A method for detecting the tilt position and tilt angle of a battery pack of a battery-swap vehicle

Abstract

The present invention provides a method for detecting the tilt position and tilt angle of a battery pack in a battery-swap vehicle. The method selects the outer edge and shape of the battery pack as preliminary identification feature points of the battery pack, uses a computer vision algorithm to extract the features of the battery pack from a preprocessed image, inputs the extracted features into a classifier, and uses a machine learning or deep learning algorithm to classify and identify the battery to determine whether the battery pack is a battery pack that can be swapped. The method captures four second feature points using a 3D depth camera and extracts the returned depth values ​​of the feature points. The method also uses a machine learning or deep learning algorithm to identify and compare the four depth values ​​and determine whether the battery pack is tilted. The present invention can quickly detect the tilt position and tilt angle values ​​of the battery pack, automatically detect and notify the RGV to make adjustments, improve battery swap efficiency, and reduce swap time.

Description

Technical Field

[0001] The present invention relates to the field of new energy battery replacement technology, and in particular to a method for detecting the tilt orientation and tilt angle of a battery pack of a battery replacement vehicle. Background Art

[0002] The battery replacement at the bottom of the heavy-duty truck battery swap station and the battery replacement of passenger cars are mainly unlocked by the unlocking mechanism on the RGV at the bottom of the battery to remove the battery. After unlocking, the battery is taken out and placed in the charging compartment of the battery swap station for charging, providing a new battery for the next battery swap vehicle, thereby achieving the purpose of fast battery replacement.

[0003] However, when a vehicle is swapping batteries at a battery swap station and the tire pressure is insufficient or one side is heavier than the other, the vehicle as a whole will tilt. As a result, the battery pack at the bottom of the vehicle will also have different degrees of tilt angles in different directions, thereby affecting the battery removal hardware equipment. There is currently no method for detecting the tilt angle of the battery pack at the bottom of a vehicle swapping batteries at a battery swap station. Summary of the Invention

[0004] The present invention aims to provide a method for detecting the tilt orientation and tilt angle of a battery pack of a battery-swap vehicle to overcome the above-mentioned problems or at least partially solve the above-mentioned problems.

[0005] To achieve the above object, the technical solution of the present invention is specifically implemented as follows:

[0006] The present invention provides a method for detecting the tilt orientation and tilt angle of a battery pack of a battery-swapping vehicle, comprising the following steps:

[0007] Step 1: Place the camera and 3D camera hardware on the RGV unlocking mechanism and drive under the vehicle;

[0008] Step 2: In the battery swapping area of ​​the battery swapping station, use a camera to capture an image of the battery pack. The camera can clearly capture the front, back, and sides of the battery.

[0009] Step 3: Preprocess the image captured by the camera to improve the subsequent recognition accuracy;

[0010] Step 4: Select the outer edge and shape of the battery pack as the initial identification feature points of the battery pack. Use a computer vision algorithm to extract the battery pack features from the pre-processed image to determine whether the battery pack is dirty or damaged.

[0011] Step 5: Input the extracted features into a classifier, and use machine learning or deep learning algorithms to classify and identify the battery pack to determine whether the battery pack is replaceable. The classifier is trained and classified based on different battery brands, models, or other attributes;

[0012] Step 6: Select the four lock holes on the top, bottom, left, and right sides of the battery pack as four second feature points, and use a 3D depth camera to photograph the four second feature points and extract the returned depth values ​​of the feature points;

[0013] Step 7: Input the extracted depth values ​​returned by the four keyholes into the processing terminal, and identify and compare the four depth values ​​through machine learning or deep learning algorithms to determine whether the battery pack is tilted;

[0014] Step 8: Use machine learning or deep learning algorithms to determine the tilt orientation of the battery pack, and transmit the judgment result to the PLC host and RGV unlocking mechanism through relevant protocols, and adjust the unlocking posture of the RGV unlocking mechanism until the PLC unlocking mechanism and the battery pack are relatively vertical and horizontal to ensure that the four unlocking gun heads on the RGV unlocking mechanism correspond to the four lock holes on the battery pack;

[0015] Step nine: Unlock the battery pack through the RGV unlocking mechanism and replace the battery.

