Vehicle tailgate adaptive control method, device, vehicle and program product

Abstract

The application discloses a vehicle tail door self-adaptive control method and device, a vehicle and a program product, and belongs to the technical field of vehicle control. The method comprises the following steps: acquiring multi-modal data of a target object; performing intention recognition according to the multi-modal data, and determining that the target object has a tail door operation intention; wherein the tail door operation intention represents the intention of the target object to open the tail door of a target vehicle; performing parameter calculation according to the multi-modal data, and obtaining tail door control parameters of the target vehicle; and performing control processing on the tail door of the target vehicle according to the tail door control parameters. When it is detected that the target object has the intention to open the tail door of the vehicle, the control parameters of the tail door of the vehicle are adaptively adjusted according to the multi-modal information associated with the target object, and the tail door of the vehicle is controlled in an optimal manner to be actively opened, so that the control precision of the tail door of the vehicle is effectively improved, and the use experience of the tail door of the user is improved.

Description

Technical Field

[0001] This application relates to the field of vehicle control technology, and in particular to a vehicle tailgate adaptive control method, device, vehicle, and program product. Background Technology

[0002] Currently, vehicles generally support multiple tailgate opening methods, such as interior button opening, exterior door handle opening, remote key opening, sensor-activated projection foot opening, and voice control opening. These diverse tailgate opening methods, to a certain extent, meet users' operational needs for the vehicle's tailgate in different usage scenarios. However, most existing tailgate opening methods can only achieve a binary action of opening and closing the tailgate, or control the tailgate to execute a single movement trajectory according to a preset opening degree after opening. They cannot dynamically adjust the tailgate's control parameters, such as opening angle and opening speed, according to actual conditions, resulting in low control precision of the vehicle's tailgate. Summary of the Invention

[0003] The main objective of this application is to propose an adaptive control method, device, vehicle, and program product for vehicle tailgates, which aims to effectively improve the control accuracy of vehicle tailgates and enhance the user experience of tailgates.

[0004] To achieve the above objectives, one aspect of this application proposes a vehicle tailgate adaptive control method, the method comprising: Obtain multimodal data of the target object; Intent recognition is performed based on the multimodal data to determine that the target object has the intent to operate the tailgate; wherein, the tailgate operation intent means the target object's intent to open the tailgate of the target vehicle; Based on the multimodal data, parameter calculations are performed to obtain the tailgate control parameters of the target vehicle; The tailgate of the target vehicle is controlled according to the tailgate control parameters.

[0005] In some embodiments, the multimodal data includes a sequence of vehicle operating parameters and a sequence of visual images; the step of performing intent recognition based on the multimodal data to determine that the target object has a tailgate operation intent includes: Hand occupancy recognition is performed based on the visual image sequence to obtain the hand load degree sequence of the target object; Pose feature extraction is performed on the visual image sequence to obtain the pose feature sequence of the target object; Intent recognition is performed based on the posture feature sequence, the hand load sequence, and the vehicle operation parameter sequence to determine that the target object has the intention to operate the tailgate.

[0006] In some embodiments, the visual image sequence includes visual images at multiple time points, and the hand load sequence includes hand load values ​​at multiple time points; the step of performing hand occupancy recognition based on the visual image sequence to obtain the hand load sequence of the target object includes: Based on the visual images at each time point, hand occupancy is identified to obtain the hand occupancy duration, hand occupancy area, and arm extension restriction parameters at each time point. The hand load at each specified moment is determined based on the hand occupancy duration, hand occupancy area, and arm extension restriction parameters.

[0007] In some embodiments, the multimodal data includes a sequence of visual images; the tailgate control parameters include a tailgate opening angle; and the step of calculating the tailgate control parameters of the target vehicle based on the multimodal data includes: The relative height of the target object is determined based on the visual image sequence; The tailgate opening angle is obtained by calculating parameters based on the relative height.

[0008] In some embodiments, the multimodal data includes a sequence of visual images; the tailgate control parameters include the tailgate opening time and the tailgate opening angle; the step of calculating the tailgate control parameters of the target vehicle based on the multimodal data includes: Based on the visual image sequence, the distance between the target object and the tailgate is determined as the target distance; The tailgate opening time is obtained by calculating parameters based on the target distance and the tailgate opening angle.

[0009] In some embodiments, the tailgate control parameters further include tailgate opening time, tailgate opening speed, and tailgate opening angle; the step of calculating the tailgate control parameters of the target vehicle based on the multimodal data further includes: The tailgate opening speed is obtained by calculating parameters based on the tailgate opening time and the tailgate opening angle.

[0010] In some embodiments, the method further includes: The tailgate control parameters are corrected based on the obstacle perception data of the target vehicle.

[0011] To achieve the above objectives, another aspect of this application provides a vehicle tailgate adaptive control device, the device comprising: The acquisition module is used to acquire multimodal data of the target object; The first processing module is used to perform intent recognition based on the multimodal data to determine that the target object has the intent to operate the tailgate; wherein, the tailgate operation intent means the intention of the target object to operate the tailgate of the target vehicle; The second processing module is used to perform parameter calculations based on the multimodal data to obtain the tailgate control parameters of the target vehicle. The third processing module is used to control the tailgate of the target vehicle according to the tailgate control parameters.

[0012] To achieve the above objectives, another aspect of this application provides a vehicle, the vehicle comprising: At least one processor; At least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the above-described adaptive control method for the vehicle tailgate.

[0013] To achieve the above objectives, another aspect of this application provides a computer program product, including a computer program that, when executed by a processor, implements the aforementioned adaptive control method for vehicle tailgates.

[0014] According to an embodiment of this application, a vehicle tailgate adaptive control method, device, vehicle, and program product are provided. The method involves acquiring multimodal data of a target object; performing intent recognition based on the multimodal data to determine if the target object intends to operate the tailgate; wherein the tailgate operation intent represents the target object's intention to open the tailgate of the target vehicle; calculating parameters based on the multimodal data to obtain tailgate control parameters for the target vehicle; and controlling the tailgate of the target vehicle according to the tailgate control parameters. According to the technical solution of this application, when the intention of a target object to open the vehicle tailgate is detected, the control parameters of the vehicle tailgate, such as opening angle, opening speed, and opening time, are adaptively adjusted based on the multimodal information associated with the target object, and the vehicle tailgate is controlled to actively open in the optimal manner. This achieves proactive prediction of the user's intention to open the vehicle tailgate and adaptive control of the vehicle tailgate. This enables the tailgate opening process to have individualized adaptation capabilities, accurately matching actual tailgate operation needs, thereby effectively improving the control accuracy of the vehicle tailgate and enhancing the user's tailgate experience.

