Anti-pinch control method, system and equipment for automobile active air inlet grille and storage medium

Abstract

The invention discloses an automobile active air inlet grille anti-pinch control method, system and device and a storage medium. The method comprises the steps that visual data and radar data within a preset range of an automobile active air inlet grille are obtained; performing data fusion on the visual data and the radar data through an extended Kalman filtering algorithm to obtain target object position information and motion track prediction information; carrying out danger early warning analysis on the target object position information and the motion trail prediction information; determining an anti-pinch early warning grade according to the danger early warning analysis result, and obtaining an anti-pinch early warning strategy corresponding to the anti-pinch early warning grade; and based on the anti-pinch early warning strategy, anti-pinch control is conducted on the automobile active air inlet grille according to the motor torque. Vision and radar are combined, the monitoring precision and robustness are improved through a Kalman filtering algorithm, then a graded early warning mechanism and the motor torque are adopted, progressive response from early warning to prevention is achieved, and the clamping risk caused by the fact that a target object enters a grid gap by mistake in the vehicle charging process is solved.

Description

Technical Field

[0001] This invention relates to the field of active grille shutter anti-pinch technology, and in particular to an active grille shutter anti-pinch control method, system, device and storage medium for automobiles. Background Technology

[0002] With the development of automotive intelligence, active grille shutters have become a key component for improving vehicle aerodynamic performance and thermal management efficiency. Traditional active grille shutter control systems mainly adjust the blade opening based on parameters such as vehicle speed and coolant temperature, lacking active protection mechanisms for personal safety. Existing anti-pinch technologies mostly rely on physical isolation or passive triggering, resulting in insufficient response speed and difficulty in handling special scenarios where the vehicle may be clamped while charging and stationary.

[0003] Therefore, how to solve the risk of objects being trapped in the gaps of the grille during vehicle charging has become an urgent problem to be solved.

[0004] The above content is only used to help understand the technical solution of the present invention and does not represent an admission that the above content is prior art. Summary of the Invention The main objective of this invention is to provide an active grille shutter anti-pinch control method, system, device, and storage medium for automobiles, aiming to solve the technical problem of the risk of being pinched when a target object accidentally enters the grille gap during vehicle charging.

[0005] To achieve the above objectives, the present invention provides an active grille shutter anti-pinch control method for automobiles, the active grille shutter anti-pinch control method comprising: During the vehicle charging process, visual and radar data within a preset range of the vehicle's active grille shutter are acquired; The visual data and the radar data are fused using the extended Kalman filter algorithm to obtain the target's location information and motion trajectory prediction information. A hazard warning analysis is performed on the target object's location information and the predicted motion trajectory information; The anti-pinch warning level is determined based on the hazard warning analysis results, and the anti-pinch warning strategy corresponding to the anti-pinch warning level is obtained. Based on the aforementioned anti-pinch warning strategy, the active air intake grille of the vehicle is controlled to prevent pinching according to the motor torque.

[0006] Optionally, acquiring visual and radar data within a preset range of the vehicle's active grille shutter includes: The vehicle uses an onboard camera to capture real-time images of the front of the car's active grille. The target object feature is identified in the foreground image, and the identified target object feature information is used as visual data. The relative information between the vehicle's active grille and the target object within a preset range is obtained using millimeter-wave radar. The relative information includes distance, speed, and angle. The distance, speed, and angle are used as radar data.

[0007] Optionally, the target feature detection of the foreground image includes: The foreground image is preprocessed, and the target object features are identified in the preprocessed image using a preset recognition model. The preset recognition module is built based on the MobileNet framework.

[0008] Optionally, the step of fusing the visual data and the radar data using an extended Kalman filter algorithm to obtain target location information and motion trajectory prediction information includes: A state vector is constructed based on the visual data and the radar data; Based on the state vector, state prediction is performed to obtain the state prediction matrix and the covariance prediction matrix. Visual updates are performed based on the state prediction matrix and the covariance prediction matrix to obtain the updated state matrix and covariance matrix. The radar is updated based on the updated state matrix and covariance matrix to obtain the optimal state estimation result. The target object's location information and trajectory prediction information are determined based on the optimal state estimation results.