[0016] As a further solution of the present invention, in step three, the image preprocessing operation includes denoising, contrast enhancement, and brightness adjustment.

[0017] As a further solution of the present invention, in step 4, the preliminary recognition feature points include the shape, color, and logo of the battery.

[0018] As a further solution of the present invention, in step 4, the computer vision algorithm used is any one of a grayscale vision algorithm, a scale-invariant feature transformation algorithm, a Haar cascade classifier algorithm, an optical flow algorithm, and a conditional random field.

[0019] As a further solution of the present invention, in step 5, the type and state of the battery pack are determined based on the output result of the classifier;

[0020] If the battery pack meets the battery swap brand and battery swap standards of the battery swap station, the normal battery swap operation will be carried out;

[0021] If the battery pack does not meet the battery swap brand and battery swap standards of the battery swap station, corresponding operation control will be carried out automatically or manually based on the identification results, or manual intervention or alarm processing will be carried out as needed.

[0022] As a further solution of the present invention, in step seven, a machine learning or deep learning algorithm is used to determine whether the battery pack is tilted;

[0023] If the depth values ​​returned by the four feature points are consistent, it is assumed that the battery pack and the RGV unlocking mechanism are vertical and horizontal, and there is no tilt in any direction, then the normal battery replacement operation is performed;

[0024] If the depth values ​​returned by the four feature points are inconsistent, it is determined that the entire battery pack is tilted in a certain direction. Based on the recognition result, the unlocking posture of the RGV unlocking mechanism is adjusted.

[0025] The present invention provides a method for detecting the tilt position and tilt angle of a battery pack of a battery-swapping vehicle, which has the following beneficial effects:

[0026] The battery pack tilt detection method for battery swapping vehicles provided by the present invention can quickly detect the tilt azimuth and tilt angle value of the battery pack, automatically detect and notify the RGV to make adjustments, improve the battery swap efficiency, reduce the battery swap time, and solve the problem of being unable to unlock and swap the battery due to tilt, without the need to manually calculate the tilt angle through measuring tools or other means;

[0027] The present invention performs preliminary identification of the battery pack's shape and contamination, thereby enabling judgment of the battery pack before unlocking, avoiding unnecessary procedures and improving accuracy and unlocking efficiency.

[0028] The vehicle battery pack tilt detection method of the present invention meets the needs of different battery swap models in battery swap stations. It can detect whether the battery packs of different models are tilted relative to the RGV unlocking mechanism hardware equipment. The different depth values ​​returned by the feature points can be used to determine whether a certain direction is tilted. Various offset angle values, including pitch angle, roll angle, yaw angle, etc., can be detected.

[0029] The battery pack tilt detection method for battery-swap vehicles provided by the present invention can transmit the detection azimuth tilt angle value result to the PLC unlocking mechanism in real time. The PLC unlocking mechanism can adjust the unlocking posture according to the detection structure, conduct real-time detection during the adjustment process, and transmit the detection results in real time until the PLC unlocking mechanism and the battery pack are relatively vertical and horizontal. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0031] Figure 1 It is a flowchart of the workflow of the present invention. DETAILED DESCRIPTION

[0032] Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0033] See also Figure 1 The embodiment of the present invention provides a method for detecting the tilt position and tilt angle of a battery pack of a battery-swap vehicle.

[0034] S1, place the camera and 3D camera hardware equipment on the RGV unlocking mechanism and drive under the vehicle.

[0035] S2, in the battery swap area of ​​the battery swap station, uses a camera to capture images of the battery pack. The camera can clearly capture the front, back, and sides of the battery.

[0036] S3 pre-processes the image captured by the camera, including denoising, contrast enhancement, brightness adjustment and other operations to improve the subsequent recognition accuracy.

[0037] S4 selects the outer edge and shape of the battery pack as the preliminary identification feature points of the battery pack, and uses a computer vision algorithm to extract the features of the battery pack from the preprocessed image to confirm whether the battery pack is dirty or damaged. These features may include information such as the shape, color, and logo of the battery.

