Parking position updating method in automatic parking, storage medium and electronic equipment

Abstract

The invention provides a parking position updating method in automatic parking, a storage medium and electronic equipment. The method comprises the steps that coordinates of four corner points of a parking space are acquired, and the four corner points of the parking space are arranged in sequence; in response to a received automatic parking instruction, establishing a coordinate system of the vehicle, and correspondingly converting the coordinates of the four angular points of the parking space into the coordinate system of the vehicle according to an arrangement sequence to form initial position coordinates of the angular points of the parking space; acquiring the real-time position of the vehicle, and updating the vehicle position coordinates under the coordinate system of the vehicle in real time based on the real-time position of the vehicle; acquiring a real-time parking space angular point position and a corresponding parking space angular point position timestamp, and updating an initial parking space angular point position coordinate under a coordinate system of a vehicle based on the real-time parking space angular point position, the parking space angular point position timestamp and a vehicle space-time array in which a vehicle position and a vehicle position timestamp are pre-stored, and forming updated parking space angular point position coordinates. The reliability and precision of automatic parking can be effectively improved.

Description

Technical Field

[0001] The present application belongs to the field of intelligent driving technology, and in particular to the field of vehicle position control technology in automatic parking. Background Art

[0002] In existing automatic parking systems, the core of automatic parking technology is to obtain real-time environmental information around the vehicle through sensors, identify parking space boundaries, and control the vehicle to park safely and efficiently in the target parking space. In order to achieve automatic parking, the system usually relies on cameras, radars, ultrasonic sensors and other devices for environmental perception and parking space identification, and uses on-board computing units for path planning and vehicle control.

[0003] The existing automatic parking system has the following main problems:

[0004] 1. Insufficient parking space detection accuracy

[0005] Most existing automatic parking systems rely on cameras or ultrasonic sensors for parking space recognition. However, in complex environments, such as poor lighting conditions, blurred parking space lines, or obstacles near the parking space boundaries, the camera may not be able to accurately identify the parking space corners. Ultrasonic sensors have low accuracy when detecting distant objects, which can easily cause errors in parking space information recognition. In the prior art, the sequential recognition and sorting of parking space corners are easily affected by environmental factors (such as obstacle occlusion, light changes, etc.), resulting in parking failures.

[0006] 2. Time delay leads to decreased parking accuracy

[0007] There is a certain delay from the time the camera captures the image to the time the automatic parking system processes the data and generates the parking path. In the case of fast-moving vehicles or changing parking spaces, this delay can cause the vehicle to miss the best parking opportunity. Especially in the real-time update phase of the parking process, the delay problem may cause the vehicle position and parking space information to be out of sync, thus affecting the parking accuracy.

[0008] 3. Lack of confidence

[0009] The confidence judgment mechanism of the current automatic parking system is relatively simple, often relying on data from a single sensor or a single condition to judge the validity of the parking corner point. This method is prone to misjudgment when multi-sensor data is inconsistent or there is a lot of environmental interference, resulting in deviations in the parking position or even collisions.

[0010] 4. Control system response lag

[0011] Existing automatic parking systems often rely on the real-time position and path planning of the vehicle in terms of vehicle control. However, due to the lag in updating sensor data during vehicle driving, the response speed of the control system may not be able to fully keep up with the dynamic changes of the vehicle. In this case, the parking trajectory of the vehicle may deviate, especially when precise adjustment of vehicle speed or steering is required, which can easily lead to parking failure or deviation.

[0012] 5. Inadequate filtering and updating mechanisms

[0013] In the prior art, the detection of vehicle position and parking space boundaries by automatic parking systems usually directly relies on the single detection results of sensors, lacking sufficient filtering or data smoothing. This approach may cause jitter or noise interference during the parking process, thereby affecting the precise alignment of the vehicle and the parking space. In addition, the corner point update mechanism is relatively passive and lacks a dynamic corner point update strategy, resulting in unstable corner point data during the parking process.

[0014] Therefore, although the current automatic parking system can provide a certain degree of automatic parking function, its reliability and accuracy in complex environments still need to be improved, especially in terms of parking space detection accuracy, time synchronization, confidence judgment and vehicle control response. The existing technology has not yet provided sufficient guarantees. Summary of the invention

[0015] The present application provides a parking position updating method, storage medium and electronic device in automatic parking, which are used to improve the reliability and accuracy of automatic parking.

[0016] In a first aspect, an embodiment of the present application provides a method for updating a parking position in automatic parking, comprising: acquiring coordinates of four corner points of a parking space, and arranging the four corner points of the parking space in sequence; in response to receiving an automatic parking instruction, establishing a coordinate system of the vehicle itself, and converting the coordinates of the four corner points of the parking space into the coordinate system of the vehicle itself in the order of arrangement, to form initial parking corner point position coordinates; acquiring a real-time position of the vehicle, and updating the vehicle position coordinates in the coordinate system of the vehicle itself based on the real-time position of the vehicle; acquiring a real-time parking corner point position and a corresponding parking corner point position timestamp, and updating the initial parking corner point position coordinates in the coordinate system of the vehicle itself based on the real-time parking corner point position, the parking corner point position timestamp, and a vehicle spatiotemporal array pre-stored with the vehicle position and the vehicle position timestamp, to form updated parking corner point position coordinates.

[0017] In an implementation of the first aspect, the initial parking space corner point position coordinates in the vehicle's own coordinate system are updated based on the real-time parking space corner point position, the parking space corner point position timestamp, and a vehicle spatiotemporal array pre-stored with vehicle positions and vehicle position timestamps to form updated parking space corner point position coordinates, including: searching the vehicle position timestamp with the smallest time difference with the parking space corner point position timestamp from the vehicle spatiotemporal array based on the parking space corner point position timestamp; obtaining the corresponding vehicle position based on the vehicle position timestamp with the smallest time difference, and obtaining the corresponding relationship between the vehicle position and the current vehicle position; obtaining the parking space corner point position corresponding to the real-time parking space corner point position based on the corresponding relationship, and converting the parking space corner point position into the parking space corner point position coordinates in the vehicle's own coordinate system.

