Method for controlling battery swapping process, edge computing system, and battery swapping station

Abstract

Disclosed are a method for controlling a battery swapping process, an edge computing system implementing the method, a battery swapping station having the edge computing system, and a computer readable storage medium implementing the method. The method is applied to a battery swapping station side edge computing system. The method comprises: in response to receiving first process information, sending a first instruction, wherein the first process information indicates that a battery swapping process enters a battery locking and unlocking stage, and the first instruction is used for instructing to turn on a light supplementing lamp and acquire battery lower surface data; using a trained visual neural network model to predict the received battery lower surface data, and generating result information indicating a blockage condition of a battery lock hole; and controlling the battery swapping process on the basis of the result information.

Description

Method for controlling battery swapping process, edge computing system, and battery swapping station

[0001] This application claims priority to Chinese patent application 202311768029.X, filed on December 20, 2023, entitled “Method, edge computing system and battery swap station for controlling battery swap process”, the entire contents of which are incorporated by reference into this application. Technical Field

[0002] The present disclosure relates to the field of battery swapping technology, and more specifically to a method for controlling a battery swapping process, an edge computing system for implementing the method, a battery swapping station equipped with the edge computing system, and a computer-readable storage medium for implementing the method. Background Art

[0003] With the popularity of battery swap stations, more and more failure scenarios have been discovered during operation. In particular, due to the complex external environment that the vehicle chassis is in contact with, there is a risk that the lock hole on the lower surface of the battery will be blocked by foreign objects. For example, as shown in Figure 1, in the northern snowy areas, when users drive in winter, the ice and snow on the ground will easily cause a layer of ice shell to adhere to the lower surface of the battery on the vehicle chassis, thereby blocking the lock hole. For another example, as shown in Figure 2, foreign objects such as plastic bags may be rolled into and block the lock hole during driving. When foreign objects block the lock hole or foreign objects appear on the battery locking components of the vehicle chassis, the unlocking mechanism on the battery swap platform will be unable to extend into the lock hole for unlocking, resulting in unlocking failure and the inability to remove the user's low-power battery.

[0004] The direct consequence of these failures is that users must wait for the on-duty specialist to deal with the blockage before they can swap batteries, which prolongs the time it takes for users to swap batteries. What's more serious is that when users encounter a locking and unlocking failure at an unmanned station, they will be forced to stay at the battery swap station, waiting for rescue by operation and maintenance personnel, which greatly affects the user experience. In addition, when a station is unable to provide service due to this failure, it will affect the subsequent queue of vehicles coming to this station for battery swapping. Such problems are more prominent in Europe, where the climate is cold and rainy and snowy. Due to the sparse population and high labor costs in Europe, the risk of vehicle detention is further magnified when encountering locking and unlocking problems. In general, the user detention problem caused by locking and unlocking failures caused by ice, snow or other foreign objects blocking the lock hole has seriously affected the user experience and reduced the efficiency of battery swap services.

[0005] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of the present disclosure, and therefore may include information that does not constitute the prior art known to ordinary technicians in this field.

[0006] Summary of the Invention

[0007] In order to solve or at least alleviate one or more of the above problems, the following technical solutions are provided. The embodiments of the present disclosure provide a method for controlling the battery swap process, an edge computing system for implementing the method, a battery swap station equipped with the edge computing system, and a computer-readable storage medium for implementing the method. This solution can prevent user detention accidents caused by locking and unlocking failures due to battery lock hole blockage, thereby ensuring the stable operation of the battery swap station.

[0008] According to a first aspect of the present disclosure, a method for controlling a battery swap process is provided, which is applied to an edge computing system on the battery swap station side, and the method includes: sending a first instruction in response to receiving first process information, wherein the first process information indicates that the battery swap process enters a battery locking and unlocking stage, and the first instruction is used to instruct to turn on the fill light and obtain battery bottom surface data; using a trained visual neural network model to predict the received battery bottom surface data and generate result information indicating the battery lock hole blockage condition; and controlling the battery swap process based on the result information.