[0009] Optionally, the step of performing hazard warning analysis on the target location information and the motion trajectory prediction information includes: Based on the target location information and the motion trajectory prediction information, determine whether the target object is in a dangerous area; If so, the radial distance between the target object and the vehicle's active air intake grille is determined, and a hazard warning analysis is performed on the radial distance.

[0010] Optionally, obtaining the anti-pinch warning strategy corresponding to the anti-pinch warning level includes: A hierarchical early warning mechanism mapping table is constructed, in which there is a one-to-one correspondence between the anti-pinch early warning level and the anti-pinch early warning mechanism; According to the anti-pinch warning level, the corresponding anti-pinch warning mechanism is matched through the hierarchical warning mechanism mapping table, and the matched anti-pinch warning mechanism is used as the anti-pinch warning strategy corresponding to the anti-pinch warning level.

[0011] Optionally, the anti-pinch warning strategy based on the motor torque for controlling the active air intake grille of the vehicle to prevent pinching includes: The anti-pinch early warning strategy is based on real-time monitoring of the motor torque of the DC motor driving the grid blades; Determine whether the motor torque meets the preset clamping conditions; If so, the motor will stop running and the blade rotation direction will be reversed until the target object is released.

[0012] Furthermore, to achieve the above objectives, the present invention also proposes an active grille shutter anti-pinch control system for automobiles, the active grille shutter anti-pinch control system comprising: The perception layer is used to acquire visual and radar data within a preset range of the vehicle's active grille during the vehicle charging process. The decision layer is used to fuse the visual data and the radar data using the extended Kalman filter algorithm to obtain target location information and motion trajectory prediction information. The decision-making layer is also used to perform hazard warning analysis on the target object's location information and the motion trajectory prediction information; The decision-making layer is also used to determine the anti-pinch warning level based on the danger warning analysis results, and to obtain the anti-pinch warning strategy corresponding to the anti-pinch warning level. An execution layer is used to perform anti-pinch control on the vehicle's active air intake grille based on the motor torque according to the anti-pinch warning strategy.

[0013] Furthermore, to achieve the above objectives, the present invention also proposes an active air intake grille anti-pinch control device for automobiles, the device comprising: a memory, a processor, and an active air intake grille anti-pinch control program stored in the memory and executable on the processor, the active air intake grille anti-pinch control program being configured to implement the steps of the active air intake grille anti-pinch control method for automobiles as described above.

[0014] In addition, to achieve the above objectives, the present invention also proposes a storage medium storing an active grille shutter anti-pinch control program for automobiles, wherein when the active grille shutter anti-pinch control program is executed by a processor, the program implements the steps of the active grille shutter anti-pinch control method for automobiles as described above.

[0015] This invention first acquires visual and radar data within a preset range of the vehicle's active air intake grille. Then, it fuses the visual and radar data using an extended Kalman filter (EKF) algorithm to obtain target location information and trajectory prediction information. A hazard warning analysis is then performed on this target location and trajectory prediction information. Based on the hazard warning analysis results, an anti-pinch warning level is determined, and a corresponding anti-pinch warning strategy is obtained. Finally, based on the anti-pinch warning strategy, the active air intake grille is controlled to prevent pinching according to the motor torque. This invention combines vision and radar, improves monitoring accuracy and robustness through the EKF algorithm, and then employs a tiered warning mechanism and motor torque to achieve a progressive response from warning to prevention, thus mitigating the risk of objects accidentally entering the grille gap during vehicle charging. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the structure of the automotive active grille shutter anti-pinch control device in the hardware operating environment involved in the embodiments of the present invention; Figure 2 This is a flowchart illustrating the first embodiment of the active grille shutter anti-pinch control method for automobiles according to the present invention; Figure 3 This is a structural block diagram of the first embodiment of the active grille shutter anti-pinch control system for automobiles of the present invention.

[0017] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0018] It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the invention.

[0019] Reference Figure 1 , Figure 1 This is a schematic diagram of the structure of an active grille shutter anti-pinch control device for automobiles, which is part of the hardware operating environment of the embodiment of the present invention.

