Tail gate control method and device and vehicle

Abstract

The invention discloses a tail gate control method and device and a vehicle, is applied to a control system and relates to the technical field of vehicle control, and the method comprises the steps that ultrasonic obstacle data corresponding to a plurality of continuous time points are collected; according to the distance between the target obstacle data and the target ultrasonic radar, generating a motion sequence formed by the distances corresponding to the multiple time points; a target time point is determined, and whether the intention of opening the tail gate exists or not is judged according to the change trend of a forward motion sequence located before the target time point and a backward motion sequence located after the target time point in the motion sequences; if yes, based on continuous frames of vehicle surrounding images collected in the pre-starting stage, target detection is carried out on the vehicle surrounding images so as to extract motion track features used for opening the tail door; and under the condition that the motion track features are matched with the preset feature library, the tail door is controlled to be opened, the whole vehicle manufacturing cost can be reduced, and the probability of foot kicking motion misrecognition is reduced.

Description

Technical Field

[0001] The present disclosure relates to the field of vehicle control technology, and in particular to a tailgate control method, device, and vehicle. Background Art

[0002] In recent years, with the continuous improvement of automobile intelligence and user experience requirements, the electric tailgate kick-opening function has gradually become a standard feature of mid-to-high-end models. This function realizes contactless tailgate control by detecting the user's foot movements, greatly improving convenience.

[0003] Traditional tailgate control requires the installation of a dedicated kick sensor at the rear of the vehicle to detect the movement. However, the high hardware procurement cost of the kick sensor, coupled with the additional sensor wiring and control system layout, increases vehicle manufacturing costs and increases assembly complexity. Furthermore, the kick sensor's detection performance is susceptible to environmental interference, and random movements by non-user targets, such as pets approaching or pedestrians passing by, can be misinterpreted by the sensor as valid commands, causing the tailgate to open unexpectedly. This not only affects the user experience but can also pose a safety hazard.

[0004] Therefore, when controlling the opening of the tailgate by kicking, how to reduce the deployment cost and the probability of misidentification has become a technical problem that technicians in this field urgently need to solve. Summary of the Invention

[0005] In view of the above problems, the present disclosure provides a tailgate control method that overcomes the above problems or at least partially solves the above problems, which is applied to a control system. The technical solution is as follows:

[0006] The ultrasonic obstacle data corresponding to multiple consecutive time points are collected through the target ultrasonic radar pre-set on the vehicle;

[0007] Determine, based on the ultrasonic obstacle data, target obstacle data closest to the target ultrasonic radar at each time point, and generate, based on the distance between the target obstacle data and the target ultrasonic radar, a motion sequence consisting of distances corresponding to multiple time points;

[0008] determining a target time point, and determining whether there is an intention to open the tailgate based on a change trend of a forward motion sequence located before the target time point and a backward motion sequence located after the target time point in the motion sequence;

[0009] If yes, performing target detection on the vehicle surrounding images based on the continuous frames of vehicle surrounding images collected during the pre-start phase to extract motion trajectory features for opening the tailgate;

[0010] When the motion trajectory feature matches a preset feature library, the tailgate is controlled to open.

[0011] The system reuses the existing ultrasonic radar at the rear of the vehicle to replace the kick sensor for kick motion recognition, eliminating the need for an additional dedicated kick sensor and significantly reducing vehicle manufacturing costs. By generating a motion sequence from obstacle data collected by the ultrasonic radar at consecutive time points and analyzing the changes in forward and backward motion trends, it initially identifies the intention to open the tailgate. Based on this, it extracts and matches motion trajectory features using images of the vehicle's surroundings collected during the pre-start phase. Visual recognition technology is used to further eliminate interference from non-target objects. Through multi-dimensional data fusion, the reliability of kick motion recognition is significantly improved, fundamentally reducing the probability of misidentification.

[0012] Optionally, judging whether there is an intention to open the tailgate based on a change trend of a forward motion sequence located before the target time point and a backward motion sequence located after the target time point in the motion sequence specifically includes:

[0013] determining a forward motion sequence located before the target time point and a backward motion sequence located after the target time point in the motion sequence;

[0014] Determine a distance difference between a distance corresponding to each time point in the forward motion sequence and the backward motion sequence and a distance corresponding to the target time point, and determine whether there is an intention to open the tailgate based on a changing trend of the distance difference.

[0015] Determine the distance difference between each time point in the forward motion sequence and the backward motion sequence and the target time point, and accurately identify the motion pattern that matches the kicking action by quantitatively analyzing the movement direction of the target obstacle data. At the same time, filter out invalid signals of interfering behavior, thereby significantly reducing safety hazards caused by misjudgment while improving the accuracy of tailgate opening intention recognition.

[0016] Optionally, judging whether there is an intention to open the tailgate based on a changing trend of the distance difference may include:

[0017] According to the distance difference, screening out a first time point and a second time point whose corresponding distance difference is greater than a preset threshold from the forward motion sequence and the backward motion sequence;

[0018] determining a first time period between the first time point and the target time point, and a second time period between the target time point and the second time point;

[0019] Determine respectively the changing trends of the distance differences corresponding to adjacent time points in the first time period and the second time period, and determine that there is an intention to open the tailgate if the changing trends meet a preset condition.

[0020] By performing a phased quantitative analysis of the distance difference between the forward and backward motion sequences, it is possible to identify whether there is a clear directional change in the motion sequence, thereby accurately capturing the dynamic characteristics of the kicking action and effectively improving the anti-interference capability of the tailgate opening intention recognition.

[0021] Optionally, when the change trend satisfies a preset condition, determining that there is an intention to open the tailgate specifically includes:

[0022] Filtering out, from the first time period, a first set of adjacent time points consisting of adjacent time points whose changing trend is a downward trend, and filtering out, from the second time period, a second set of adjacent time points consisting of adjacent time points whose changing trend is an upward trend;

[0023] Determine a ratio between the number of the first adjacent time point set and the second adjacent time point set and the total number of adjacent time point sets consisting of all adjacent time points;

[0024] In a case where the ratio satisfies a preset ratio threshold, it is determined that there is an intention to open the tailgate.

[0025] By screening out a set of adjacent time points with a downward trend in the first time period and a set of adjacent time points with an upward trend in the second time period, and calculating their ratio to the total adjacent time points, if the ratio reaches a preset ratio threshold, it means that the target motion follows the motion trajectory of the kicking action at most time points, and it is very likely that the kicking action was intentional by the user. This avoids the randomness of judging a single time point or a local trend, and improves the accuracy of identifying the tailgate control intention.