[0009] As an alternative or supplement to the above solution, in a method according to an embodiment of the present disclosure, the battery lower surface data is acquired by an image acquisition device provided on a rail-guided vehicle for transporting batteries in the battery swap station.

[0010] As an alternative or supplement to the above solution, in a method according to an embodiment of the present disclosure, the fill light is arranged on the front edge and / or rear edge of the track-guided vehicle, and the lighting direction of the fill light is toward the ground and at a preset angle to the ground.

[0011] As an alternative or supplement to the above scheme, in a method according to an embodiment of the present disclosure, controlling the battery replacement process based on the result information includes: if the result information indicates that the battery lock hole is blocked, generating a second instruction for stopping the battery replacement process and a manual intervention request; and if the result information indicates that the battery lock hole is not blocked, generating a third instruction for continuing the battery replacement process.

[0012] As an alternative or supplement to the above scheme, in a method according to an embodiment of the present disclosure, controlling the battery replacement process based on the result information includes: if the result information indicates that the battery lock hole is blocked and the battery replacement station is an unmanned station, sending risk warning information and charging suggestions to the vehicle computer; if the blockage predicted by the visual neural network model belongs to the first category that is easy to remove and the battery replacement station is a manned station, sending waiting for processing information to the vehicle computer; if the blockage predicted by the visual neural network model belongs to the second category that is not easy to remove and the battery replacement station is a manned station, sending waiting for processing information and risk warning information to the vehicle computer.

[0013] As an alternative or supplement to the above solution, in a method according to an embodiment of the present disclosure, the method further includes: uploading the battery lower surface data to a cloud server for archiving.

[0014] As an alternative or supplement to the above scheme, in a method according to an embodiment of the present disclosure, the visual neural network model is constructed based on training data containing sample images and annotation information, and the sample images include images captured by the image acquisition device when the battery lock hole is not blocked and when it is blocked by different categories of blockages.

[0015] According to a second aspect of the present disclosure, an edge computing system is provided, which is arranged at the battery swap station side, and the edge computing system includes: a central control unit, which is configured to: send a first instruction in response to receiving first process information, wherein the first process information indicates that the battery swap process enters the battery locking and unlocking stage, and the first instruction is used to instruct to turn on the fill light and obtain the battery lower surface data; an intelligent reasoning unit, which is configured to: receive the battery lower surface data; use a trained visual neural network model to predict the battery lower surface data and generate result information indicating the battery lock hole blockage status; and send the result information to the central control unit, so that the central control unit can control the battery swap process based on the result information.

[0016] As an alternative or supplement to the above solution, in an edge computing system according to an embodiment of the present disclosure, the battery lower surface data is acquired by an image acquisition device installed on a rail-guided vehicle for transporting batteries in the battery swap station.

[0017] As an alternative or supplement to the above solution, in an edge computing system according to an embodiment of the present disclosure, the fill light is arranged on the front edge and / or rear edge of the track-guided vehicle, and the lighting direction of the fill light is toward the ground and at a preset angle to the ground.

[0018] As an alternative or supplement to the above scheme, in an edge computing system according to an embodiment of the present disclosure, the central control unit is further configured to: if the result information indicates that the battery lock hole is blocked, generate a second instruction for stopping the battery replacement process and a request for manual intervention; and if the result information indicates that the battery lock hole is not blocked, generate a third instruction for continuing the battery replacement process.

[0019] As an alternative or supplement to the above scheme, in an edge computing system according to an embodiment of the present disclosure, the central control unit is further configured to: if the result information indicates that the battery lock hole is blocked and the battery swap station is an unmanned station, then risk warning information and charging suggestions are sent to the vehicle computer; if the blockage predicted by the visual neural network model belongs to the first category that is easy to remove and the battery swap station is a manned station, then waiting for processing information is sent to the vehicle computer; if the blockage predicted by the visual neural network model belongs to the second category that is not easy to remove and the battery swap station is a manned station, then waiting for processing information and risk warning information are sent to the vehicle computer.