[0015] Other features and advantages of this application will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the application. The objectives and other advantages of this application may be realized and obtained by means of the structures particularly pointed out in the description, claims and drawings. Attached Figure Description

[0016] Figure 1 This is a flowchart of an adaptive control method for a vehicle tailgate provided in this application; Figure 2 This is a diagram illustrating the specific implementation process of an adaptive control method for a vehicle tailgate provided in this application; Figure 3 This is a structural diagram of a vehicle tailgate adaptive control device provided in this application; Figure 4 This is an example image of a vehicle provided in this application. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit it. In the following description, when referring to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with those of this application; they are merely examples of apparatuses and methods consistent with some aspects of the embodiments of this application as detailed in the appended claims.

[0018] It is understood that the terms “first,” “second,” etc., used in this application may be used herein to describe various concepts, but unless otherwise stated, these concepts are not limited by these terms. These terms are only used to distinguish one concept from another. For example, without departing from the scope of the embodiments of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the words “if,” “when,” or “in response to a determination” as used herein may be interpreted as “when…” or “when…” or “in response to a determination.”

[0019] As used in this application, the terms "at least one", "multiple", "each", "any", etc., "at least one" includes one, two or more, "multiple" includes two or more, "each" refers to each of the corresponding multiples, and "any" refers to any one of the multiples.

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0021] It should be noted that in various specific embodiments of this application, when processing data related to user identity or characteristics, such as multimodal data (e.g., visual images), user permission or consent is obtained first. Furthermore, the collection, use, and processing of this data comply with relevant laws, regulations, and standards. In addition, when embodiments of this application require access to sensitive personal information of users, separate permission or consent from the user is obtained through pop-ups or redirection to confirmation pages. Only after obtaining the user's separate permission or consent is the necessary user-related data required for the proper functioning of these embodiments acquired.

[0022] With the continuous improvement of vehicle intelligence, the methods of opening vehicle tailgates have become increasingly diversified, becoming one of the important indicators for measuring vehicle convenience and user experience. Currently, vehicles generally support multiple tailgate opening methods, such as interior button opening, exterior door handle opening, remote key opening, sensor-activated projection foot opening, and voice control opening. These diversified tailgate opening methods, to a certain extent, meet users' operational needs for the tailgate in different usage scenarios.

[0023] However, most existing tailgate opening methods can only achieve a binary action of opening and closing the tailgate, or control the tailgate to perform a single movement trajectory according to a preset opening degree after opening. Specifically, regardless of how the user opens the tailgate, the vehicle responds with a uniform action strategy, and cannot dynamically adjust the tailgate control parameters, such as opening angle and opening speed, based on information such as the user's height, user position, and the position of surrounding obstacles. This leads to situations such as the tailgate colliding with obstacles, or inconvenience for the user to take out or put down the tailgate due to insufficient opening degree, resulting in low control precision of the vehicle's tailgate.

[0024] To address this, embodiments of this application provide a vehicle tailgate adaptive control method, device, vehicle, and program product. When it detects that a target object intends to open the vehicle tailgate, it adaptively adjusts the control parameters of the vehicle tailgate according to the actual situation, such as the opening angle, opening speed, and opening time, thereby effectively improving the control accuracy of the vehicle tailgate and enhancing the user's experience with the tailgate.

[0025] First, the implementation steps of a vehicle tailgate adaptive control method provided in this application embodiment will be described in detail below with reference to the accompanying drawings.

[0026] This application provides a vehicle tailgate adaptive control method, which can be applied to a terminal, a server, or software running on either a terminal or a server. The terminal can be a tablet, laptop, desktop computer, etc., but is not limited to these. The server can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDN), and big data and artificial intelligence platforms. Furthermore, the server can be a node server in a blockchain network, but is not limited to these. Blockchain is a new application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanisms, and encryption algorithms.

[0027] Reference Figure 1 , Figure 1 This is a flowchart of a vehicle tailgate adaptive control method provided in this application, which may include the following steps S101-S104.

[0028] S101, Obtain multimodal data of the target object.

[0029] It is understood that the target object refers to the object that triggers the tailgate opening behavior. In some embodiments, the target object may specifically be a real user.

[0030] It should be noted that multimodal data may include, but is not limited to, visual image sequences and vehicle operation parameter sequences. Visual image sequences may include visual images from multiple moments, each containing a target object. Vehicle operation parameter sequences may include vehicle operation parameters from multiple moments. Multiple moments refer to the current moment and multiple historical moments preceding the current moment. The vehicle operation parameters are not specifically limited; for example, they may include the signal status of the target vehicle's key, such as a remote control key or Bluetooth key. If the key's location is within a preset range (e.g., 20 meters) of the target vehicle, the key's signal status is assigned a first value (e.g., 1); otherwise, it is assigned a second value (e.g., 0), but this is not a limitation.

[0031] In this step, it is detected whether the target object is close to the target vehicle. If so, the multimodal data of the target object is acquired to provide data input for subsequent processing; otherwise, it returns to the step of detecting whether the target object is close to the target vehicle, thus achieving cyclic detection.

[0032] The method for detecting whether a target object is close to the target vehicle is not specifically limited. For example, it can be to use an in-vehicle camera to perform human detection in the area behind the target vehicle, and at the same time detect whether the key of the target vehicle is within a preset range (such as 20 meters) of the target vehicle. If a human figure is detected in the area behind the target vehicle and the key of the target vehicle is within the preset range of the target vehicle, it is determined that the target object is close to the target vehicle; otherwise, it is determined that the target object is not close to the target vehicle. However, it is not limited to this method.

[0033] S102, based on multimodal data, perform intent recognition to determine that the target object has the intent to operate the tailgate; wherein, the tailgate operation intent means the target object's intent to open the tailgate of the target vehicle.