[0020] like Figure 1As shown, the active grille shutter anti-pinch control device for automobiles may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. The communication bus 1002 is used to enable communication between these components. The user interface 1003 may include a display screen or an input unit such as a keyboard; optionally, the user interface 1003 may also include a standard wired interface or a wireless interface. The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a Wireless-Fidelity (Wi-Fi) interface). The memory 1005 may be high-speed random access memory (RAM) or stable non-volatile memory (NVM), such as a disk storage device. Optionally, the memory 1005 may also be a storage system independent of the aforementioned processor 1001.

[0021] Those skilled in the art will understand that Figure 1 The structure shown does not constitute a limitation on the active grille shutter control device for automobiles, and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0022] like Figure 1 As shown, the memory 1005, which serves as a storage medium, may include an operating system, a network communication module, a user interface module, and an active grille shutter anti-pinch control program for automobiles.

[0023] exist Figure 1 In the illustrated active grille shutter anti-pinch control device for automobiles, the network interface 1004 is mainly used for data communication with the network server; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 and memory 1005 in the active grille shutter anti-pinch control device of the present invention can be set in the active grille shutter anti-pinch control device for automobiles. The active grille shutter anti-pinch control device for automobiles calls the active grille shutter anti-pinch control program stored in the memory 1005 through the processor 1001 and executes the active grille shutter anti-pinch control method for automobiles provided in the embodiments of the present invention.

[0024] This invention provides a method for controlling the active grille shutter anti-pinch mechanism in automobiles, referring to... Figure 2 , Figure 2 This is a flowchart illustrating the first embodiment of the active grille shutter anti-pinch control method for automobiles according to the present invention.

[0025] In this embodiment, the active grille shutter anti-pinch control method for automobiles includes the following steps: S1 acquires visual and radar data within a preset range of the vehicle's active grille shutter during the vehicle charging process.

[0026] It is easy to understand that the executing entity of this embodiment can be an active grille shutter control system for automobiles with functions such as data processing, network communication and program operation, or other computer devices with similar functions. This embodiment does not limit it.

[0027] Furthermore, the processing method for acquiring visual and radar data within a preset range of the vehicle's active air intake grille is as follows: real-time acquisition of the front image of the vehicle's active air intake grille via an onboard camera; target feature recognition of the front image, and use the recognized target feature information as visual data; acquisition of relative information between the vehicle's active air intake grille and the target within a preset range via millimeter-wave radar, the relative information including distance, speed, and angle; and use the distance, speed, and angle as radar data.

[0028] In the specific implementation, it is necessary to acquire the front image of the car's active air intake grille in real time through the camera at the front of the car, then preprocess the front image, and use a preset recognition model to identify the target features of the preprocessed image. The preset recognition module is built based on the MobileNet framework.

[0029] It should also be noted that the target object can be the user's hand, or other pre-set obstacles.

[0030] In this embodiment, the target object is a hand as an example. A lightweight convolutional neural network is constructed based on the MobileNet framework. Its input is a preprocessed image of the front and its output is hand feature information, which includes shape, corresponding bounding box, and confidence score.

[0031] In the specific implementation, the preprocessing method for the foreground image is as follows: 1. Fast noise reduction and filtering: Sensor noise (Gaussian noise, salt and pepper noise) and subtle environmental interference can be removed by bilateral filtering, while preserving the target edge to the maximum extent. The bilateral filtering method can be efficiently implemented using libraries such as OpenCV.

[0032] 2. Adaptive Lighting Compensation: By limiting contrast adaptive histogram equalization, it solves the problem of local underexposure or overexposure of images caused by backlighting, shadows, insufficient light at night, etc., and balances the overall contrast.

[0033] The specific implementation method is as follows: divide the image into several small grids (such as 8x8), and perform histogram equalization on each grid to enhance local contrast.

[0034] To prevent local noise from being excessively amplified, a limiting threshold (e.g., 2.0) is set, and portions of the histogram exceeding this threshold are cropped and redistributed. This is key to avoiding "blocky noise" in uniform regions.

[0035] Finally, bilinear interpolation is used to eliminate boundary effects between meshes.

[0036] 3. Color space conversion and target feature enhancement: Convert the image to a color space that better highlights the features of the "hand" and perform targeted enhancement.

[0037] step: 3.1 Convert to RGB: Convert the camera's original YUV422 format to standard RGB.