[0026] Optionally, based on the continuous frames of vehicle surrounding images collected during the pre-start phase, target detection is performed on the vehicle surrounding images to extract motion trajectory features for opening the tailgate, specifically including:

[0027] When the vehicle key is detected entering the detection area, the tailgate is controlled to enter the pre-start phase and the vehicle is powered on;

[0028] Based on the pre-start phase, collecting continuous frames of vehicle surrounding images, and performing target detection on the vehicle surrounding images to determine whether there is a user's movement track during the pre-start phase;

[0029] If so, based on the moving direction of the motion trajectory, if it is determined that the user has an intention to approach the vehicle, motion trajectory features corresponding to each key feature area of the user are extracted.

[0030] By detecting objects in successive frames, the system can dynamically track the user's movements, not only determining whether the user intends to approach the vehicle but also extracting motion trajectory features in key feature areas such as the feet. Combining image recognition with ultrasonic radar motion sequence analysis significantly improves the accuracy of tailgate opening intention recognition and reduces accidental tailgate openings caused by environmental interference or misjudgments. Furthermore, by detecting the vehicle key entering the detection area to trigger the pre-start phase, unauthorized users are restricted from operating. Unauthorized users without the key cannot activate the subsequent image acquisition and motion recognition processes. Even if they might kick the tailgate, they cannot open it, thus enhancing the safety of tailgate control.

[0031] Optionally, when the motion trajectory feature matches a preset feature library, controlling the tailgate to open specifically includes:

[0032] matching the motion trajectory feature with a preset feature library to determine whether the motion trajectory feature matches the specified motion feature required to open the tailgate;

[0033] If so, determining a matching degree between the motion trajectory feature and the specified motion feature based on feature weights corresponding to different specified motion features;

[0034] When the matching degree is greater than a preset matching degree, the vehicle door is controlled to be opened; wherein the preset matching degree is set according to the tailgate control frequency.

[0035] By matching the extracted motion trajectory features with the preset specified motion features, the standard actions that meet the tailgate opening requirements can be strictly screened out, effectively eliminating interference from illegal user imitation or unexpected actions. At the same time, the motion trajectory features are differentiated and evaluated through different feature weights to further improve the recognition accuracy and avoid misjudgment due to similar local features.

[0036] Optionally, ultrasonic obstacle data corresponding to multiple consecutive time points may be collected through a pre-set target ultrasonic radar on the vehicle, specifically including:

[0037] According to the relative position between the motion trajectory and a plurality of ultrasonic radars pre-set on the vehicle, selecting an ultrasonic radar closest to the user from the plurality of ultrasonic radars as a target ultrasonic radar;

[0038] The target ultrasonic radar collects ultrasonic obstacle data corresponding to a plurality of consecutive time points.

[0039] Based on the relative position of the user's motion trajectory and multiple ultrasonic radars pre-set on the vehicle, the nearest radar is selected in real time for data collection, ensuring that the acquired ultrasonic obstacle data most directly reflects the dynamic characteristics of the user's foot movements. The closer the distance, the higher the radar signal strength, and the stronger the positioning accuracy and motion trajectory capture ability of the target obstacle data, thereby reducing misjudgments caused by signal attenuation or mixing of multiple radar data and improving the recognition accuracy of motion sequences.

[0040] Optionally, before collecting ultrasonic obstacle data corresponding to a plurality of consecutive time points using a target ultrasonic radar pre-set on the vehicle, the method further includes:

[0041] generating a data collection area for collecting ultrasonic obstacle data according to the location of the target ultrasonic radar;

[0042] After detecting that the vehicle key enters the detection area and controlling the tailgate to enter the pre-start phase, the method further includes:

[0043] collecting facial information of the user, and determining whether the user is a frequent user based on the facial information and preset tailgate opening record information;

[0044] If so, the data collection area is modified according to the tailgate opening record information, so that the modified data collection area satisfies the user's kicking habit.

[0045] A dedicated data collection area is generated based on the target ultrasonic radar's location, effectively filtering out interference from non-relevant areas within the radar's coverage area. This improves the pertinence and effectiveness of ultrasonic data collection, reduces invalid data processing, and reduces system computational complexity. Secondly, after the vehicle enters the pre-start phase, facial information is collected and combined with tailgate opening records to determine the identity of frequent users. The data collection area is then modified based on the user's historical kicking habits. This allows the control system to adapt to the movement frequencies and preferences of different users, providing them with a more appropriate data collection area. This not only improves the user experience, but also significantly enhances the accuracy of kicking motion detection results.

[0046] A tailgate control device, applied to a control system, comprising:

[0047] An acquisition module is used to collect ultrasonic obstacle data corresponding to multiple consecutive time points through a target ultrasonic radar pre-set on the vehicle;

[0048] a motion sequence generating module, configured to determine, based on the ultrasonic obstacle data, target obstacle data closest to the target ultrasonic radar at each time point, and generate, based on the distance between the target obstacle data and the target ultrasonic radar, a motion sequence consisting of distances corresponding to a plurality of time points;

[0049] an intention determination module, configured to determine a target time point and determine whether there is an intention to open the tailgate based on a change trend of a forward motion sequence located before the target time point and a backward motion sequence located after the target time point in the motion sequence;

[0050] an object detection module for, when there is an intention to open the tailgate, performing object detection on the vehicle surrounding images based on successive frames of vehicle surrounding images acquired during the pre-start phase to extract motion trajectory features for opening the tailgate;

[0051] The control module is used to control the opening of the tailgate when the motion trajectory feature matches a preset feature library.

[0052] A vehicle comprising:

[0053] at least one processor; and,

[0054] a memory communicatively connected to the at least one processor; wherein,

[0055] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to: perform any one of the above-mentioned tailgate control methods.

[0056] By leveraging the aforementioned technical solution, the present disclosure provides a tailgate control method, device, and vehicle. These methods reuse the vehicle's existing ultrasonic radar to replace the kick sensor for kick motion recognition, eliminating the need for an additional dedicated kick sensor and significantly reducing vehicle manufacturing costs. By generating a motion sequence based on obstacle data collected by the ultrasonic radar at consecutive time points and analyzing changes in forward and backward motion trends, the system initially identifies the presence of tailgate opening intent. Based on this, motion trajectory features are extracted and matched using images of the vehicle's surroundings collected during the pre-start phase. Visual recognition technology is used to further eliminate interference from non-target objects. Through multi-dimensional data fusion, the reliability of kick motion recognition is significantly improved, fundamentally reducing the probability of misidentification.