[0020] As an alternative or supplement to the above solution, in an edge computing system according to an embodiment of the present disclosure, the central control unit is further configured to upload the battery lower surface data to a cloud server for archiving.

[0021] As an alternative or supplement to the above scheme, in an edge computing system according to an embodiment of the present disclosure, the visual neural network model is constructed based on training data containing sample images and annotation information, and the sample images include images captured by the image acquisition device when the battery lock hole is not blocked and when it is blocked by different categories of blockages.

[0022] According to a third aspect of the present disclosure, a battery swap station is provided, which includes: a rail-guided vehicle for transporting batteries, the rail-guided vehicle being provided with an image acquisition device and a fill light for acquiring data of the lower surface of the battery; and any one of the edge computing systems described in the second aspect of the present disclosure.

[0023] According to a fourth aspect of the present disclosure, a computer-readable storage medium is provided, wherein the computer-readable storage medium includes instructions, and the instructions, when run, execute any one of the methods described in the first aspect of the present disclosure.

[0024] On the one hand, the solution for controlling the battery swap process according to one or more embodiments of the present disclosure utilizes a visual neural network model deployed at the edge of the battery swap station to infer and predict the battery bottom surface data. This does not require a cloud server to perform inference or comparison with pre-stored images, and can be directly embedded in the current normal battery swap process. When there is no obstruction, this can be performed without any obstruction, without affecting the user's battery swap experience. On the other hand, this solution automatically determines the blockage status of the battery lock hole by turning on the fill light and obtaining the battery bottom surface data when the battery swap process enters the battery locking and unlocking stage, thereby preventing user detention accidents caused by locking and unlocking failures caused by battery lock hole blockage, and ensuring the stable operation of the battery swap station. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The above and / or other aspects and advantages of the present disclosure will become clearer and easier to understand through the following description of various aspects in conjunction with the accompanying drawings, in which the same or similar elements are represented by the same reference numerals. In the accompanying drawings:

[0026] FIG1 is an exemplary schematic diagram of a battery lock hole being blocked by ice and snow;

[0027] FIG2 is an exemplary schematic diagram of a battery lock hole being blocked by a plastic bag;

[0028] FIG3 is a schematic flow chart of a method 30 for controlling a battery swapping process according to one or more embodiments of the present disclosure;

[0029] FIG4 illustrates an exemplary arrangement of cameras according to one or more embodiments of the present disclosure;

[0030] FIG5 is an exemplary schematic diagram showing the difference between imaging with and without a fill light;

[0031] FIG6 is an exemplary schematic diagram of an original image of battery lower surface data according to one or more embodiments of the present disclosure;

[0032] FIG7 is an exemplary schematic diagram of a cropped image according to one or more embodiments of the present disclosure;

[0033] FIG8 is an exemplary block diagram of an edge computing system 80 according to one or more embodiments of the present disclosure;

[0034] FIG9 is an exemplary diagram of a system architecture according to one or more embodiments of the present disclosure. DETAILED DESCRIPTION

[0035] The description of the following specific embodiments is merely exemplary in nature and is not intended to limit the disclosed technology or the application and use of the disclosed technology. In addition, there is no intention to be bound by any express or implied theory presented in the foregoing technical field, background technology or the following specific embodiments.

[0036] In the following detailed description of the embodiments, numerous specific details are set forth to provide a more thorough understanding of the disclosed technology. However, it will be apparent to one of ordinary skill in the art that the disclosed technology can be practiced without these specific details. In other instances, well-known features are not described in detail to avoid unnecessarily complicating the description.

[0037] Terms such as "comprising" and "including" indicate that, in addition to the units and steps directly and explicitly stated in the specification, the technical solution of the present disclosure does not exclude the possibility of having other units and steps not directly or explicitly stated. Terms such as "first" and "second" do not indicate the order of units in terms of time, space, size, etc., but are simply used to distinguish between units.

[0038] The battery swap station mentioned in this article can be understood as a battery swap platform or device used to swap vehicle power batteries. It can also be understood as a building-type battery swap station isolated from the outside world. In addition, it is understood that the battery swap station mentioned in this article includes integrated charging and swapping stations or pure battery swap stations.