[0038] 3.2 Secondary Conversion: Based on the strategy, the conversion can be selected to: The YCbCr color space, a two-dimensional subspace composed of Cb (blue component) and Cr (red component), is a classic region for skin color detection. Skin-like areas can be non-linearly stretched within this space to make them more prominent.

[0039] HSV / HSL color space: The hue (H) channel is relatively insensitive to changes in light intensity, making it more suitable for color-based segmentation. It can be used to stably extract skin tone ranges.

[0040] Motion history enhancement (optional): If processing consecutive frames, you can calculate inter-frame differences or construct a motion history image and use motion-salient regions as weights to overlay them with features of the original image (such as edges or colors) to highlight a moving hand against a static background.

[0041] 4. Perspective correction and coordinate normalization: Eliminate perspective distortion caused by the camera installation angle and map the image coordinates to a real-world two-dimensional plane coordinate system based on the grid.

[0042] The implementation method is as follows: 4.1 Offline calibration: After system installation, use a checkerboard calibration board to determine the camera's intrinsic parameter matrix (focal length, optical center) and extrinsic parameter matrix (rotation and translation relative to the grid plane).

[0043] 4.2 Online Transformation: For each frame of image, the transformation matrix H is calculated using the above parameters to distort the image from the "camera view" to the "bird's-eye view" or the "view from the grid plane".

[0044]

[0045] In the formula, (u,v) are the original image coordinates, and (x',y') are the corrected planar coordinates, so that the pixel distance in the image output by the preprocessing unit can directly correspond to the centimeter distance in the physical world (for example, 10 pixels correspond to 1cm).

[0046] The defined “danger zone” (such as 10cm away from the grid) can be drawn directly on the image with a fixed pixel radius.

[0047] The target position (x, y) detected by vision can be directly used for EKF fusion with radar data (unit: meters / cm) without the need for complex unit conversion and coordinate alignment.

[0048] 5. Image standardization and tensor construction: Converting images into a standardized input format required by subsequent neural network models.

[0049] step: 5.1 Scaling: Use bilinear interpolation to scale the corrected image to the network input size (e.g., 224x224 pixels).

[0050] 5.2 Pixel value normalization: Normalize pixel values ​​from [0, 255] to the range of [-1, 1] or [0, 1] to accelerate network convergence.

[0051] 5.3 Channel Arrangement: According to the requirements of the deep learning framework, the image data is arranged into a tensor in the format of (C, H, W) or (H, W, C) (C = number of channels, H = height, W = width).

[0052] The output of the preprocessing unit is a normalized four-dimensional tensor: (BatchSize, Channels, Height, Width), for example (1, 3, 224, 224). This tensor will be directly fed into a lightweight CNN (such as MobileNet) for hand object detection.

[0053] It should also be noted that by preprocessing the lighting, color, and noise, the detection model can obtain stable input under various extreme environments, which greatly improves the system's all-weather working capability.

[0054] In this embodiment, by removing interference and enhancing the target in the preprocessing stage, the recognition difficulty of the subsequent neural network model is reduced, thereby allowing the use of a lighter and faster model (such as MobileNet) to meet the real-time requirements of the vehicle system.

[0055] The preset recognition model adopts a two-stage design of "backbone feature extraction network + lightweight detection head", which is optimized to balance speed and accuracy.

[0056] Backbone network: MobileNetv3, efficiently extracting multi-level features from input images (e.g., 224x224 pixels). It uses depthwise separable convolutions as basic units. It decomposes standard convolution into "channel-wise convolution" and "pointwise convolution," significantly reducing computation and parameters. It can also incorporate attention mechanisms (SE modules) and network structures designed based on Neural Architecture Search (NAS), improving feature quality while reducing computation.

[0057] The detection head, SSDLite, receives feature maps from multiple scales in the backbone and directly predicts the location and confidence level of the hand region. It is a lightweight version of the standard SSD detector. It uses depthwise separable convolutions instead of all ordinary convolutions to build the detection layers, further reducing the number of parameters.

[0058] The entire workflow can be summarized as follows: Image → Feature extraction from MobileNetv3 backbone → Multi-scale prediction by SSDLite detector head → Output of hand feature information.