[0057] The above description is only an overview of the technical solution of the present disclosure. In order to more clearly understand the technical means of the present disclosure, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present disclosure more obvious and easy to understand, the specific implementation methods of the present disclosure are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Various other advantages and benefits will become apparent to those skilled in the art upon reading the detailed description of the preferred embodiment below. The accompanying drawings are for illustration purposes only and are not to be considered as limiting the present disclosure. The same reference symbols are used throughout the drawings to represent the same components. In the drawings:

[0059] Figure 1 A schematic flow chart of a tailgate control method provided in an embodiment of the present application is shown;

[0060] Figure 2 A schematic diagram of a tailgate structure provided by an embodiment of the present application is shown;

[0061] Figure 3 A schematic diagram of a motion sequence provided by an embodiment of the present application is shown;

[0062] Figure 4 A schematic diagram of an obstacle provided in an embodiment of the present application is shown;

[0063] Figure 5 A schematic structural diagram of a tailgate control device provided in an embodiment of the present application is shown;

[0064] Figure 6 A structural schematic diagram of a vehicle provided in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0065] Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0066] In recent years, with the continuous improvement of automobile intelligence and user experience requirements, the electric tailgate kick-opening function has gradually become a standard feature of mid-to-high-end models. This function realizes contactless tailgate control by detecting the user's foot movements, greatly improving convenience.

[0067] Traditional tailgate control requires the installation of a dedicated kick sensor at the rear of the vehicle to detect the movement. However, the high hardware procurement cost of the kick sensor, coupled with the additional sensor wiring and control system layout, increases vehicle manufacturing costs and increases assembly complexity. Furthermore, the kick sensor's detection performance is susceptible to environmental interference, and random movements by non-user targets, such as pets approaching or pedestrians passing by, can be misinterpreted by the sensor as valid commands, causing the tailgate to open unexpectedly. This not only affects the user experience but can also pose a safety hazard.

[0068] Therefore, when controlling the opening of the tailgate by kicking, how to reduce the deployment cost and the probability of misidentification has become a technical problem that technicians in this field urgently need to solve.

[0069] Based on the above application scenarios, in order to solve the technical problems of high vehicle manufacturing cost and easy misidentification of kicking actions, this application provides a tailgate control method, such as Figure 1 As shown, Figure 1 : is a schematic flow chart of a tailgate control method provided in an embodiment of the present application. The method can be applied to a control system. The method includes:

[0070] S101: Collect ultrasonic obstacle data corresponding to multiple consecutive time points through a target ultrasonic radar pre-set on the vehicle.

[0071] The control system opens the tailgate by identifying whether there is a kicking action. Traditional tailgate control methods require an additional kick sensor to be installed at the rear of the vehicle. However, the kick sensor not only increases the deployment cost, but also has certain deviations in the recognition of the kicking action. If the kicking signal is generated by an illegal user or a moving object such as an animal or floating garbage, the tailgate will be opened by mistake, reducing the safety of the vehicle. At present, some vehicles personalize the sensitivity of the kick sensor to prevent the misidentification of the kicking action. However, this method often fails to capture the user's subtle movements and requires the user to try the kicking action multiple times before the tailgate can be opened, thereby reducing the opening efficiency of the tailgate. Therefore, the embodiment of the present application reuses the vehicle's existing ultrasonic radar to identify whether there is a kicking action at the rear of the current vehicle through the ultrasonic obstacle data detected by the ultrasonic radar.

[0072] Generally, a vehicle is pre-installed with multiple ultrasonic radars, such as Figure 2 The tailgate structure diagram shown in the figure shows that there are four ultrasonic radars installed at the rear of the vehicle, namely R A ~R D In general, any ultrasonic radar is used as the target ultrasonic radar to detect the kick signal generated by the user. By detecting the ultrasonic obstacle data corresponding to multiple consecutive time points and then analyzing the motion characteristics of the ultrasonic obstacle data, it is possible to identify whether the motion characteristics of the current obstacle are consistent with the kicking action. If the motion characteristics of the kicking action are met, it can be determined that the user may be using the kicking action to open the tailgate.

[0073] It should be noted that when detecting obstacle data, the following preconditions need to be met: the tailgate kick-open is enabled, the gear is in P or N and EPB is valid, the vehicle speed signal is valid and the vehicle is stationary, the smart key is in the valid area (within the range of the tailgate low-frequency antenna), the power mode is non-crank mode (ignition gear), the vehicle is in non-transport mode and low-power mode, and the rear tailgate is stationary and in the fully closed position.

[0074] S102: Determine the target obstacle data closest to the target ultrasonic radar at each time point based on the ultrasonic obstacle data, and generate a motion sequence consisting of distances corresponding to multiple time points based on the distance between the target obstacle data and the target ultrasonic radar.

[0075] When performing kick-to-open detection, ultrasonic obstacle data within a time window is required. Because the target ultrasonic radar may collect multiple ultrasonic obstacle data at each time point due to the detection range when performing obstacle detection. In order to reduce the computational complexity, only one ultrasonic obstacle data with significant significance is selected for reference at each time point. Considering that the closer the obstacle signal is to the ultrasonic radar, the higher the credibility, when screening the target obstacle data for reference, it is necessary to screen out the obstacle closest to the target ultrasonic radar from all the obstacle data collected at each time point. Figure 3 As shown in the figure, the points connected by lines represent the target obstacle data corresponding to each time point, while the other discrete points are irrelevant obstacle data. After filtering out the target obstacle data, it is arranged in chronological order. At the same time, the distance between the target obstacle data and the target ultrasonic radar is calculated based on the coordinates of the target obstacle data and the target ultrasonic radar. In this way, the corresponding distance corresponding to each time point can be used to generate a corresponding motion sequence, which consists of the closest distances to the target ultrasonic radar at multiple time points.

[0076] like Figure 4 The diagram of an obstacle is shown in the figure. Points A, B, and C in the detection area represent the obstacles at three different time points. C The nearest target obstacle data is calculated with the vehicle's front direction as the x-axis and the vehicle's left rear wheel direction as the y-axis. Based on the coordinates of the target obstacle data and the target ultrasonic radar, the distance between the target obstacle data and the target ultrasonic radar can be calculated accordingly. This can be achieved using the following calculation formula:

[0077]

[0078] Among them, d A d B d CRespectively represent the distances between obstacles A, B, C and the target ultrasonic radar, (X A ,Y A )、(X B ,Y B )、(X C ,Y C )、(X rc ,Y rc ) represent the coordinates of obstacles A, B, C and target ultrasonic radar respectively.

[0079] The calculation formula of the above distance can be simplified to the distance difference on the x-axis:

[0080] d A =|X A -X rc |

[0081] d B =|X B -X rc |

[0082] d C =|X C -X rc |

[0083] S103: Determine a target time point, and determine whether there is an intention to open the tailgate based on the change trends of the forward motion sequence before the target time point and the backward motion sequence after the target time point in the motion sequence.

[0084] When a user attempts to open the tailgate with a kick while the vehicle is stationary, the typical motion sequence is as follows: the user lifts their foot toward the rear of the vehicle, gradually approaching the area below the bumper; the foot briefly remains within the radar's detection range; and then the foot naturally retracts, moving away from the vehicle. During this process, the ultrasonic radar continuously collects obstacle distance data, forming a complete "far-near-far" motion sequence.