[0034] In this step, multimodal data is used as a benchmark for intent recognition, thereby determining that the target object has the intent to open the tailgate of the target vehicle. This intent is defined as the tailgate operation intent.

[0035] S103, calculate parameters based on multimodal data to obtain the tailgate control parameters of the target vehicle.

[0036] It should be noted that the tailgate control parameters may include, but are not limited to, tailgate opening angle, tailgate opening speed, and tailgate opening time.

[0037] In this step, after determining that the target object intends to open the tailgate of the target vehicle, multimodal data is used as a reference for parameter calculation. The aim is to adaptively adjust the control parameters of the vehicle tailgate based on the multimodal information associated with the target object, thereby obtaining the tailgate control parameters of the target vehicle, such as opening angle, opening speed, opening time, etc., so that the tailgate state of the target vehicle meets the actual tailgate operation requirements of the target object.

[0038] S104, based on the tailgate control parameters, controls the tailgate of the target vehicle.

[0039] In this step, the tailgate of the target vehicle is controlled based on the tailgate control parameters to ensure that the tailgate opens and reaches a state that matches the tailgate control parameters, thereby achieving high-precision control of the vehicle's tailgate. In some examples, the tailgate control parameters may include, but are not limited to, tailgate opening angle, tailgate opening speed, tailgate opening time, etc., controlling the tailgate to open at the tailgate opening speed and to reach the tailgate opening angle at the tailgate opening time.

[0040] Therefore, when the present application's embodiments detect that a target object intends to open the vehicle's tailgate, they adaptively adjust the control parameters of the vehicle's tailgate, such as opening angle, opening speed, and opening time, based on the multimodal information associated with the target object. Accordingly, they control the vehicle's tailgate to actively open in the optimal manner. This achieves proactive prediction of the user's intention to open the vehicle's tailgate and adaptive control of the vehicle's tailgate. This gives the tailgate opening process an individualized adaptation capability, accurately matching the actual tailgate operation requirements, thereby effectively improving the control accuracy of the vehicle's tailgate and enhancing the user's tailgate experience.

[0041] The steps described above will be explained in further detail below.

[0042] In some implementations, step S102 above, which involves performing intent recognition based on multimodal data to determine that the target object has the intent to operate the tailgate, may include the following steps S201-S203.

[0043] S201, perform hand occupancy recognition based on the visual image sequence to obtain the hand load sequence of the target object.

[0044] It should be noted that the hand load sequence may include, but is not limited to, hand load at multiple moments. The hand load represents the actual load-bearing capacity of the target object's hand when it is carrying an object. A higher hand load indicates a higher actual load-bearing capacity of the hand, and vice versa.

[0045] In this step, pose and hand / object occupancy recognition (Pose & Hand / Load Estimation) are performed based on the visual images at each time point to obtain the hand load degree of the target object at each time point, and these are integrated into a hand load degree sequence according to the time sequence.

[0046] S202, extract pose features from the visual image sequence to obtain the pose feature sequence of the target object.

[0047] It should be noted that the posture feature sequence can include posture features at multiple time points. Posture features may include, but are not limited to, approach speed, approach angle, and the rate of change of hand posture relative to the tailgate. Here, approach speed refers to the speed at which the target object approaches the tailgate; approach angle refers to the orientation deviation angle of the target object relative to the tailgate; and the rate of change of hand posture relative to the tailgate represents the change in hand posture relative to the tailgate between two adjacent time points. The hand posture relative to the tailgate can be measured by the approach angle.

[0048] In this step, people detection and tracking (Detect & Track) are performed based on visual image sequences to determine the position and velocity vector of the target object at each time step. The velocity vector is the approach velocity.

[0049] Define any given moment as the target moment. Based on the position of the target object and the tailgate of the target vehicle at the target moment, determine the hand orientation vector and tailgate reference vector at the target moment. The hand orientation vector represents the vector pointing from the wrist to the tip of the middle finger, and the tailgate reference vector represents the normal vector of the tailgate doorway plane. Project the hand orientation vector at the target moment onto the tailgate coordinate system, and calculate the spatial angle between the tailgate reference vector and the projected hand orientation vector at the target moment as the orientation deviation angle, i.e., the approach angle.

[0050] The backward difference method is further used to calculate the rate of change of the hand's attitude relative to the tailgate at the target moment. Specifically, the difference between the approach angle at the target moment and the approach angle at the previous moment is calculated as the angle difference, the difference between the target moment and the previous moment is calculated as the time difference, and the ratio of the angle difference to the time difference is calculated as the rate of change of the hand's attitude relative to the tailgate at the target moment. The approach angle at the previous moment is initialized to zero.

[0051] S203, based on the posture feature sequence, hand load sequence and vehicle operation parameter sequence, perform intent recognition to determine that the target object has the intent to operate the tailgate.

[0052] In this step, for each time point, the posture features, hand load, and vehicle operation parameters are aligned and concatenated into multi-dimensional features. This yields multi-dimensional features at multiple time points, forming a multi-dimensional feature sequence. This multi-dimensional feature sequence is then input into a pre-trained intent recognition model, which outputs the probability that the target object has the intent to operate the tailgate at the current time. If the probability is greater than a preset probability threshold (e.g., 0.8), the target object is determined to have the intent to operate the tailgate; otherwise, it is determined that the target object does not have the intent to operate the tailgate.

[0053] It should be understood that the intent recognition model can be trained using multiple pre-set training samples and the corresponding label information for each training sample. The training samples can include three types of data: posture features, hand load, and vehicle operating parameters. The label information indicates whether the target object has the intent to operate the tailgate; a first value (e.g., 1) indicates that the target object has the intent to operate the tailgate, while a second value (e.g., 0) indicates that the target object does not have the intent to operate the tailgate.

[0054] Furthermore, the type of intent recognition model is not specifically limited. It can be a machine learning model such as a random forest or support vector machine, or a deep learning model such as a long short-term memory (LSTM) network, but is not limited to these. In one example, the intent recognition model may include a sequentially connected long short-term memory network and a fully connected layer. A softmax function is placed after the fully connected layer. The long short-term memory network processes the input sequence, outputting the hidden state at the last time step. This hidden state is then processed by the fully connected layer and the softmax function to obtain the probability that the target object has the intent to perform a tailgate operation at the current time step.