[0059] The Compressed-Excited (SE) attention module integrated in MobileNetv3 can adaptively calibrate the channel feature response, allowing the network to focus more on key hand-related features and improve accuracy.

[0060] It should be understood that after identifying hand features, it is necessary to use millimeter-wave radar to detect the distance, speed and angle of the target within a preset range (e.g., within 0-50cm in front of the grille) to compensate for the blind spots of the vision system in low light and bad weather.

[0061] S2, the visual data and the radar data are fused using the extended Kalman filter algorithm to obtain target location information and motion trajectory prediction information.

[0062] Furthermore, a state vector is constructed based on visual and radar data; state prediction is performed based on the state vector to obtain a state prediction matrix and a covariance prediction matrix; visual updates are performed based on the state prediction matrix and the covariance prediction matrix to obtain an updated state matrix and a covariance matrix; radar updates are performed based on the updated state matrix and the covariance matrix to obtain the optimal state estimation result; and the target's position information and trajectory prediction information are determined based on the optimal state estimation result.

[0063] In the specific implementation, , For state vectors, and The position of the target object on a two-dimensional plane in physical world coordinates. and Let be the velocity of the target object in the two-dimensional plane in the physical world coordinate system, and k be the k-th discrete time step. , .

[0064] Camera observation vector: In the formula and These are the measured values ​​of camera image pixel coordinates mapped to the physical world coordinate system after calibration and coordinate transformation.

[0065] Radar observation vector: In the formula The radial distance to the target as measured by radar. The azimuth angle of the target as measured by radar.

[0066] Assume the target moves at approximately a constant velocity in a straight line for a short period of time: State prediction matrix :

[0067] Wherein, the state transition matrix is:

[0068] In the formula, This represents the sampling time interval.

[0069] Covariance prediction matrix :

[0070] Where P is the covariance matrix of the state estimation error, representing the estimation uncertainty, and Q is the process noise covariance matrix, representing the uncertainty of the motion model (such as sudden acceleration).

[0071] Visual update (linear observation): 1. Kalman gain:

[0072] The camera observation matrix includes:

[0073] 2. Status Update:

[0074] 3. Covariance Update:

[0075] Radar update (nonlinear observation, EKF linearization): Use the updated and Alternative and This is used as input for the radar update step in the calculation.

[0076] 1. Nonlinear observation function:

[0077] 2. Jacobian Matrix

[0078] 3. Kalman gain

[0079] In the formula, This is the radar observation noise covariance matrix.

[0080] 4. Status Update:

[0081] 5. Covariance Update:

[0082] In practical implementation, within the Kalman filter framework for multi-sensor fusion, the optimal fused state estimate is continuously output. It includes target location information and motion trajectory prediction information (i.e. finger prediction trajectory), and will be dynamically updated as new observation data is continuously input.

[0083] S3, perform hazard warning analysis on the target object's location information and the predicted motion trajectory information.

[0084] Furthermore, based on the target location information and motion trajectory prediction information, it is determined whether the target is in a dangerous area; if so, the radial distance between the target and the vehicle's active air intake grille is determined, and a hazard warning analysis is performed on the radial distance.

[0085] It should also be noted that a fan-shaped area with a radial distance of ≤10cm centered on the grid blade can be defined as a danger zone. When a target object is detected within the defined fan-shaped area, it is determined that the target object is in the danger zone.

[0086] When the radial distance is detected to be within the preset range (e.g., 5cm-10cm), the anti-pinch warning level is determined to be Level 1.

[0087] When the radial distance is detected to be within a preset range (e.g., distance < 5cm), the anti-pinch warning level is determined to be a Level 2 warning.

[0088] The preset range is a user-defined setting, and this embodiment does not impose any limitations on it.

[0089] S4. Determine the anti-pinch warning level based on the danger warning analysis results, and obtain the anti-pinch warning strategy corresponding to the anti-pinch warning level.

[0090] It should be noted that the hazard warning analysis results include radial distance and anti-pinch warning level.

[0091] In the specific implementation, a hierarchical early warning mechanism mapping table is constructed, in which there is a one-to-one correspondence between the anti-pinch early warning level and the anti-pinch early warning mechanism. Based on the anti-pinch early warning level, the corresponding anti-pinch early warning mechanism is matched through the hierarchical early warning mechanism mapping table, and the matched anti-pinch early warning mechanism is used as the anti-pinch early warning strategy corresponding to the anti-pinch early warning level.