[0085] like Figure 2 As shown, for this motion sequence, the target time point refers to the time point T1 when the obstacle closest to the vehicle's rear ultrasonic radar appears in the continuously collected ultrasonic obstacle data (i.e., the timestamp when the obstacle distance reaches the minimum value). This time point usually corresponds to the moment when the foot is closest to the vehicle during the kicking action and is the key node for the execution of the kicking action. The distance between the ultrasonic obstacle data corresponding to the target time point and the target ultrasonic radar needs to be less than the preset distance threshold S th, the preset distance threshold represents the minimum effective recognition distance of the target ultrasonic radar. When detecting a kicking action, it can only be considered a valid kicking action when the distance is less than this value. Otherwise, even if the motion characteristics of the motion sequence are met, it will be considered an irrelevant random action when the distance is too far. The preset distance threshold can be set according to the actual kicking habits of the user, and this application does not limit this. The forward motion sequence refers to a data sequence of distance change between the obstacle and the ultrasonic radar in a continuous time window before the target time point, reflecting the movement process of the foot from the starting position to the vehicle. The backward motion sequence refers to a data sequence of distance change between the obstacle and the ultrasonic radar in a continuous time window after the target time point, reflecting the movement process of the foot away from the vehicle after completing the approach action.

[0086] Therefore, after generating the motion sequence through the above steps, it is necessary to filter out the target time point with the smallest distance based on the distance corresponding to each time point. With the target time point as the center, the motion sequence is intercepted forward and backward respectively to form a forward motion sequence and a backward motion sequence. Based on the changing trends of the forward and backward motion sequences, it is possible to determine whether there is an intention to open the tailgate. In other words, by determining whether the forward and backward motion sequences follow the "far-near-far" changing trend, it is possible to determine whether there is a valid kicking action at the current vehicle's tailgate. For example, if a pet passes quickly under the car, the ultrasonic radar will detect a sudden drop in distance but no subsequent trend of moving away. Such a motion sequence will be judged as noise and the response will be rejected. By observing the changing trends of the motion sequence, it is possible to eliminate interference caused by motion signals generated by moving obstacles other than humans, thereby improving the recognition accuracy of the kicking action.

[0087] S104: If yes, perform target detection on the vehicle surrounding images based on the continuous frames of vehicle surrounding images collected in the pre-start phase to extract motion trajectory features for opening the tailgate.

[0088] After the ultrasonic radar initially determines that there is a kicking intention at the tailgate, the ultrasonic radar alone cannot detect the identity of the person performing the kicking action, and cannot verify whether the person is the car owner or an illegal user. In some cases, some moving objects may even form a motion sequence with motion characteristics that match the kicking action. In such cases, the effectiveness of the kicking action still needs to be verified through visual information.

[0089] Visual information refers to images of the vehicle's surroundings captured by the vehicle's cameras. These images are collected during the pre-start phase, which occurs when the vehicle detects the user entering a preset range via the key fob and automatically activates the standby state of some sensors and control modules. During the pre-start phase, even if the user does not approach the tailgate, the control system will control the vehicle's cameras to capture multiple, continuous images of the vehicle's surroundings at a fixed frequency for dynamic analysis of user movements. The vehicle cameras here generally refer to surround-view cameras, which are capable of capturing all-round image information of the vehicle's surroundings. If a vehicle is not equipped with a surround-view camera, target detection can be performed on the images of the vehicle's surroundings obtained by stitching together multiple cameras.

[0090] After capturing images of the vehicle's surroundings, computer vision algorithms are used to identify the location and category of specific targets within the images. First, using object detection models such as YOLOv5s, the human body and key feature areas within it are detected, outputting bounding boxes and confidence scores. Detection results with confidence scores below a threshold are filtered to eliminate noise interference. Next, the detected objects between consecutive frames are associated, and the SORT algorithm is used to match the detection boxes of adjacent frames. Using the detection boxes in adjacent frame images, the location information from multiple images of the vehicle's surroundings is captured to form corresponding motion trajectory features. Motion trajectory features include the user's overall motion trajectory as well as motion features of key feature areas such as the head and legs. Visual information is used to verify the validity of the kicking motion detected by the ultrasonic radar, preventing the tailgate from opening accidentally due to misjudgment by a single sensor.

[0091] S105: When the motion trajectory feature matches the preset feature library, control the tailgate to open.

[0092] The preset feature library is a collection of standardized tailgate opening motion features constructed based on extensive experimental data. These standardized tailgate opening motion features are used to assist in determining whether the tailgate was kicked open. These motion features include kicking, bending over to approach the tailgate, lifting a leg, and lifting cargo. The real-time motion trajectory features are then compared to the feature templates in the feature library for similarity. By determining whether the two match, the ultrasonic radar can be used to determine whether the motion sequence detected is valid. If the visual information indicates that the kick is a legitimate user action, the control system will determine that the user intends to open the tailgate and will open it.

[0093] In one embodiment, when detecting the intention to open the tailgate through a motion sequence, the trend of the forward and backward motion sequences segmented at the target time point is used to determine the presence of the tailgate opening intention. Since the motion sequence is a collection of the distances between the obstacle and the target ultrasonic radar at different time points, the trend of change is essentially determined by the distance trend between different time points. This distance trend refers to the trend of the difference between the distance at each time point and the distance at the target time point.

[0094] Specifically, after determining the distance difference between each time point and the target time point, traverse forward from the target time point to find the first time point that satisfies the distance difference greater than the preset threshold. At the same time, traverse backward from the target time point to find the second time point that satisfies the distance difference greater than the preset threshold. The first time point is the starting point of the approaching action, and the second time point is the end point of the moving away action. Figure 3 As shown, the first time point refers to T2, and the second time point refers to T0. It should be noted that the preset threshold is pre-set. In low-noise environments such as rain and snow, the preset threshold can be adaptively relaxed to avoid missed detections.

[0095] After selecting the first and second time points, a first time period between the first and target time points, i.e., T2-T1, and a second time period between the target and second time points, i.e., T1-T0, are determined. The first and second time periods divide the motion sequence into two sub-motion sequences based on the direction of motion. Based on the changing trend of the distance difference between adjacent time points in the first and second time periods, an intention to open the tailgate can be determined if the changing trend meets a preset condition.

[0096] In one embodiment, the changing trends that need to be satisfied in the first time period and the second time period are different. Since the user's foot tends to approach and move away from the vehicle when performing a kicking action, based on the foot characteristics, the changing trend of the distance difference in the above two time periods should also be consistent with the foot action, that is, it shows a downward trend in the first time period and an upward trend in the second time period. Therefore, it is necessary to filter out a first set of adjacent time points consisting of adjacent time points with a downward trend in the first time period, and to filter out a second set of adjacent time points consisting of adjacent time points with an upward trend in the second time period. The distance difference between each adjacent time point in the first set of adjacent time points gradually decreases, and the foot continues to approach the vehicle, while the distance difference between each adjacent time point in the second set of adjacent time points gradually increases, and the foot moves away from the vehicle.