[0055] Therefore, this embodiment acquires a visual image sequence of the target object, performs hand occupancy recognition to obtain a hand load sequence, and extracts posture features to obtain a posture feature sequence. Combined with vehicle operation parameter sequences, a multi-dimensional feature sequence is constructed. Furthermore, a pre-trained intent recognition model is used to accurately determine whether the target object intends to operate the tailgate. In this process, the hand load sequence reflects the actual load-bearing capacity of the target object's hands in real time, while the posture feature sequence covers key dynamic information such as approach speed, approach angle, and the rate of change of hand posture relative to the tailgate. The introduction of vehicle operation parameters further enriches the contextual awareness dimension. This allows the multi-dimensional features obtained by fusing the above features to comprehensively characterize the target object's behavioral patterns and tailgate opening tendency, accurately distinguishing between high-load intent tailgate opening operations and casual passing operations, thereby effectively improving the accuracy of intent recognition.

[0056] In some embodiments, step S201 above, which involves performing hand occupancy recognition based on the visual image sequence to obtain the hand load degree sequence of the target object, may include: Hand occupancy is identified based on visual images at each time point, and parameters such as hand occupancy duration, hand occupancy area, and arm extension restriction are obtained at each time point. The hand load at each time point is determined based on the duration of hand occupation, the area occupied by the hand, and the parameters of limited arm extension.

[0057] In this embodiment, after obtaining the visual image sequence, the timer is initialized by setting the current count to zero and the total count to the number of moments. For each moment, we have: Human keypoint estimation is performed on the visual image to obtain the shoulder, elbow, and wrist keypoints, as well as multiple hand keypoints (such as thumb and index finger keypoints). Based on the lateral maxima, lateral minima, lateral maxima, and lateral minima of the coordinates of these hand keypoints, a hand coordinate range is set. Human recognition is then performed on the visual image to obtain the target object's bounding box. Furthermore, object recognition is performed on the visual image to obtain the bounding boxes for each object, retaining only the bounding boxes for objects whose coordinates fall within the hand coordinate range.

[0058] When a detection box contains an object whose coordinates fall within the hand's coordinate range, the current count is incremented by one; otherwise, the current count is not modified. The ratio of the current count to the total count is determined as the hand's occupancy time, which indicates the duration for which the target object's hand carries the object.

[0059] The sum of the areas of the detection boxes of each object is calculated as the object area, and the area of ​​the detection box of the target object is calculated as the object area. The ratio of the object area to the object area is determined as the area occupied by the hand, which is used to indicate the total area occupied by the target object's hand by the objects.

[0060] Using the coordinates of the shoulder, elbow, and wrist key points, a first vector pointing from the elbow to the shoulder and a second vector pointing from the elbow to the wrist are determined. The spatial angle between the first and second vectors is calculated as the joint angle. The ratio of the joint angle to the preset maximum joint range of motion is determined as the arm extension limitation parameter, which is used to indicate the degree of limitation of the target object's hand when carrying an object. The larger the value, the lower the degree of limitation, and vice versa.

[0061] After obtaining parameters such as hand occupancy time, hand occupancy area, and arm extension restriction, these parameters are fused to obtain the hand load intensity. The specific method of fusion is not limited. For example, machine learning methods can be used to process these parameters to obtain the hand load intensity. This machine learning method learns the mapping relationship between data such as hand occupancy time, hand occupancy area, and arm extension restriction parameters and the hand load intensity. The type of machine learning method can be flexibly set according to the actual situation; for example, it can be a support vector machine, random forest, etc., but is not limited to these. Another example is to perform a weighted sum of the hand occupancy time, hand occupancy area, and arm extension restriction parameters to obtain the hand load intensity. The weight values ​​of each data point can be set according to the actual situation, and the sum of the weight values ​​of each data point is one, but is not limited to this.

[0062] Therefore, this implementation first utilizes human keypoint estimation and object detection algorithms to obtain the hand coordinate range, human detection box, and object detection box, respectively. Then, based on these, it calculates the hand occupancy duration, hand occupancy area, and arm extension restriction parameters. The hand occupancy duration, by statistically analyzing the proportion of consecutive frames carrying the object, characterizes the continuity of the load-bearing task. The hand occupancy area, using the ratio of the object's area to the human detection box area, indirectly reflects the physical extent to which the object's volume occupies the hand. The arm extension restriction parameter, through shoulder-elbow-wrist joint angle normalization, quantifies the degrees of freedom of posture when carrying the object. These three parameters characterize the load state from three aspects: temporal persistence, spatial occupancy, and physiological comfort, and are ultimately fused into a hand load degree. Thus, through multi-dimensional features and their fusion mechanism, the accuracy of the hand load degree can be effectively improved. This helps in subsequent intent recognition to accurately distinguish between high-load intents such as opening the tailgate and accidental passing through, improving the accuracy of intent recognition.

[0063] In some implementations, step S103 above, which calculates parameters based on multimodal data to obtain the tailgate control parameters of the target vehicle, may include the following steps S301-S302: S301, Determine the relative height of the target object based on the visual image sequence; S302, the tailgate opening angle is obtained by calculating parameters based on relative height.

[0064] In this embodiment, firstly, the last frame (i.e., the current moment) of the visual image sequence is extracted and human body recognition is performed on it to obtain the detection box of the target object at the current moment. The absolute value of the difference between the vertical minimum and the vertical maximum of the detection box is calculated as the pixel height of the target object. The pixel height is transformed to the world coordinate system using a preset transformation matrix to obtain the first height of the target object. Simultaneously, the second height of the target object is measured directly through camera calibration, ground plane assumption, or depth camera. All of the above processes can be implemented using existing technologies. The first height and the second height are fused (e.g., by weighted averaging, average, etc.) to obtain the relative height of the target object. In this way, by measuring the height of the target object through both visual perception and camera measurement, the error of the target object's height can be kept within an acceptable range (e.g., ±5cm), thereby effectively improving the accuracy of the target object's height measurement.