[0039] Hereinafter, various exemplary embodiments according to the present disclosure will be described in detail with reference to the accompanying drawings.

[0040] FIG3 is a schematic flowchart of a method 30 for controlling a battery swapping process according to one or more embodiments of the present disclosure.

[0041] It should be noted that the step names mentioned herein are merely used to distinguish one step from another and facilitate reference to the steps, and do not represent a sequential order between the steps. The flowcharts, including the accompanying figures, are merely examples of how to execute this method. The steps may be executed in any order or simultaneously unless there is an obvious conflict.

[0042] It should also be noted that method 30 is applied to the edge computing system on the battery swap station side. Overall, with the help of the edge computing system, the battery swap anomaly of the vehicle, especially the battery lock hole anomaly, is detected close to the data source (that is, in the local edge computing layer), so as to ensure the real-time and stability of the detection. In addition, by using the edge computing system to perform computer vision judgments that require high processing power, the computing load on the cloud can be reduced to a certain extent, and since the edge detection is only performed on a single battery swap station, the speed can reach milliseconds, thereby improving the user experience.

[0043] As shown in FIG. 3 , in step S310 , a first instruction is sent in response to receiving first process information.

[0044] In this article, the battery swap process or process information can be used to indicate the time node or time period of the battery swap process. Specifically, it indicates the status of the power battery (depleted battery or service battery) on the vehicle chassis. Specifically, the entire battery swap process can be divided into, for example, a parking phase (i.e., parking the vehicle in a preset position on the parking platform), a depleted battery removal phase, a service battery installation phase, and a departure phase (i.e., the battery swap process ends and the vehicle leaves the battery swap station).

[0045] To prevent battery lock and unlock failures caused by a clogged battery lock hole, it is necessary to detect battery abnormalities during the battery swap process, especially before removing a depleted battery. In step S310, in response to receiving the first process information indicating that the battery swap process has entered the battery lock and unlock phase, the edge computing system at the battery swap station sends a first instruction to activate the fill light and obtain data on the battery's lower surface.

[0046] Optionally, the battery lower surface data is acquired by an image acquisition device on a rail guided vehicle (RGV) that is provided in the battery swap station for transporting batteries. Exemplarily, the image acquisition device may include a two-dimensional or three-dimensional camera, a gun-type camera, a fisheye camera, an infrared camera, a hemispherical camera, or any other form of camera device. It should be understood that in the embodiments of the present disclosure, cameras, video cameras, and cameras, etc. all refer to devices that can acquire images or images within the coverage range, and their meanings are similar and interchangeable, and the present disclosure does not limit this. Preferably, in one or more embodiments according to the present disclosure, the image acquisition device includes one or more battery lower surface fisheye cameras, and the specific number of battery lower surface fisheye cameras on the RGV can be selected according to the distance between the RGV and the vehicle chassis, and adjusted according to the actual situation of the battery swap station.

[0047] Exemplarily, before executing method 30, it is necessary to reserve an installation position for an image acquisition device (i.e., an image sensor) on the RGV. In particular, since the fisheye camera has a large field of view, installing one or two fisheye cameras can cover the entire lower surface of the battery. For an exemplary arrangement of sensors (e.g., fisheye cameras), please refer to FIG4 , wherein the left side figure shows an exemplary arrangement of one sensor and the right side figure shows an exemplary arrangement of two sensors. Exemplarily, after the sensor is installed, it needs to be calibrated with the reference object on the RGV. It should be noted that since the image acquisition device set at the reserved position on the RGV can cover the entire lower surface of the battery and can obtain complete data on the lower surface of the battery, during the execution of method 30, the image acquisition device does not need to be lifted and moved when collecting data, and thus no new mechanical movement steps are required.