[0097] When detecting obstacles, the ultrasonic radar may acquire some interference data. Therefore, after selecting the first and second sets of adjacent time points, it is necessary to determine whether the trend of change meets the preset conditions by calculating the ratio of valid adjacent time points. First, the number of sets of the first and second sets of adjacent time points is determined. Then, the ratio between the sum of these numbers and the total number of adjacent time point sets consisting of all adjacent time points is calculated. This ratio refers to the ratio of the number of valid trend point pairs to the total number of adjacent point pairs. If the ratio meets the preset ratio threshold, it is determined that there is an intention to open the tailgate. If the ratio is less than the preset ratio threshold, the ultrasonic obstacle data can be identified as an interference signal, and the control system will refuse to open the tailgate.

[0098] The unit of the motion sequence is centimeters. For example, when the motion sequence is [60, 43, 30, 18, 7, 15, 28, 25, 57], in this motion sequence, the time point when the distance value is 7 is the target time point. Then the forward motion sequence divided according to the target time point is [60, 43, 30, 18, 7], and the backward motion sequence is [7, 15, 28, 25, 57]. Each distance represents the closest distance to the ultrasonic radar at each time point. The difference between the corresponding distance of each time point and the target time point is calculated in turn, and the corresponding distance difference sequences [53, 36, 23, 11, 0] and [0, 8, 21, 18, 50] can be obtained. The preset threshold is set to 50 cm. The first time point selected based on this threshold is when the distance difference is 53, and the second time point is when the distance difference is 50. Within the first time period, the number of first adjacent time point sets whose change trend satisfies a downward trend is 4, and within the second time period, the number of second adjacent time point sets whose change trend satisfies an upward trend is 3. The total number of adjacent time point sets in the two time periods is 8. The calculated ratio is 7 / 8 = 87.5%. Since the preset ratio threshold is set to 80%, this ratio is greater than the preset ratio threshold, indicating that the user has a genuine intention to open the tailgate, and the vehicle needs to be controlled to open the tailgate.

[0099] In one embodiment, when the ultrasonic radar recognizes that there is an intention to open the tailgate, the authenticity and legitimacy of the intention to open the tailgate need to be verified in combination with the vehicle's surrounding images to ensure that the current motion sequence is generated by a real user, rather than being mistakenly identified as an illegal user, a pet, or a moving obstacle.

[0100] When the vehicle's wireless key detection module detects a legitimate key entering the detection area, it triggers the tailgate pre-activation phase. After the pre-activation signal is issued, the vehicle completes circuit initialization and activates related hardware such as the tailgate camera and image processor, providing power and computing power for subsequent image acquisition and analysis. This step ensures that the control system is ready before the user initiates the kick, preventing image data loss caused by temporarily activating the device.

[0101] During the pre-startup phase, the vehicle's camera continuously captures images of the vehicle's surroundings at a fixed frame rate, covering the rear and surrounding areas. These images form a continuous video stream, providing time-series data for subsequent analysis. Object detection is performed on each frame, identifying various objects in the image, such as people, vehicles, ground, and plants. A bounding box, category label, and confidence score are generated for each detected object. For example, if the camera captures an image of a user standing behind the rear of a vehicle, a bounding box is generated around the user, labeled "person," and the confidence score confirms the detection reliability. Trajectories of the detected user objects in consecutive frames are associated to generate a trajectory of the user's movement in the image sequence. Based on the generated trajectory, the user's movement direction and trend are analyzed. If the trajectory shows that the user is continuously moving toward the vehicle, it is determined that the user intends to approach the vehicle, triggering the feature extraction process for subsequent actions.

[0102] After confirming the user's approach intent, the system identifies the coordinates of key feature areas on the human body in each frame. The coordinates of key points in consecutive frames are linked in chronological order to form a motion trajectory reflecting the key feature areas. Dynamic features of the trajectory are extracted, including speed, acceleration, and rate of change of trajectory direction, as well as static features such as maximum height and width. These features constitute the motion trajectory features, which are subsequently used for matching with pre-set templates.

[0103] In one embodiment, after extracting the motion trajectory features, it is necessary to calculate the similarity between the real-time detected motion trajectory features and the features in the feature library. By judging the degree of matching between the features, it is determined whether the obstacle motion sequence detected by the current ultrasonic radar is a valid behavior generated by the kicking action.

[0104] The multiple extracted motion trajectory features are quantized and integrated to form a feature vector corresponding to each specified motion feature. This feature vector contains numerical representations of all motion features related to tailgate opening, such as velocity, acceleration, and rate of change of trajectory direction. This vector is then used for subsequent matching with a preset feature library. The constructed motion trajectory feature vector is then matched against the preset feature library. The preset feature library is a standardized set of tailgate opening motion features built based on extensive experimental data, containing various possible examples of specified motion features for tailgate opening. The matching process involves comparing the motion trajectory feature vector with each sample feature vector in the feature library one by one to identify the most similar sample. During the matching process, similarity between the motion trajectory feature and the feature library can be measured using methods such as Euclidean distance and cosine similarity. If the resulting matching degree exceeds the preset matching degree, it can be determined that the user is performing a kicking motion and intends to open the tailgate, and the vehicle will be controlled to open the tailgate.

[0105] It should be noted that the preset matching degree is set based on the frequency of tailgate control. The control system can determine whether kicking the tailgate is a common behavior based on tailgate opening records. If a vehicle's tailgate is most often opened by kicking, it indicates that users have a certain preference for this method. In this case, the preset matching degree can be appropriately lowered to improve user convenience and shorten the tailgate opening response time. Furthermore, the preset matching degree should be determined based on geographic location to determine whether the vehicle is in a high-frequency tailgate opening scenario, such as a parking lot or supermarket unloading area. If the scenario matches this high-frequency scenario, the preset matching degree should be lowered to increase the sensitivity of the kicking action to trigger tailgate opening. Conversely, in atypical scenarios such as highways, the preset matching degree should be increased.

[0106] For example, when a user approaches a vehicle carrying cargo, the system detects a 35° torso tilt, a hand holding cargo, and a kicking motion. The similarities between each action feature and the feature library are 0.5, 0.8, and 0.3, respectively. The weights are 0.4 for carrying, 0.2 for bending, and 0.4 for kicking, for a total match of 0.48. The preset threshold is 0.7, which prevents the tailgate from opening.