[0065] Then, based on the relative height, the tailgate opening angle of the target vehicle is calculated. The range of the tailgate opening angle can be determined by the vehicle's mechanical limits, for example, its value range can be [40°, 60°], but it is not limited to this.

[0066] In some examples, the tailgate opening angle is determined solely based on relative height to improve decision-making efficiency and reduce computational load. For instance, a first mapping table is invoked, storing multiple preset height values ​​and their corresponding opening angles. The table is then looked up based on the relative height to select the opening angle corresponding to that height as the tailgate opening angle. Alternatively, a machine learning method can be used to process the relative height to obtain the tailgate opening angle. This method learns the mapping relationship between relative height and tailgate opening angle, and its type can be flexibly set according to the actual situation; for example, it could be a support vector machine, random forest, etc., but is not limited to these.

[0067] In other examples, when the target object provides a corresponding authorization code, the target object's preference data is further introduced, such as pre-set preferred tailgate opening angles, to make the tailgate opening angle more consistent with user preferences, thereby improving the decision accuracy of the tailgate opening angle. For example, a second mapping table is called, which stores multiple preset composite samples (including preference data and relative height) and the opening angle corresponding to each composite sample. The mapping table is looked up using relative height and preference data, and the opening angle corresponding to the preference data and relative height is selected as the tailgate opening angle. Another example is using machine learning methods to process the preference data and relative height to obtain the tailgate opening angle. This machine learning method learns the mapping relationship between data such as preference data and relative height and the tailgate opening angle. Its type can be flexibly set according to the actual situation; for example, it can be a support vector machine, random forest, etc., but is not limited to these.

[0068] Therefore, this implementation introduces a height-sensing-based adaptive tailgate opening control mechanism. This mechanism determines the height of the target object through a sequence of visual images and adaptively decides the tailgate opening angle accordingly. This achieves adaptive matching between the tailgate opening angle and the target object's height, effectively improving the accuracy of the tailgate opening angle. This allows the tailgate of the target vehicle to actively open to the angle that best matches the target object's posture in subsequent control, meeting practical needs such as closing the tailgate. This effectively improves the control precision of the vehicle's tailgate and enhances the user's tailgate experience. In some implementations, step S103 above, which calculates parameters based on multimodal data to obtain the tailgate control parameters of the target vehicle, may include the following steps S303-S304: S303, Based on the visual image sequence, determine the distance between the target object and the tailgate as the target distance; S304 calculates the tailgate opening time based on the target distance and tailgate opening angle.

[0069] In this embodiment, firstly, personnel detection and tracking are performed based on visual image sequences to determine the position of the target object at the current moment. Then, combined with the position of the tailgate of the target vehicle, the distance between the target object and the tailgate is calculated as the target distance.

[0070] The distance between the target object and the tailgate can be calculated using a distance function. The type of distance function can be flexibly set according to the actual situation; for example, the distance function can be Euclidean distance function, Manhattan distance function, etc., but it is not limited to these. For example, the Euclidean distance function can be used as the distance function, and the Euclidean distance between the target object and the tailgate can be calculated based on the current position of the target object and the position of the tailgate of the target vehicle. Another example is that both the Euclidean and Manhattan distance functions can be used as distance functions. Based on the current position of the target object and the position of the tailgate of the target vehicle, the Euclidean and Manhattan distances between the target object and the tailgate can be calculated, and the Euclidean and Manhattan distances can be averaged to obtain the target distance.

[0071] Then, based on the target distance, the time required for the target object to walk to the tailgate is calculated as the first time. For example, the first time could be calculated as the ratio of the target distance to the target object's approach speed at the current moment. Alternatively, the average approach speed of the target object at each moment could be calculated first as the average speed of the target object, and then the ratio of the target distance to the average speed of the target object could be calculated as the first time.

[0072] Simultaneously, the time required for the tailgate of the target vehicle to open from zero to the tailgate opening angle under acceleration limitations is obtained as the second time. This second time can be obtained through a third mapping table, which stores multiple preset opening angles and their corresponding opening times. The tailgate opening angle is used to look up the corresponding opening time in this mapping table, and the second time is selected from the selected time.

[0073] Finally, the difference between the first duration and the second duration is calculated as the duration difference. If the duration difference is less than zero, it means the target object is very close to the tailgate of the target vehicle, and the tailgate opening is urgent. In this case, the tailgate opening time is determined as the current time to control the tailgate to open immediately or provide a half-open quick opening action to avoid delay. If the duration difference is greater than or equal to zero, it means there is still a certain distance between the target object and the tailgate of the target vehicle. In this case, the sum of the current time and the difference is calculated as the tailgate opening time to ensure that the tailgate is opened to the correct opening angle when the target object arrives.

[0074] Therefore, this embodiment, by fully considering the position of the target object and the time required for the tailgate to be driven to a specific opening angle, adaptively decides the opening time of the target vehicle's tailgate. This ensures that the action of opening the tailgate to the tailgate opening angle is precisely synchronized with the time when the target object arrives at the tailgate, thereby matching the actual tailgate operation requirements and effectively improving the accuracy of the tailgate opening time.

[0075] In some implementations, step S103 above, which calculates parameters based on multimodal data to obtain the tailgate control parameters of the target vehicle, may further include the following step S305: S305, the tailgate opening speed is obtained by calculating parameters based on the tailgate opening time and tailgate opening angle.

[0076] In this embodiment, the tailgate opening speed of the target vehicle is adaptively determined based on the tailgate opening time and tailgate opening angle to ensure that the tailgate is opened to the correct opening angle when the target arrives.

[0077] In some examples, a fourth mapping table is invoked, which stores multiple preset composite samples (including tailgate opening time and tailgate opening angle) and the opening speed corresponding to each composite sample. The mapping table is looked up using the tailgate opening time and tailgate opening angle, and the opening speed corresponding to the tailgate opening time and tailgate opening angle is selected as the tailgate opening speed.

[0078] In other examples, machine learning methods are used to process the tailgate opening time and tailgate opening angle to obtain the tailgate opening speed. This machine learning method learns the mapping relationship between data such as tailgate opening time and tailgate opening angle and tailgate opening speed. Its type can be flexibly set according to the actual situation. For example, it can be support vector machine, random forest, etc., but is not limited to this.