[0048] Furthermore, before executing method 30, it is necessary to reserve an installation position for a fill light on the RGV to optimize the data collection effect. For example, Figure 5 shows the difference in imaging without using a fill light (left picture) and using a fill light (right picture). As can be seen from Figure 5, the use of a fill light can improve the clarity of the collected data on the lower surface of the battery, thereby improving the accuracy of subsequent data processing. Optionally, the fill light is set at the front edge and / or rear edge of the RGV, and the lighting direction of the fill light is toward the ground and at a preset angle to the ground. For example, the lighting direction is inclined toward the ground and the angle between the light and the horizontal ground is 30 degrees. The main reason for downward illumination is to avoid the reflection of the metal on the bottom surface of the battery to form a large light spot, thereby affecting the accuracy of subsequent data processing.

[0049] In step S320, the trained visual neural network model is used to predict the received battery bottom surface data and generate result information indicating the battery lock hole blockage condition.

[0050] The visual neural network model utilizes deep neural network technology from computer vision. The model is pre-trained and embedded into the software of the battery swap station's edge computing system (for example, the software of the intelligent inference unit). The model possesses generalization and compatibility, allowing the trained model to be applied to untrained data environments, maintaining high-precision judgments.

[0051] Furthermore, the visual neural network model is iterative, allowing it to be updated based on the increasing number of images in the database. Specifically, after the camera and fill light are installed, images of the vehicle chassis can be taken and stored in the database. Once a sufficient number of images have been accumulated, the model can be trained. After training, the visual neural network model can be iterated over time using the latest images from the database.

[0052] Optionally, the visual neural network model is constructed based on training data containing sample images and annotation information, and the sample images include images captured by the image acquisition device when the battery lock hole is not blocked and when it is blocked by different types of blockages (for example, ice and snow, plastic bags, and mud). Exemplarily, during the training process, the original image of the battery lower surface data can be collected first, and then the original image can be cropped and labeled, the cropped image and label can be input into the model, and training can be performed in a supervised manner. Figure 6 shows an exemplary schematic diagram of the original image of the battery lower surface data, which includes the original image of the front battery lower surface (left image) and the original image of the rear battery lower surface (right image) taken by two fisheye cameras responsible for shooting the front chassis of the vehicle and the rear chassis of the vehicle, in which the area to be cropped including the battery lock hole is marked. Figure 7 shows an exemplary schematic diagram of the image after the detection area is cropped, wherein the lock hole in the left image is not blocked, and a label = 0 can be added; the lock hole in the right image is blocked, and a label = 1 can be added.

[0053] For example, the visual neural network model in the present disclosure can be regarded as an equation F, and the data image can be regarded as x; the training of the deep learning network first includes forward propagation, that is, x is mapped to Then use the loss function L to calculate the predicted value The loss with the true label y (the image annotation described above) Backpropagation then iterates the parameters of the neural network model F using gradient descent until training is terminated or the loss function output converges. Once the model is trained, it can be saved in a callable model file format, such as .pth, then converted to .engine format and packaged into the edge computing software at each battery swap station. Furthermore, model training is not restricted by programming languages ​​or frameworks (such as PyTorch or TensorFlow), nor by computer hardware or software systems, and can be freely selected based on actual circumstances.

[0054] Optionally, the result information generated in step S320 may indicate whether the battery lock hole is blocked, and may further indicate the type of blockage, for example, whether the blockage belongs to the first category that is easy to remove (such as a plastic bag), or the second category that is difficult to remove (such as ice and snow).

[0055] In step S330, the battery replacement process is controlled based on the result information.

[0056] Exemplarily, the edge computing system on the battery swap station side can confirm whether the vehicle lock hole is blocked by ice, snow or other foreign objects based on the result information generated by the visual neural network model, and control the subsequent battery swap process accordingly. In one or more embodiments according to the present disclosure, if the result information indicates that the battery lock hole is blocked, a second instruction for stopping the battery swap process and a manual intervention request are generated; and if the result information indicates that the battery lock hole is not blocked, a third instruction for continuing the battery swap process is generated. The above-mentioned manual intervention request refers to a request sent to the operation and maintenance specialist of the battery swap station for on-site processing.