[0107] In one embodiment, since the vehicle is equipped with multiple ultrasonic radars at the rear, and the motion sequence provided in the embodiment of the present application is generated for a kicking motion approaching the vehicle from the front or rear direction, generally, only a single target ultrasonic radar is used to collect obstacle data, which can avoid interference between the signals of different radars in the front or rear direction. Therefore, based on the relative position between the user's motion trajectory collected in real time and the multiple ultrasonic radars pre-set in the vehicle, the ultrasonic radar closest to the user is selected from the multiple ultrasonic radars as the target ultrasonic radar for final obstacle detection. In this way, no matter where the user stands at the rear of the vehicle, the control system can dynamically switch the target ultrasonic radar, ensuring that high-quality ultrasonic data can be obtained in various usage scenarios (such as narrow parking spaces, diagonal approach). If the user's motion trajectory is approaching the vehicle from the rear, the target ultrasonic radar can be selected based on the straight-line distance. If the user approaches the vehicle from the front or side, the target ultrasonic radar needs to be selected based on the user's last position. Generally, selecting the ultrasonic radar in the middle position can effectively reduce the impact of interfering obstacle signals on the recognition results. By dynamically selecting the radar closest to the user, the collected ultrasonic data is ensured to reflect the kicking action most directly and clearly. The close-range radar has higher signal strength and stronger resolution of foot movement, and can capture more subtle changes in movement.

[0108] In one embodiment, to reduce the probability of false positives by ultrasonic radar, a preset data collection area is set when collecting obstacle data through ultrasonic radar detection. This means that only obstacle data within the set range is collected, while obstacle data outside the data collection area is considered interference signals. This effectively avoids interference caused by nearby obstacles. For example, a pedestrian may repeatedly kick near a vehicle, but due to their distance from the vehicle, such obstacle signals will not be picked up by the ultrasonic radar. By setting a data collection area, the data collection range is effectively narrowed while focusing on areas where the user's kicking is likely to occur, improving the effectiveness and accuracy of data collection and reducing unnecessary data interference.

[0109] The data collection area needs to be set according to the target ultrasonic radar to be used, such as Figure 2 As shown, using R C As a target ultrasonic radar, the data collection area generally uses a regular shape, such as a rectangle. C As the midpoint of the rectangular side, a data collection area for collecting ultrasonic obstacle data is generated.

[0110] The data collection area defines the range for ultrasonic obstacle data collection. When the vehicle key is detected entering the detection area, the vehicle automatically powers on and activates the vehicle camera. The camera then collects facial information from the user as they approach the vehicle. After acquiring the facial information, the control system compares it with pre-set tailgate opening records to determine whether the user is a frequent user. The vehicle's storage system stores records of past tailgate openings, including the user's facial information, the time, and location of each tailgate opening. After collecting the facial information, key facial features, such as facial contours and facial proportions, are extracted and their similarity with the facial information in the tailgate opening records is calculated. If the facial comparison is successful and the user has opened the tailgate at a set threshold, the user is considered a frequent user. For example, the system may define a user as a frequent user if they have opened the tailgate more than 15 times in the past month. In order to reduce computing power and improve recognition efficiency, for users who open the tailgate frequently, their corresponding user identification can be generated in the tailgate opening record in advance. In this way, after obtaining the facial information, the opening record information carrying the user identification can be directly extracted to compare with the facial information, which can effectively improve the efficiency of facial comparison.

[0111] If facial comparison identifies the user as a frequent user, the control system will modify the data collection area based on the tailgate opening record to accommodate the user's kicking habits. The control system extracts the specific kicking location data for each frequent user from the tailgate opening record. Statistical analysis of this data is performed to calculate the average and distribution of the user's kicking locations. Based on the results of the kicking habit analysis, the original data collection area is adjusted accordingly. If the user's preferred kicking area is to one side of the original data collection area, the data collection area is shifted toward that side. If the user's kicking height is characteristic, the vertical range of the area is adjusted accordingly. For example, if the user tends to kick at a lower position, the lower limit of the data collection area is lowered. If the user's kicking amplitude is small, the preset threshold for the distance difference when detecting the tailgate opening intention through the motion sequence can be appropriately adjusted. The modified data collection area better covers the user's kicking habits, increasing the probability that the ultrasonic radar will capture valid kicking motion data, thereby improving the accuracy and reliability of tailgate opening.

[0112] It should be noted that the data collection area is adjusted based on the vehicle's actual parking location and current environmental conditions. If the vehicle is parked in a cluttered area, it is likely to collect a large amount of interference signals. In this case, the data collection area can be appropriately reduced to prevent accidental opening of the tailgate. Alternatively, in harsh environmental conditions, such as when the vehicle is parked in snow or mud, this will also have a certain impact on the collection of ultrasonic obstacle data. Users may have difficulty in getting close to the vehicle's tailgate to kick. In this case, the data collection area can be appropriately expanded. This allows the ultrasonic radar to collect effective ultrasonic signals to control the tailgate opening even when the user is slightly away from the vehicle, effectively improving operational convenience.

[0113] In one possible implementation, in addition to detecting the tailgate opening intention by kicking the tailgate perpendicularly to the ultrasonic radar, in some cases, the user may also trigger the tailgate opening by swiping horizontally. Unlike a vertical kick, swiping horizontally detects the changing trend of the obstacle's motion sequence along the y-axis. In this case, one or more ultrasonic radars pre-installed on the vehicle can be used for detection. By comparing the ultrasonic obstacle data from each ultrasonic radar at consecutive time points, it can be determined whether the obstacle tends to approach and then retreat, thereby determining whether the user intends to open the tailgate.

[0114] In summary, reusing the existing ultrasonic radar at the rear of the vehicle to replace the kick sensor for kick motion recognition eliminates the need for an additional dedicated kick sensor, significantly reducing vehicle manufacturing costs. By generating a motion sequence using obstacle data collected by the ultrasonic radar at consecutive time points and analyzing the changes in forward and backward motion trends, the system can initially identify the presence of tailgate opening intentions. Based on this, motion trajectory features are extracted and matched using images of the vehicle's surroundings collected during the pre-start phase. Visual recognition technology is used to further eliminate interference from non-target objects. Through multi-dimensional data fusion, the reliability of kick motion recognition is significantly improved, fundamentally reducing the probability of false positives.

[0115] In addition, if Figure 5 As shown, Figure 5 : is a structural diagram of a tailgate control device provided in an embodiment of the present application, the device comprising:

[0116] The acquisition module 501 is used to collect ultrasonic obstacle data corresponding to multiple consecutive time points through a target ultrasonic radar pre-set by the vehicle;

[0117] The motion sequence generation module 502 is configured to determine the target obstacle data closest to the target ultrasonic radar at each time point based on the ultrasonic obstacle data, and generate a motion sequence consisting of the distances corresponding to multiple time points based on the distances between the target obstacle data and the target ultrasonic radar;

[0118] Intention determination module 503 is used to determine a target time point and determine whether there is an intention to open the tailgate based on the change trends of the forward motion sequence before the target time point and the backward motion sequence after the target time point in the motion sequence;

[0119] An object detection module 504 is configured to perform object detection on the vehicle surrounding image based on the continuous frames of vehicle surrounding image acquired during the pre-start phase when there is an intention to open the tailgate, so as to extract motion trajectory features for opening the tailgate;

[0120] The control module 505 is used to control the opening of the tailgate when the motion trajectory feature matches a preset feature library.