[0079] Therefore, this embodiment, by fully considering the position of the target object and the time required for the tailgate to be driven to a specific opening angle, adaptively determines the opening speed of the target vehicle's tailgate. This ensures that the action of opening the tailgate to the tailgate opening angle is precisely synchronized with the moment the target object arrives at the tailgate, thereby matching the actual tailgate operation requirements and effectively improving the accuracy of the tailgate opening speed.

[0080] In some embodiments, the above method may further include the following step S105: S105, based on the obstacle perception data of the target vehicle, corrects the tailgate control parameters.

[0081] In this embodiment, during the calculation of the tailgate opening angle, a lookup process is performed on the fifth mapping table based on the currently calculated tailgate opening angle. This mapping table stores multiple preset opening angles and the corresponding safe distance thresholds for each opening angle. The safe distance threshold corresponding to the tailgate opening angle can be obtained through the lookup process. Simultaneously, the position of the obstacle is extracted from the obstacle perception data of the target vehicle, and combined with the tailgate position of the target vehicle, the distance between the obstacle and the tailgate is calculated as the obstacle distance. The calculation method for this distance can follow the target distance calculation method described in the previous embodiment, and will not be repeated here. If the obstacle distance is less than the safe distance threshold corresponding to the tailgate opening angle, it indicates a high risk of the tailgate colliding with the obstacle during opening. In this case, opening will be prohibited, i.e., the tailgate opening angle will be set to zero, or the tailgate opening angle will be limited to a smaller angle. Otherwise, it indicates a low risk of the tailgate colliding with the obstacle during opening, and the tailgate opening angle will not be corrected.

[0082] In some examples, a lookup process is performed on the sixth mapping table based on the obstacle distance. This mapping table stores multiple preset distance values ​​and the corresponding tailgate opening angle correction coefficients, where each coefficient is greater than zero and less than one. The lookup process yields the tailgate opening angle correction coefficient corresponding to the obstacle distance. The product of this correction coefficient and the currently calculated tailgate opening angle is then used to determine the new tailgate opening angle. The process then returns to the step of looking up the preset fifth mapping table based on the currently calculated tailgate opening angle, thus achieving cyclic detection.

[0083] In other examples, a lookup process is performed on the seventh mapping table based on the obstacle distance. This mapping table stores multiple preset distance values ​​and corresponding tailgate opening angle correction values, where the opening angle correction values ​​are less than zero. The lookup process yields the opening angle correction value corresponding to the obstacle distance. The sum of this correction value and the currently calculated tailgate opening angle is used to determine the new tailgate opening angle. The process then returns to the step of looking up the fifth mapping table based on the currently calculated tailgate opening angle, thus achieving cyclic detection.

[0084] Furthermore, considering that the position of obstacles may change, during the tailgate control process, the obstacle's position is extracted from the obstacle perception data of the target vehicle, and combined with the tailgate position of the target vehicle, the distance between the obstacle and the tailgate is calculated as the obstacle distance. Based on the obstacle distance and the tailgate opening speed of the target vehicle, the collision time between the obstacle and the tailgate is calculated. If the collision time is less than a preset time threshold (e.g., 0.5 seconds), and / or an abnormal collision torque is detected (increased motor current), it indicates that the obstacle has collided with the tailgate. In this case, the tailgate opening behavior is immediately stopped, that is, the tailgate opening angle is set to zero, or the tailgate opening angle is set to a preset negative value, causing it to slightly reverse. Otherwise, it indicates that the obstacle has not collided with the tailgate, and the tailgate opening angle is not corrected.

[0085] Therefore, this embodiment introduces an obstacle safety constraint mechanism in the process of calculating the tailgate opening angle or controlling the tailgate. This mechanism can actively reduce the risk of collision between the tailgate and obstacles while meeting the actual tailgate operation needs of the target object as much as possible. Even if a collision occurs, it can brake and buffer in time to protect the vehicle structure to the maximum extent, thereby effectively improving the safety of tailgate opening.

[0086] To facilitate understanding of the vehicle tailgate adaptive control method described above in this application, an example of a practical application scenario of the vehicle tailgate adaptive control method described above in this application is provided below.

[0087] Reference Figure 2 In this application scenario, the target object walks towards the target vehicle with a shopping bag in one hand. The adaptive control process of the tailgate of the target vehicle is as follows: steps S401-S407.

[0088] S401, Idle Detection: The vehicle's camera performs human detection in the area behind the target vehicle's rear, while simultaneously checking if the vehicle's key is within a preset range (e.g., 20 meters). If a human figure is detected in the area behind the target vehicle's rear, and the key is within the preset range, it is determined that the target is close to the vehicle, and the process proceeds to step S402. Otherwise, it is determined that the target is not close to the vehicle, and the process returns to the previous steps of performing human detection in the area behind the target vehicle's rear and checking if the key is within the preset range (e.g., 20 meters), thus achieving cyclical detection.

[0089] S402, Personnel Detection and Tracking: After detecting the target object, a target tracking (Multi-Object Tracker) is established to record its position, approach speed and other information within a preset time period (multiple moments), and to acquire its visual images within the preset time period.

[0090] S403, Pose & Hand / Load Estimation: For each moment, the process involves: first, estimating key points of the target object based on the visual image, acquiring key points such as shoulder key points, elbow key points, wrist key points, and multiple hand key points (e.g., thumb key points, index finger key points, etc.); then, recognizing hand occupancy based on the key points and the target object's bounding box, obtaining parameters such as hand occupancy duration, hand occupancy area, and arm extension restriction; and finally, weighted summation of these three parameters to obtain the hand load degree. Simultaneously, based on the visual image, data such as the target object's approach speed, approach angle, and the rate of change of hand posture relative to the tailgate are extracted as posture features, and the key's state signal is collected as vehicle operation parameters.

[0091] S404, Intention Recognition: For each time step, the posture features, hand load, and vehicle operation parameters are aligned and concatenated into multi-dimensional features. This yields multi-dimensional features for multiple time steps, forming a multi-dimensional feature sequence. This multi-dimensional feature sequence is input into a pre-trained intent recognition model, which outputs the probability that the target object has the intent to operate the tailgate at the current time. If the probability is greater than a preset probability threshold, the target object is determined to have the intent to operate the tailgate, and the process proceeds to step S405; otherwise, the target object is determined not to have the intent to operate the tailgate, and the process returns to step S401, thus achieving cyclic detection.

