the current position of the charging robot, and plan the movement path of the charging robot on the overhead rail based on the current position and the obtained charging demand parking space position; S202, control the charging robot to proceed to the work point along the stated movement path.

10. A vehicle charging control system based on a rail-mounted charging robot, characterized in that, Used to implement the vehicle charging control method according to any one of claims 1-9.

Images