[0121] Optionally, the intention determination module 503 determines a forward motion sequence located before the target time point and a backward motion sequence located after the target time point in the motion sequence;

[0122] Determine the distance corresponding to each time point in the forward motion sequence and the backward motion sequence, and the distance difference between the distance corresponding to the target time point. Based on the changing trend of the distance difference, determine whether there is an intention to open the tailgate.

[0123] Optionally, the intention determination module 503 selects, from the forward motion sequence and the backward motion sequence, a first time point and a second time point corresponding to a distance difference greater than a preset threshold value based on the distance difference;

[0124] determining a first time period between the first time point and the target time point, and a second time period between the target time point and the second time point;

[0125] A change trend of distance differences corresponding to adjacent time points in the first time period and the second time period is determined respectively, and if the change trend satisfies a preset condition, it is determined that there is an intention to open the tailgate.

[0126] Optionally, the intention determination module 503 filters out a first set of adjacent time points consisting of adjacent time points whose change trend is a downward trend from the first time period, and filters out a second set of adjacent time points consisting of adjacent time points whose change trend is an upward trend from the second time period;

[0127] Determine a ratio between the number of the first adjacent time point set and the second adjacent time point set and the total number of adjacent time point sets consisting of all adjacent time points;

[0128] When the ratio satisfies a preset ratio threshold, it is determined that there is an intention to open the tailgate.

[0129] Optionally, the target detection module 504 controls the tailgate to enter a pre-start phase and powers on the vehicle when detecting that the vehicle key enters the detection area;

[0130] During the pre-startup phase, continuous frames of vehicle surrounding images are collected and object detection is performed on the vehicle surrounding images to determine whether there is a user's movement trajectory during the pre-startup phase;

[0131] If so, based on the moving direction of the action trajectory, if it is determined that the user has the intention to approach the vehicle, the action trajectory features corresponding to each key feature area of the user are extracted.

[0132] Optionally, the target detection module 504 matches the motion trajectory features with a preset feature library to determine whether the motion trajectory features match the specified motion features required to open the tailgate;

[0133] If so, determining the matching degree between the motion trajectory feature and the specified motion feature based on the feature weights corresponding to different specified motion features;

[0134] When the matching degree is greater than a preset matching degree, the vehicle door is controlled to open; wherein the preset matching degree is set according to the tailgate control frequency.

[0135] Optionally, the acquisition module 501 selects an ultrasonic radar closest to the user from the multiple ultrasonic radars as a target ultrasonic radar based on the relative position between the motion trajectory and multiple ultrasonic radars pre-set on the vehicle;

[0136] The ultrasonic obstacle data corresponding to multiple consecutive time points are collected through the target ultrasonic radar.

[0137] Optionally, before the device collects ultrasonic obstacle data corresponding to a plurality of consecutive time points through a target ultrasonic radar preset by the vehicle, the device further includes:

[0138] According to the location of the target ultrasonic radar, a data collection area for collecting ultrasonic obstacle data is generated;

[0139] When the vehicle key is detected to enter the detection area, after the tailgate is controlled to enter the pre-start phase, the method further includes:

[0140] Collect the user's facial information and determine whether the user is a frequent user based on the facial information and the preset tailgate opening record information;

[0141] If so, the data collection area is modified according to the tailgate opening record information so that the modified data collection area meets the user's kicking habit.

[0142] Regarding the apparatus in the above embodiment, the specific manner in which each unit performs operations has been described in detail in the embodiment of the method, and will not be elaborated on here.

[0143] Figure 6 It is a structural schematic diagram of a vehicle provided in an embodiment of the present application.

[0144] For example, Figure 6 As shown, the vehicle includes: a memory 601 and a processor 602, wherein the memory 601 stores an executable program code 6011, and the processor 602 is used to call and execute the executable program code 6011 to perform the tailgate control method.

[0145] This embodiment can divide the vehicle into functional modules based on the above-described method example. For example, each functional module can be mapped to a specific function, or two or more functions can be integrated into a single processing module. The integrated module can be implemented in hardware. It should be noted that the module division in this embodiment is illustrative and represents only one logical functional division. In actual implementation, other division methods may be used.

[0146] In the case of dividing each functional module into corresponding functional modules, the vehicle may include:

[0147] An acquisition module is used to collect ultrasonic obstacle data corresponding to multiple consecutive time points through a target ultrasonic radar pre-set on the vehicle;

[0148] A motion sequence generation module is used to determine the target obstacle data closest to the target ultrasonic radar at each time point based on the ultrasonic obstacle data, and generate a motion sequence consisting of the distances corresponding to multiple time points based on the distance between the target obstacle data and the target ultrasonic radar;

[0149] An intention judgment module is used to determine the target time point and judge whether there is an intention to open the tailgate based on the changing trends of the forward motion sequence before the target time point and the backward motion sequence after the target time point in the motion sequence;

[0150] an object detection module for, when there is an intention to open the tailgate, performing object detection on the vehicle surrounding image based on consecutive frames of vehicle surrounding image acquired during the pre-start phase to extract motion trajectory features for opening the tailgate;

[0151] The control module is used to control the opening of the tailgate when the motion trajectory characteristics match the preset feature library.

[0152] It should be noted that all relevant contents of each step involved in the above method embodiment can be referred to the functional description of the corresponding functional module and will not be repeated here.

[0153] The vehicle provided in this embodiment is used to execute the above-mentioned tailgate control method, and thus can achieve the same effect as the above-mentioned implementation method.

[0154] In the case of an integrated unit, the vehicle may include a processing module and a storage module. The processing module may be used to control and manage the vehicle's movements, while the storage module may be used to support the vehicle's execution of program codes and data.

[0155] The processing module may be a processor or controller that implements or executes various exemplary logic blocks, modules, and circuits disclosed herein. The processor may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a digital signal processing (DSP) and a microprocessor, and the storage module may be a memory.

[0156] This embodiment also provides a computer-readable storage medium, which stores computer program code (including but not limited to disk storage, CD-ROM, optical storage, etc.). When the computer program code runs on a computer, the computer executes the above-mentioned related method steps to implement a vehicle remote control method provided in the above embodiment.

[0157] This embodiment also provides a computer program product. When the computer program product is run on a computer, it enables the computer to execute the above-mentioned related steps to implement a vehicle remote control method provided by the above embodiment.

[0158] Among them, the beneficial effects of the above embodiments can refer to the beneficial effects of the corresponding methods provided above, and will not be repeated here.