[0057] Exemplarily, the edge computing system at the battery swap station can also control the subsequent battery swap process based on whether there is someone on duty. In one or more embodiments according to the present disclosure, if the result information indicates that the battery lock hole is blocked and the battery swap station is unmanned, risk warning information and charging suggestions are sent to the vehicle computer. For example, the predicted results and pictures are sent to the user to inform the user of the risks and suggest the user to choose charging, and the battery swap process is actively terminated; if the obstruction predicted by the visual neural network model belongs to the first category that is easy to remove and the battery swap station is manned, a waiting information is sent to the vehicle computer, and a specialist may be prompted to resolve it on site; if the obstruction predicted by the visual neural network model belongs to the second category that is not easy to remove and the battery swap station is manned, a waiting information and risk warning information are sent to the vehicle computer, and a specialist may be prompted to handle it.

[0058] Optionally, method 30 further includes: uploading the battery lower surface data to a cloud server for archiving.

[0059] Reference is now made to Figure 8, which is an exemplary block diagram of an edge computing system 80 according to one or more embodiments of the present disclosure. Edge computing system 80 is deployed at a battery swap station and includes a central control unit 810 and an intelligent reasoning unit 820. For example, central control unit 810 is an industrial control computer, and intelligent reasoning unit 820 is an edge reasoning computing platform.

[0060] The central control unit 810 is configured to: send a first instruction in response to receiving the first process information, wherein the first process information indicates that the battery replacement process enters the battery locking and unlocking stage, and the first instruction is used to instruct to turn on the fill light and obtain the battery lower surface data; and control the battery replacement process based on the result information. Optionally, the battery lower surface data is obtained by an image acquisition device provided on a rail-guided vehicle for transporting batteries in the battery replacement station. Optionally, the fill light is provided on the front edge and / or rear edge of the rail-guided vehicle, and the lighting direction of the fill light is toward the ground and at a preset angle to the ground. Optionally, the central control unit 810 is further configured to: if the result information indicates that the battery lock hole is blocked, generate a second instruction for stopping the battery replacement process and a manual intervention request; and if the result information indicates that the battery lock hole is not blocked, generate a third instruction for continuing the battery replacement process. Optionally, the central control unit 810 is further configured to: if the result information indicates that the battery lock hole is blocked and the battery swap station is unmanned, then send a risk warning message and charging advice to the vehicle computer; if the visual neural network model predicts that the blockage belongs to the first category that is easy to remove and the battery swap station is manned, then send a waiting for processing message to the vehicle computer; if the visual neural network model predicts that the blockage belongs to the second category that is difficult to remove and the battery swap station is manned, then send a waiting for processing message and a risk warning message to the vehicle computer. Optionally, the central control unit 810 is further configured to: upload the battery bottom surface data to a cloud server for archiving.

[0061] Intelligent reasoning unit 820 is configured to: receive battery lower surface data; use a trained visual neural network model to predict the battery lower surface data and generate result information indicating the battery lock hole obstruction status; and send the result information to central control unit 810. Optionally, the visual neural network model is constructed based on training data comprising sample images and annotation information, wherein the sample images include images captured by an image capture device when the battery lock hole is unobstructed and when it is obstructed by different types of obstructions.

[0062] Figure 9 is an exemplary schematic diagram of the system architecture according to one or more embodiments of the present disclosure. As shown in Figure 9, there are battery swap stations sharing a cloud server, and each battery swap station can communicate with the cloud server to archive the image data uploaded by all battery swap stations on the server side. It should be noted that the cloud server does not participate in algorithm prediction and is only used for the storage of images and battery swap information. In addition to the newly added cameras and fill lights, in order to implement the present disclosure, it is also necessary to utilize the central control unit (for example, the central control unit 810) and the intelligent reasoning unit (for example, the intelligent reasoning unit 820) that each battery swap station edge side currently has. The cloud server (or server cluster) can communicate with the central control unit on the edge side of each battery swap station. The central control unit and the intelligent reasoning unit can communicate through a wired connection via a switch.