[0092] S405, Height & ID: The relative height of the target object is estimated based on camera geometric calibration and human detection bounding boxes (obtained from the visual image at the current moment).

[0093] S406, Control parameters calculation: (1) Tailgate opening angle: If the authorization code of the target object is determined, the preference data of the target object is obtained. The second mapping table is called, which stores multiple preset composite samples (including preference data and relative height) and the opening angle corresponding to each composite sample. The mapping table is looked up by relative height and preference data, and the opening angle corresponding to preference data and relative height is selected as the tailgate opening angle.

[0094] In calculating the tailgate opening angle, a lookup process is performed on a fifth mapping table based on the currently calculated tailgate opening angle. This mapping table stores multiple preset opening angles and their corresponding safe distance thresholds. The safe distance threshold corresponding to the tailgate opening angle can be obtained through the lookup process. Simultaneously, the position of obstacles is extracted from the obstacle perception data of the target vehicle, and combined with the tailgate position of the target vehicle, the distance between the obstacle and the tailgate is calculated as the obstacle distance. If this obstacle distance is less than the safe distance threshold corresponding to the tailgate opening angle, the tailgate opening angle is reduced; otherwise, the tailgate opening angle is not corrected. (2) Tailgate opening time: First, personnel detection and tracking are performed based on the visual image sequence to determine the position of the target object at the current time. Combined with the position of the tailgate of the target vehicle, the distance between the target object and the tailgate is calculated as the target distance. Next, based on the target distance, the time required for the target object to walk to the tailgate is calculated as the first time. At the same time, the time required for the tailgate of the target vehicle to open from zero to the tailgate opening angle under acceleration limitation is obtained as the second time. After that, the difference between the first time and the second time is calculated as the time difference. If the time difference is less than zero, the tailgate opening time is determined as the current time; otherwise, the sum of the current time and the difference is calculated as the tailgate opening time, that is, the tailgate opening time is t_now + (t_user_arrival - t_motor_needed), where t_now is the current time, t_user_arrival is the first time, and t_motor_needed is the second time.

[0095] (3) Tailgate opening speed: Call the fourth mapping table, which stores multiple preset composite samples (including tailgate opening time and tailgate opening angle) and the opening speed corresponding to each composite sample. The mapping table is looked up by the tailgate opening time and tailgate opening angle, and the opening speed corresponding to the tailgate opening time and tailgate opening angle is selected as the tailgate opening speed.

[0096] In some examples, the tailgate opening angle and tailgate opening speed can be adjusted based on the task type (loading large items, placing shopping bags, retrieving child seats).

[0097] S407, Execute & Monitor: An opening command is sent to the Electronic Control Unit (ECU) of the tailgate drive, including the tailgate opening time, tailgate opening speed, and tailgate opening angle. During the tailgate control process, data returned by the camera and ultrasonic radar are continuously monitored. Based on the obstacle distance and the tailgate opening speed of the target vehicle, the collision time between the obstacle and the tailgate is calculated. If the collision time is less than a preset time threshold, and / or an abnormal collision torque is detected, the tailgate opening angle is set to a preset negative value to slightly reverse it; otherwise, the tailgate opening angle is not corrected. If the target object is detected to have suddenly left, or the tailgate operation intention is canceled, the tailgate can be maintained or reversed according to the configuration. After opening is completed, the tailgate will automatically close or remain open based on the target object's action (such as placing items or leaving) and timeout.

[0098] In some examples, the data acquisition module may include: 1) A camera module: a rear-facing exterior camera (wide-angle / fisheye) covering the pedestrian area behind the vehicle, with pixels and frame rate sufficient for real-time detection (at least 720p@25fps recommended), optionally paired with near-infrared to enhance low-light performance; optional depth / binocular or structured light to directly acquire depth information. 2) Near-field obstacle sensors: multi-point ultrasonic sensors or short-range millimeter-wave radars are positioned on the tailgate edge, bumper, etc., to measure the distance between the tailgate and obstacles, and to compensate for near-field blind spots behind the vehicle. The camera calibration parameters and vehicle coordinate system must be calibrated at the factory or during maintenance, and the depth camera / binocular system must be timestamped.

[0099] In some examples, the unit used to perform calculations and control may be: an on-board domain control unit equipped with GPU / AI acceleration, running multi-threaded perception algorithms and decision logic, and outputting Controller Area Network (CAN) messages to the tailgate drive ECU or directly controlling the drive.

[0100] In some examples, the tailgate drive actuator is controlled by the tailgate drive ECU, which can be a variable speed motor with a position / angle encoder and torque / current monitoring; it supports software-level stop / reverse and mechanical limit protection, meeting ISO safety requirements.

[0101] In some examples, the above process is preferably handled locally on the vehicle side. If remote uploading is required, only the de-identified feature vectors are uploaded with user authorization.

[0102] In some examples, in this application scenario, a target person approaches a target vehicle carrying a shopping bag. The target vehicle detects this intention to open the tailgate, adaptively decides the tailgate opening time, speed, and angle, and controls the tailgate opening accordingly. The entire process takes 1.2 seconds, ensuring the tailgate is open to 55° when the target person arrives (the target person is 1.7m tall). Furthermore, during the tailgate opening process, if an obstacle is detected at a distance of 0.25 meters behind, which is less than a safety threshold, the vehicle immediately retracts 5° and remains hovering.

[0103] In addition, refer to Figure 3 This application also provides a vehicle tailgate adaptive control device, which may include: Acquisition module 501 is used to acquire multimodal data of the target object; The first processing module 502 is used to perform intent recognition based on multimodal data to determine that the target object has the intent to operate the tailgate; wherein, the tailgate operation intent means the intention of the target object to operate the tailgate of the target vehicle; The second processing module 503 is used to perform parameter calculations based on multimodal data to obtain the tailgate control parameters of the target vehicle. The third processing module 504 is used to control the tailgate of the target vehicle according to the tailgate control parameters.