[0159] Through the description of the above implementation methods, technical personnel in the relevant field can understand that for the convenience and simplicity of description, only the division of the above-mentioned functional modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0160] In the embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of modules or units is only a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0161] In the description of the present disclosure, it should be understood that if the terms "up", "down", "front", "back", "left" and "right" are used to indicate directions or positional relationships, they are based on the directions or positional relationships shown in the accompanying drawings. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the positions or elements referred to must have a specific direction, be constructed and operate in a specific direction. Therefore, they should not be understood as limitations of the present disclosure.

[0162] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. It should also be noted that the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, commodity, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, commodity, or device comprising the element.

[0163] The above are merely examples of the present disclosure and are not intended to limit the present disclosure. Various modifications and variations are possible for those skilled in the art. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present disclosure are intended to be included within the scope of the claims of the present disclosure.

Claims

1. A tailgate control method, characterized in that: Applied to a control system, the method includes: The ultrasonic obstacle data corresponding to multiple consecutive time points are collected through the target ultrasonic radar pre-set on the vehicle; Determine, based on the ultrasonic obstacle data, target obstacle data closest to the target ultrasonic radar at each time point, and generate, based on the distance between the target obstacle data and the target ultrasonic radar, a motion sequence consisting of distances corresponding to multiple time points; determining a target time point, and determining whether there is an intention to open the tailgate based on a change trend of a forward motion sequence located before the target time point and a backward motion sequence located after the target time point in the motion sequence; If yes, performing target detection on the vehicle surrounding images based on the continuous frames of vehicle surrounding images collected during the pre-start phase to extract motion trajectory features for opening the tailgate; When the motion trajectory feature matches a preset feature library, the tailgate is controlled to open.

2. The tailgate control method according to claim 1, characterized in that: Determining whether there is an intention to open the tailgate based on a change trend of a forward motion sequence located before the target time point and a backward motion sequence located after the target time point in the motion sequence specifically includes: determining a forward motion sequence located before the target time point and a backward motion sequence located after the target time point in the motion sequence; Determine a distance difference between a distance corresponding to each time point in the forward motion sequence and the backward motion sequence and a distance corresponding to the target time point, and determine whether there is an intention to open the tailgate based on a changing trend of the distance difference.

3. The tailgate control method according to claim 2, characterized in that: Judging whether there is an intention to open the tailgate based on the changing trend of the distance difference includes: According to the distance difference, screening out a first time point and a second time point whose corresponding distance difference is greater than a preset threshold from the forward motion sequence and the backward motion sequence; determining a first time period between the first time point and the target time point, and a second time period between the target time point and the second time point; Determine respectively the changing trends of the distance differences corresponding to adjacent time points in the first time period and the second time period, and determine that there is an intention to open the tailgate if the changing trends meet a preset condition.

4. The tailgate control method according to claim 3, characterized in that: When the change trend satisfies a preset condition, determining that there is an intention to open the tailgate includes: Filtering out, from the first time period, a first set of adjacent time points consisting of adjacent time points whose changing trend is a downward trend, and filtering out, from the second time period, a second set of adjacent time points consisting of adjacent time points whose changing trend is an upward trend; Determine a ratio between the number of the first adjacent time point set and the second adjacent time point set and the total number of adjacent time point sets consisting of all adjacent time points; In a case where the ratio satisfies a preset ratio threshold, it is determined that there is an intention to open the tailgate.

5. The tailgate control method according to claim 1, characterized in that: Based on the continuous frames of vehicle surrounding images collected during the pre-start phase, target detection is performed on the vehicle surrounding images to extract motion trajectory features for opening the tailgate, specifically including: When the vehicle key is detected entering the detection area, the tailgate is controlled to enter the pre-start phase and the vehicle is powered on; Based on the pre-start phase, collecting continuous frames of vehicle surrounding images, and performing target detection on the vehicle surrounding images to determine whether there is a user's movement track during the pre-start phase; If so, based on the moving direction of the motion trajectory, if it is determined that the user has an intention to approach the vehicle, motion trajectory features corresponding to each key feature area of the user are extracted.

6. The tailgate control method according to claim 5, characterized in that: When the motion trajectory feature matches a preset feature library, controlling the tailgate to open specifically includes: matching the motion trajectory feature with a preset feature library to determine whether the motion trajectory feature matches the specified motion feature required to open the tailgate; If so, determining a matching degree between the motion trajectory feature and the specified motion feature based on feature weights corresponding to different specified motion features; When the matching degree is greater than a preset matching degree, the vehicle door is controlled to be opened; wherein the preset matching degree is set according to the tailgate control frequency.

7. The tailgate control method according to claim 5, characterized in that: The vehicle's pre-set target ultrasonic radar collects ultrasonic obstacle data corresponding to multiple consecutive time points, specifically including: selecting, from the plurality of ultrasonic radars, an ultrasonic radar closest to the user as a target ultrasonic radar based on the relative position between the motion trajectory and a plurality of ultrasonic radars pre-set on the vehicle; The target ultrasonic radar collects ultrasonic obstacle data corresponding to a plurality of consecutive time points.

8. The tailgate control method according to claim 5, characterized in that: Before collecting ultrasonic obstacle data corresponding to a plurality of consecutive time points using a target ultrasonic radar pre-set on the vehicle, the method further includes: generating a data collection area for collecting ultrasonic obstacle data according to the location of the target ultrasonic radar; After detecting that the vehicle key enters the detection area and controlling the tailgate to enter the pre-start phase, the method further includes: collecting facial information of the user, and determining whether the user is a frequent user based on the facial information and preset tailgate opening record information; If so, the data collection area is modified according to the tailgate opening record information, so that the modified data collection area satisfies the user's kicking habit.

9. A tailgate control device, characterized in that: Applied to a control system, the device comprises: An acquisition module is used to collect ultrasonic obstacle data corresponding to multiple consecutive time points through a target ultrasonic radar pre-set on the vehicle; a motion sequence generating module, configured to determine, based on the ultrasonic obstacle data, target obstacle data closest to the target ultrasonic radar at each time point, and generate, based on the distance between the target obstacle data and the target ultrasonic radar, a motion sequence consisting of distances corresponding to a plurality of time points; an intention determination module, configured to determine a target time point and determine whether there is an intention to open the tailgate based on a change trend of a forward motion sequence located before the target time point and a backward motion sequence located after the target time point in the motion sequence; an object detection module for, when there is an intention to open the tailgate, performing object detection on the vehicle surrounding images based on successive frames of vehicle surrounding images acquired during the pre-start phase to extract motion trajectory features for opening the tailgate; The control module is used to control the opening of the tailgate when the motion trajectory feature matches a preset feature library.

10. A vehicle, characterized in that: include: at least one processor; as well as, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to: the tailgate control method according to any one of claims 1 to 8.

Images