[0063] As mentioned above, the visual neural network model can be deployed on the local intelligent inference unit at the battery swap station, with each intelligent inference unit hosting a specific model. When the battery swap process reaches the unlocking stage, the RGV's fisheye camera captures a real-time image, and the central control unit calls the model in the intelligent inference unit. The intelligent inference unit performs inference within milliseconds and sends the inference results to the central control unit.

[0064] The present disclosure can also be implemented as a battery swap station, which includes an RGV for transporting batteries. The RGV is provided with an image acquisition device and a fill light for acquiring data of the lower surface of the battery. The battery swap station also includes an edge computing system 80 as shown in FIG8 .

[0065] In addition, the present disclosure may also be implemented as a computer-readable storage medium, which stores therein information for causing a computer to execute the method 30 shown in Figure 3. Here, as the computer-readable storage medium, various computer-readable storage media can be used, such as disks (e.g., magnetic disks, optical disks, etc.), cards (e.g., memory cards, optical cards, etc.), semiconductor memories (e.g., ROMs, non-volatile memories, etc.), and tapes (e.g., magnetic tapes, cassettes, etc.).

[0066] The solution for controlling the battery swap process according to one or more embodiments of the present disclosure solves the problem of being unable to unlock and remove a depleted battery due to ice, snow or foreign objects blocking the lock hole, and improves the vehicle chassis abnormality detection function. In addition, this solution only requires adding one or two fisheye cameras and fill lights to the RGV and can be seamlessly embedded in the current battery swap process. After the camera is installed, there is no need to lift or move the camera, there is no significant cost increase, and no additional mechanical battery swap steps are required. Furthermore, this solution uses deep convolutional neural networks in computer vision to implement computer vision judgment, without the need for comparison with pre-saved images. After training, the model can be directly deployed in the intelligent reasoning unit on each battery swap station, and the detection is completed locally at the battery swap station. The detection speed is fast (milliseconds), and the user will not feel it when the lock hole is not blocked. This solution also fully considers the energy replenishment capacity of the battery swap station, and can provide different feedback to users and specialists for manned or unmanned scenarios, and scenarios that can be handled on-site or cannot be handled on-site.

[0067] Where applicable, hardware, software, or a combination of hardware and software may be used to implement the various embodiments provided by the present disclosure. Moreover, where applicable, without departing from the scope of the present disclosure, the various hardware components and / or software components set forth herein may be combined into composite components comprising software, hardware, and / or both. Where applicable, without departing from the scope of the present disclosure, the various hardware components and / or software components set forth herein may be divided into subcomponents comprising software, hardware, or both. In addition, where applicable, it is contemplated that software components may be implemented as hardware components, and vice versa.

[0068] Software according to the present disclosure (such as program code and / or data) can be stored on one or more computer-readable storage media. It is also contemplated that the software identified herein can be implemented using one or more general or special computers and / or computer systems that are networked and / or otherwise. The embodiments and examples proposed herein are provided to best illustrate the embodiments according to the present disclosure and its specific applications, and to enable those skilled in the art to implement and use the present disclosure. However, it will be appreciated by those skilled in the art that the above description and examples are provided only for the sake of ease of explanation and example. The description proposed is not intended to cover all aspects of the present disclosure or to limit the present disclosure to the disclosed precise form.

Claims

1. A method for controlling a battery replacement process, characterized in that: The method is applied to an edge computing system at a battery swap station, and the method includes: Sending a first instruction in response to receiving the first process information, wherein the first process information indicates that the battery replacement process enters the battery locking and unlocking stage, and the first instruction is used to instruct to turn on the fill light and obtain the battery bottom surface data; Using the trained visual neural network model to predict the received battery bottom surface data and generate result information indicating the battery lock hole blocking condition; and The battery replacement process is controlled based on the result information.

2. The method according to claim 1, wherein: The battery lower surface data is acquired by an image acquisition device installed on a rail-guided vehicle for transporting batteries in the battery swap station.