[0038] S5, inputs the extracted features into the classifier, uses machine learning or deep learning algorithms to classify and identify the battery, and determines whether the battery pack is a battery pack that can be replaced. The classifier is trained and classified according to different battery brands, models or other attributes.

[0039] Based on the output results of the classifier, the type and status of the battery pack are judged. For example, it is determined whether the battery meets the prescribed standards of the battery swap station and whether the battery swap can be performed normally.

[0040] If the battery pack meets the battery swap brand and standards of the battery swap station, normal battery swap operations will be performed.

[0041] If the battery pack does not meet the battery swap brand and battery swap standards of the battery swap station, corresponding operation control will be carried out automatically or manually based on the identification results, or manual intervention or alarm processing will be carried out as needed.

[0042] S6, selecting four lock holes on the top, bottom, left, and right sides of the battery pack as four second feature points, photographing the four second feature points using a 3D depth camera, and extracting returned depth values ​​of the feature points;

[0043] S7, inputting the extracted depth values ​​returned by the four keyholes into a processing terminal, and identifying and comparing the four depth values ​​using a machine learning or deep learning algorithm to determine whether the battery pack is tilted;

[0044] If the depth values ​​returned by the four feature points are consistent, it is assumed that the battery pack and the RGV unlocking mechanism are vertical and horizontal, and there is no tilt in any direction, then the normal battery replacement operation is performed;

[0045] If the depth values ​​returned by the four feature points are inconsistent, it is determined that the entire battery pack is tilted in a certain direction. Based on the recognition result, the unlocking posture of the RGV unlocking mechanism is adjusted.

[0046] S8 uses machine learning or deep learning algorithms to determine the tilt orientation of the battery pack, and transmits the detection results to the PLC host and RGV unlocking mechanism through relevant protocols, and adjusts the unlocking posture of the RGV unlocking mechanism until the PLC unlocking mechanism and the battery pack are relatively vertical and horizontal to ensure that the four unlocking gun heads on the RGV unlocking mechanism correspond to the four lock holes on the battery pack.

[0047] S9, unlock the battery pack through the RGV unlocking mechanism and replace the battery.

[0048] During use, the present invention selects the outer edge of the battery pack and the shape of the battery pack as the initial recognition feature points of the battery pack, and uses the Haar cascade classifier algorithm to

[0049] The Haar cascade classifier algorithm formula is as follows:

[0050] H(x) = ∑(w * f(x))

[0051] Among them, H(x) represents the output result of the cascade classifier, w is the weight, and f(x) is the feature vector;

[0052] Extract the battery pack features from the preprocessed image and determine whether the battery pack is dirty or damaged. If not, proceed to the next step. If so, automatically or manually intervene to perform corresponding operational controls. The extracted features are then input into a classifier, which uses machine learning or deep learning algorithms to classify the batteries and determine whether the battery pack can be replaced.

[0053] Then, the positions of the four lock holes on the upper, lower, left and right sides of the battery pack are selected as reference points for the depth values. The depth values ​​returned by the four depth feature points are consistent. When the depth values ​​returned by the four feature points are consistent, it is assumed that the battery pack is vertically horizontal to the RGV unlocking mechanism, and there is no tilt in a certain direction. If the depth values ​​returned by the four feature points are inconsistent, it is determined whether the entire battery pack is tilted in a certain direction. For example, the depth values ​​returned by the upper left and lower left feature points are inconsistent with the depth values ​​returned by the upper right / lower right feature points. The depth values ​​returned by the upper right and lower right sides are the distance between the battery pack and the RGV unlocking mechanism relative to the vertical level. It can be determined that the left side of the battery pack is tilted relative to the unlocking mechanism. When the vehicle enters the designated battery swap area and the vehicle is lifted, the RGV unlocking mechanism hardware device drives into the default position under the vehicle. The RGV posture at this position is the default posture. The battery pack at the default vehicle parking position is vertically horizontal to the RGV unlocking mechanism. When one side is detected to be tilted, the tilt direction and tilt angle are given to the unlocking mechanism. The unlocking mechanism adjusts its posture according to the tilt direction and angle value. Finally, the RGV unlocking mechanism is used to perform the battery swap operation on the battery pack.