[0092] It should be understood that the anti-pinch warning level and anti-pinch warning mechanism in the hierarchical warning mechanism mapping table can be customized by the user.

[0093] For example, the anti-pinch warning mechanism corresponding to the first-level warning can trigger an audible and visual alarm (short whistle) to remind people to stay away; the anti-pinch warning mechanism corresponding to the second-level warning can automatically reduce the blade movement speed to 50% of the normal speed and enter the anti-pinch preparation state in advance.

[0094] S5, based on the anti-pinch warning strategy, the active air intake grille of the vehicle is controlled to prevent pinching according to the motor torque.

[0095] Furthermore, based on the anti-pinch warning strategy, the motor torque of the DC motor driving the grille blades is monitored in real time; it is determined whether the motor torque meets the preset clamping conditions; if so, the motor is controlled to stop running and the rotation direction of the blades is reversed (e.g., the reverse angle is ≥30°) until the target object is released, and a notification is sent to the owner's mobile APP via the CAN bus; if not, the anti-pinch warning strategy is returned to be executed until the vehicle is far away from the danger zone.

[0096] In practice, after the anti-pinch warning strategy is implemented, the motor torque of the DC motor driving the grid blades is monitored in real time. The DC motor can be a permanent magnet brushed DC motor or a brushless DC motor.

[0097] The preset clamping condition is that when the torque exceeds a threshold (e.g., 1.5 times the normal torque, which can be dynamically adjusted based on historical data) and lasts for more than 100ms for a preset time, and the target object is confirmed to be in a dangerous area by the sensor, it is determined to be a clamping event.

[0098] In this embodiment, visual and radar data within a preset range of the vehicle's active air intake grille are first acquired. Then, the visual and radar data are fused using an extended Kalman filter algorithm to obtain target location information and motion trajectory prediction information. A hazard warning analysis is then performed on the target location information and motion trajectory prediction information. Following this, an anti-pinch warning level is determined based on the hazard warning analysis results, and a corresponding anti-pinch warning strategy is obtained. Finally, based on the anti-pinch warning strategy, the active air intake grille is controlled to prevent pinching according to the motor torque. This embodiment combines vision and radar, improves monitoring accuracy and robustness through the EKF algorithm, and then employs a tiered warning mechanism and motor torque to achieve a progressive response from warning to prevention, addressing the risk of trapping caused by a target object accidentally entering the grille gap during vehicle charging.

[0099] Reference Figure 3 , Figure 3 This is a structural block diagram of the first embodiment of the active grille shutter anti-pinch control system for automobiles of the present invention.

[0100] like Figure 3 As shown, the active grille shutter anti-pinch control system for automobiles proposed in this embodiment of the invention includes: The perception layer 3001 is used to acquire visual and radar data within a preset range of the vehicle's active air intake grille during the vehicle charging process. The decision layer 3002 is used to perform data fusion on the visual data and the radar data through the extended Kalman filter algorithm to obtain target location information and motion trajectory prediction information. The decision layer 3002 is also used to perform hazard warning analysis on the target object's location information and the motion trajectory prediction information; The decision layer 3002 is also used to determine the anti-pinch warning level based on the danger warning analysis results, and to obtain the anti-pinch warning strategy corresponding to the anti-pinch warning level. Execution layer 3003 is used to perform anti-pinch control on the active air intake grille of the vehicle based on the motor torque according to the anti-pinch warning strategy.

[0101] Other embodiments or specific implementations of the active air intake grille anti-pinch control system of the present invention can be referred to the above-described method embodiments, and will not be repeated here.

[0102] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0103] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0104] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as read-only memory / random access memory, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0105] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.

Claims

1. A method for controlling the active grille shutter anti-pinch function in automobiles, characterized in that, The method includes the following steps: During the vehicle charging process, real-time visual and radar data within a preset range of the vehicle's active grille shutter are acquired. The visual data and the radar data are fused using the extended Kalman filter algorithm to obtain the target's location information and motion trajectory prediction information. A hazard warning analysis is performed on the target object's location information and the predicted motion trajectory information; The anti-pinch warning level is determined based on the hazard warning analysis results, and the anti-pinch warning strategy corresponding to the anti-pinch warning level is obtained. Based on the aforementioned anti-pinch warning strategy, the active air intake grille of the vehicle is controlled to prevent pinching according to the motor torque.