3. The method according to claim 1 or 2, wherein: The fill light is arranged at the front edge and / or the rear edge of the track-guided vehicle, and the illumination direction of the fill light is toward the ground and forms a preset angle with the ground.

4. The method according to claim 1, wherein: Controlling the battery swapping process based on the result information includes: If the result information indicates that the battery lock hole is blocked, generating a second instruction for stopping the battery replacement process and a manual intervention request; and If the result information indicates that the battery lock hole is not blocked, a third instruction is generated for continuing the battery replacement process.

5. The method according to claim 1, wherein: Controlling the battery swapping process based on the result information includes: If the result information indicates that the battery lock hole is blocked and the battery swap station is unmanned, a risk warning message and charging suggestions are sent to the vehicle computer; If the obstruction predicted by the visual neural network model belongs to the first category that is easy to remove and the battery swap station is a manned station, sending a waiting-for-processing message to the vehicle computer; If the obstruction predicted by the visual neural network model belongs to the second category that is not easy to remove and the battery swap station is a manned station, waiting for processing information and risk warning information are sent to the vehicle computer.

6. The method according to claim 1, wherein: The method further comprises: The battery bottom surface data is uploaded to a cloud server for archiving.

7. The method according to claim 1, wherein: The visual neural network model is constructed based on training data including sample images and annotation information, wherein the sample images include images captured by the image acquisition device when the battery lock hole is not blocked and is blocked by different categories of blockages.

8. An edge computing system, characterized in that: The edge computing system is arranged at the battery swap station side, and the edge computing system includes: A central control unit configured to: Sending a first instruction in response to receiving the first process information, wherein the first process information indicates that the battery replacement process enters the battery locking and unlocking stage, and the first instruction is used to instruct to turn on the fill light and obtain the battery bottom surface data; An intelligent reasoning unit configured to: Receiving the battery bottom surface data; Using the trained visual neural network model to predict the battery bottom surface data and generate result information indicating the battery lock hole blockage condition; and The result information is sent to the central control unit so that the central control unit can control the battery replacement process based on the result information.

9. The edge computing system according to claim 8, wherein: The battery lower surface data is acquired by an image acquisition device installed on a rail-guided vehicle for transporting batteries in the battery swap station.

10. The edge computing system according to claim 8 or 9, wherein: The fill light is arranged at the front edge and / or the rear edge of the track-guided vehicle, and the illumination direction of the fill light is toward the ground and forms a preset angle with the ground.

11. The edge computing system according to claim 8, wherein: The central control unit is further configured to: If the result information indicates that the battery lock hole is blocked, generating a second instruction for stopping the battery replacement process and a manual intervention request; and If the result information indicates that the battery lock hole is not blocked, a third instruction is generated for continuing the battery replacement process.

12. The edge computing system according to claim 8, wherein: The central control unit is further configured to: If the result information indicates that the battery lock hole is blocked and the battery swap station is unmanned, a risk warning message and charging suggestions are sent to the vehicle computer; If the obstruction predicted by the visual neural network model belongs to the first category that is easy to remove and the battery swap station is a manned station, sending a waiting-for-processing message to the vehicle computer; If the obstruction predicted by the visual neural network model belongs to the second category that is not easy to remove and the battery swap station is a manned station, waiting for processing information and risk warning information are sent to the vehicle computer.

13. The edge computing system according to claim 8, wherein: The central control unit is further configured to upload the battery lower surface data to a cloud server for archiving.

14. According to the edge computing system of claim 8, the visual neural network model is constructed based on training data containing sample images and annotation information, and the sample images include images captured by the image acquisition device when the battery lock hole is not blocked and when it is blocked by different categories of blockages.

15. A battery swap station, characterized in that: The battery swap station comprises: A rail-guided vehicle for conveying batteries, wherein the rail-guided vehicle is provided with an image acquisition device and a fill light for acquiring data of the lower surface of the battery; and An edge computing system according to any one of claims 8 to 14.

16. A computer-readable storage medium, characterized in that: The computer-readable storage medium comprises instructions which, when executed, perform the method according to any one of claims 1-7.

Images