[0104] The content of the above method embodiments is applicable to the device embodiments. The specific functions implemented by the device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0105] Moreover, refer to Figure 4 This application also provides a vehicle, which may include: At least one processor 601; At least one memory 602 is used to store at least one program; When at least one program is executed by at least one processor 601, the at least one processor 601 implements the above-described adaptive control method for a vehicle tailgate.

[0106] The aforementioned vehicles can be private cars, such as sedans, sport utility vehicles (SUVs), multi-purpose vehicles (MPVs), or pickup trucks, or commercial vehicles, such as vans, buses, small trucks, or large trailers, or gasoline vehicles or new energy vehicles such as hybrid or pure electric vehicles.

[0107] The aforementioned memory 602, as a non-transitory network system, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory 602 may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory 602 may optionally include memory 602 remotely located relative to processor 601, and these remote memories 602 can be connected to processor 601 via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0108] The aforementioned memory 602 can be implemented as a read-only memory (ROM), static storage device, dynamic storage device, or random access memory (RAM). Memory 602 can store the operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in memory 602 and called by processor 601 to execute the methods of the embodiments of this application.

[0109] The processor 601 described above can be implemented using a general-purpose central processing unit (CPU), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this application.

[0110] In some embodiments, the vehicle may further include: Input / output interfaces are used to implement information input and output; The communication interface is used to enable communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.). The bus transmits information between various components of the device (such as processor 601, memory 602, input / output interfaces, and communication interfaces); The processor 601, memory 602, input / output interface, and communication interface can communicate with each other within the device via a bus.

[0111] The content of the above method embodiments is applicable to this vehicle embodiment. The specific functions implemented in this vehicle embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0112] Finally, this application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described adaptive control method for a vehicle tailgate.

[0113] The content of the above method embodiments is applicable to the embodiments of this program product. The specific functions implemented by the embodiments of this program product are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0114] The embodiments described in this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided by the embodiments of this application. As those skilled in the art will know, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of this application are also applicable to similar technical problems.

[0115] Those skilled in the art will understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of this application, and may include more or fewer steps than shown, or combine certain steps, or different steps.

[0116] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0117] Those skilled in the art will understand that all or some of the steps in the methods disclosed above, as well as the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, or suitable combinations thereof.

[0118] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0119] It should be understood that in this application, "at least one (item)" means one or more, and "more than" means two or more. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (item) of a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0120] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0121] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0122] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0123] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes multiple instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing programs, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0124] The preferred embodiments of the present application have been described above with reference to the accompanying drawings, but this does not limit the scope of the claims of the present application. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and substance of the embodiments of the present application shall be within the scope of the claims of the present application.

Claims

1. An adaptive control method for a vehicle tailgate, characterized in that, The method includes: Obtain multimodal data of the target object; Intent recognition is performed based on the multimodal data to determine that the target object has the intent to operate the tailgate; wherein, the tailgate operation intent means the target object's intent to open the tailgate of the target vehicle; Based on the multimodal data, parameter calculations are performed to obtain the tailgate control parameters of the target vehicle; The tailgate of the target vehicle is controlled according to the tailgate control parameters.

2. The method according to claim 1, characterized in that, The multimodal data includes vehicle operating parameter sequences and visual image sequences; The step of determining that the target object has the intent to operate the tailgate based on the multimodal data includes: Hand occupancy recognition is performed based on the visual image sequence to obtain the hand load degree sequence of the target object; Pose feature extraction is performed on the visual image sequence to obtain the pose feature sequence of the target object; Intent recognition is performed based on the posture feature sequence, the hand load sequence, and the vehicle operation parameter sequence to determine that the target object has the intention to operate the tailgate.

3. The method according to claim 2, characterized in that, The visual image sequence includes visual images at multiple time points, and the hand load sequence includes hand load values ​​at multiple time points; the step of performing hand occupancy recognition based on the visual image sequence to obtain the hand load sequence of the target object includes: Based on the visual images at each time point, hand occupancy is identified to obtain the hand occupancy duration, hand occupancy area, and arm extension restriction parameters at each time point. The hand load at each specified moment is determined based on the hand occupancy duration, hand occupancy area, and arm extension restriction parameters.

4. The method according to claim 1, characterized in that, The multimodal data includes a sequence of visual images; the tailgate control parameters include the tailgate opening angle; the step of calculating the tailgate control parameters of the target vehicle based on the multimodal data includes: The relative height of the target object is determined based on the visual image sequence; The tailgate opening angle is obtained by calculating parameters based on the relative height.

5. The method according to claim 1, characterized in that, The multimodal data includes a sequence of visual images; the tailgate control parameters include the tailgate opening time and tailgate opening angle; the calculation of parameters based on the multimodal data to obtain the tailgate control parameters of the target vehicle includes: Based on the visual image sequence, the distance between the target object and the tailgate is determined as the target distance; The tailgate opening time is obtained by calculating parameters based on the target distance and the tailgate opening angle.

6. The method according to claim 1, characterized in that, The tailgate control parameters also include tailgate opening time, tailgate opening speed, and tailgate opening angle; the step of calculating the tailgate control parameters of the target vehicle based on the multimodal data further includes: The tailgate opening speed is obtained by calculating parameters based on the tailgate opening time and the tailgate opening angle.

7. The method according to claim 1, characterized in that, The method further includes: The tailgate control parameters are corrected based on the obstacle perception data of the target vehicle.

8. A vehicle tailgate adaptive control device, characterized in that, The device includes: The acquisition module is used to acquire multimodal data of the target object; The first processing module is used to perform intent recognition based on the multimodal data to determine that the target object has the intent to operate the tailgate; wherein, the tailgate operation intent means the intention of the target object to operate the tailgate of the target vehicle; The second processing module is used to perform parameter calculations based on the multimodal data to obtain the tailgate control parameters of the target vehicle. The third processing module is used to control the tailgate of the target vehicle according to the tailgate control parameters.

9. A vehicle, characterized in that, The vehicles include: At least one processor; At least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the vehicle tailgate adaptive control method as described in any one of claims 1-7.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the vehicle tailgate adaptive control method as described in any one of claims 1-7.

Images