2. The method as described in claim 1, characterized in that, The acquisition of visual and radar data within a preset range of the vehicle's active grille shutter includes: The vehicle uses an onboard camera to capture real-time images of the front of the car's active grille. The target object feature is identified in the foreground image, and the identified target object feature information is used as visual data. The relative information between the vehicle's active grille and the target object within a preset range is obtained using millimeter-wave radar. The relative information includes distance, speed, and angle. The distance, speed, and angle are used as radar data.

3. The method as described in claim 2, characterized in that, The target feature detection of the foreground image includes: The foreground image is preprocessed, and the target object features are identified in the preprocessed image using a preset recognition model. The preset recognition module is built based on the MobileNet framework.

4. The method as described in claim 1, characterized in that, The process of fusing the visual data and radar data using an extended Kalman filter algorithm to obtain target location information and motion trajectory prediction information includes: A state vector is constructed based on the visual data and the radar data; Based on the state vector, state prediction is performed to obtain the state prediction matrix and the covariance prediction matrix. Visual updates are performed based on the state prediction matrix and the covariance prediction matrix to obtain the updated state matrix and covariance matrix. The radar is updated based on the updated state matrix and covariance matrix to obtain the optimal state estimation result. The target object's location information and trajectory prediction information are determined based on the optimal state estimation results.

5. The method as described in claim 1, characterized in that, The hazard warning analysis of the target location information and the motion trajectory prediction information includes: Based on the target location information and the motion trajectory prediction information, determine whether the target object is in a dangerous area; If so, the radial distance between the target object and the vehicle's active air intake grille is determined, and a hazard warning analysis is performed on the radial distance.

6. The method as described in claim 1, characterized in that, The step of obtaining the anti-pinch warning strategy corresponding to the anti-pinch warning level includes: A hierarchical early warning mechanism mapping table is constructed, in which there is a one-to-one correspondence between the anti-pinch early warning level and the anti-pinch early warning mechanism; According to the anti-pinch warning level, the corresponding anti-pinch warning mechanism is matched through the hierarchical warning mechanism mapping table, and the matched anti-pinch warning mechanism is used as the anti-pinch warning strategy corresponding to the anti-pinch warning level.

7. The method as described in claim 1, characterized in that, The anti-pinch warning strategy based on the motor torque for anti-pinch control of the vehicle's active air intake grille includes: The anti-pinch early warning strategy is based on real-time monitoring of the motor torque of the DC motor driving the grid blades; Determine whether the motor torque meets the preset clamping conditions; If so, the motor will stop running and the blade rotation direction will be reversed until the target object is released.

8. A vehicle active grille shutter anti-pinch control system, characterized in that, The system includes: The perception layer is used to acquire visual and radar data within a preset range of the vehicle's active grille during the vehicle charging process. The decision layer is used to fuse the visual data and the radar data using the extended Kalman filter algorithm to obtain target location information and motion trajectory prediction information. The decision-making layer is also used to perform hazard warning analysis on the target object's location information and the motion trajectory prediction information; The decision-making layer is also used to determine the anti-pinch warning level based on the danger warning analysis results, and to obtain the anti-pinch warning strategy corresponding to the anti-pinch warning level. An execution layer is used to perform anti-pinch control on the vehicle's active air intake grille based on the motor torque according to the anti-pinch warning strategy.

9. A vehicle active grille shutter anti-pinch control device, characterized in that, The device includes: a memory, a processor, and an active grille shutter anti-pinch control program for automobiles stored in the memory and executable on the processor, the active grille shutter anti-pinch control program for automobiles configured to implement the steps of the active grille shutter anti-pinch control method for automobiles as described in any one of claims 1 to 7.

10. A storage medium, characterized in that, The storage medium stores an active grille shutter anti-pinch control program for automobiles, which, when executed by a processor, implements the steps of the active grille shutter anti-pinch control method for automobiles as described in any one of claims 1 to 7.

Images