[0054] The above are merely embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should all be included within the scope of the claims of the present application.

Claims

1. A method for detecting the tilt position and tilt angle of a battery pack of a battery-swap vehicle, characterized in that: The following steps are involved: Step 1: Place the camera and 3D camera hardware on the RGV unlocking mechanism and drive under the vehicle; Step 2: In the battery swapping area of ​​the battery swapping station, use a camera to capture an image of the battery pack. The camera can clearly capture the front, back, and sides of the battery. Step 3: Preprocess the image captured by the camera to improve the subsequent recognition accuracy; Step 4: Select the outer edge and shape of the battery pack as the initial identification feature points of the battery pack. Use a computer vision algorithm to extract the battery pack features from the pre-processed image to determine whether the battery pack is dirty or damaged. Step 5: Input the extracted features into a classifier, and use machine learning or deep learning algorithms to classify and identify the battery pack to determine whether the battery pack is replaceable. The classifier is trained and classified based on different battery brands, models, or other attributes; Step 6: Select the four lock holes on the top, bottom, left, and right sides of the battery pack as four second feature points, and use a 3D depth camera to photograph the four second feature points and extract the returned depth values ​​of the feature points; Step 7: Input the extracted depth values ​​returned by the four keyholes into the processing terminal, and identify and compare the four depth values ​​through machine learning or deep learning algorithms to determine whether the battery pack is tilted; Step 8: Use machine learning or deep learning algorithms to determine the tilt orientation of the battery pack, and transmit the judgment result to the PLC host and RGV unlocking mechanism through relevant protocols, and adjust the unlocking posture of the RGV unlocking mechanism until the PLC unlocking mechanism and the battery pack are relatively vertical and horizontal to ensure that the four unlocking gun heads on the RGV unlocking mechanism correspond to the four lock holes on the battery pack; Step nine: Unlock the battery pack through the RGV unlocking mechanism and replace the battery.

2. A method for detecting the tilt position and tilt angle of a battery pack of a battery-swap vehicle according to claim 1, characterized in that: In the step three, the image preprocessing operations include denoising, contrast enhancement, and brightness adjustment.

3. The method for detecting the tilt position and tilt angle of a battery pack of a battery-swap vehicle according to claim 1, characterized in that: In the step 4, the preliminary recognition feature points include the shape, color, and logo of the battery.

4. The method for detecting the tilt position and tilt angle of a battery pack of a battery-swap vehicle according to claim 3, characterized in that: In the step 4, the computer vision algorithm used is any one of a grayscale vision algorithm, a scale-invariant feature transformation algorithm, a Haar cascade classifier algorithm, an optical flow algorithm, and a conditional random field.

5. The method for detecting the tilt position and tilt angle of a battery pack of a battery-swap vehicle according to claim 4, characterized in that: In step five, the type and state of the battery pack are determined based on the output of the classifier; If the battery pack meets the battery swap brand and battery swap standards of the battery swap station, the normal battery swap operation will be carried out; If the battery pack does not meet the battery swap brand and battery swap standards of the battery swap station, corresponding operation control will be carried out automatically or manually based on the identification results, or manual intervention or alarm processing will be carried out as needed.

6. The method for detecting the tilt position and tilt angle of a battery pack of a battery-swap vehicle according to claim 1, characterized in that: In step seven, determining whether the battery pack is tilted is performed using a machine learning or deep learning algorithm; If the depth values ​​returned by the four feature points are consistent, it is assumed that the battery pack and the RGV unlocking mechanism are vertical and horizontal, and there is no tilt in any direction, then the normal battery replacement operation is performed; If the depth values ​​returned by the four feature points are inconsistent, it is determined that the entire battery pack is tilted in a certain direction. Based on the recognition result, the unlocking posture of the RGV unlocking mechanism is adjusted.

Images