[0018] In an implementation method of the first aspect, it also includes: respectively obtaining the initial parking space corner point position coordinates and the updated parking space corner point position coordinates of the four corner points of the parking space; detecting whether each corner point in the updated parking space corner point position coordinates matches the corresponding corner points in the initial parking space corner point position coordinates according to a preset matching algorithm: if so, maintaining the current arrangement order of each corner point corresponding to the updated parking space corner point position coordinates; if not, adjusting the arrangement order of each corner point corresponding to the updated parking space corner point position coordinates according to the matching result.

[0019] In an implementation method of the first aspect, it also includes: verifying whether the updated parking space corner point position coordinates simultaneously meet at least two preset confidence conditions; if so, confirming that the updated parking space corner point position coordinates are credible; if not, confirming that the updated parking space corner point position coordinates are unreliable, and re-acquiring the initial parking space corner point position coordinates and the updated parking space corner point position coordinates; wherein the confidence conditions include: whether any parking space corner point is within the field of view of the rear, left or right camera; whether two parking space corner points relatively close to the vehicle are located outside the camera stitching area; whether the vehicle steering wheel angle is within a threshold range.

[0020] In an implementation method of the first aspect, it also includes: constructing a quadrilateral area based on the corner points corresponding to the updated parking space corner point position coordinates; detecting whether the vehicle is appropriately located inside the quadrilateral area according to the vehicle position coordinates: if so, filtering the updated parking space corner point position coordinates using a fixed time window, and outputting the updated parking space corner point position coordinates after filtering; if not, directly outputting the updated parking space corner point position coordinates.

[0021] In an implementation method of the first aspect, it also includes: accumulating the number of consistency between the currently updated parking space corner point position coordinates and the last updated parking space corner point position; in response to the number of consistency being greater than or equal to a preset number, using the currently updated parking space corner point position coordinates for parking; in response to the number of consistency being less than a preset number, using the historically updated parking space corner point position coordinates for parking.

[0022] In an implementation of the first aspect, obtaining the real-time position of the vehicle includes: obtaining body data; establishing a vehicle motion equation based on the body data; obtaining the current speed of the vehicle, wheel status and sensor measurement data, and fusing the current speed of the vehicle, the wheel status, the sensor measurement data and the vehicle motion equation according to a filtering algorithm to obtain vehicle state parameters; wherein, in obtaining the wheel status, each wheel speed is mapped to a virtual speed relative to the center of the rear axle of the vehicle, and the wheel status is determined based on the difference between each of the virtual speeds; and the real-time position of the vehicle is obtained based on the vehicle state parameters.

[0023] In an implementation of the first aspect, the real-time position of the vehicle is obtained based on at least two of vehicle navigation data, image acquisition data, Internet of Vehicles data, positioning data, and vehicle body data.

[0024] In a second aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the parking position updating method in automatic parking described in any one of the first aspect of the present application.

[0025] In a third aspect, an embodiment of the present application provides an electronic device, comprising a processor and a memory; the memory stores program instructions; the processor is used to run the program instructions to execute the parking position update method in automatic parking described in any one of the first aspect of the present application.

[0026] The parking position updating method in automatic parking provided by the embodiment of the present application has the following beneficial effects:

[0027] This application accurately sorts the corner points according to the geometric relationship between the actual position of the vehicle and the corner points of the parking space, and can accurately identify the corner points of the parking space in different environments, greatly improving the accuracy of parking space detection. In actual parking scenarios, it shows a higher success rate and stability than the existing technology, especially in narrow and poorly lit parking environments. It can accurately identify the corner points of the parking space, and has higher accuracy and reliability in parking space detection, path planning and vehicle control, so it can effectively improve the reliability and accuracy of automatic parking. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 Shown is a flowchart of a method for updating parking position in automatic parking according to an embodiment of the present application.

[0029] Figure 2 Shown is a flow chart of updating the coordinates of parking space corner points in a method for updating parking position in automatic parking according to an embodiment of the present application.

[0030] Figure 3 Shown is a flow chart of ensuring the consistency of updated parking space corner point coordinates by sorting corner points in a parking position update method in automatic parking according to an embodiment of the present application.

[0031] Figure 4 Shown is a flow chart of ensuring the consistency of updated parking space corner point coordinates through confidence conditions in a parking position update method in automatic parking according to an embodiment of the present application.

[0032] Figure 5 Shown is a flow chart of ensuring the consistency of updated parking space corner point coordinates through filtering in a parking position update method in automatic parking according to an embodiment of the present application.

[0033] Figure 6 A flowchart showing whether to use updated parking space corner points during parking in a parking position updating method during automatic parking according to an embodiment of the present application is provided.

[0034] Figure 7 Shown is a schematic diagram of the overall implementation process of a method for updating parking position in automatic parking according to an embodiment of the present application.

[0035] Figure 8 Shown is a schematic diagram of the structure of an electronic device in an embodiment of the present application.

[0036] Component number description

[0037] 100 Electronic devices

[0038] 101 Memory

[0039] 102 Processor

[0040] 103 Display

[0041] S100~S400 steps

[0042] Steps S410 to S430

[0043] Steps S510 to S540

[0044] Steps S610 to S640

[0045] Steps S710 to S730 DETAILED DESCRIPTION

[0046] The following describes the embodiments of the present application through specific examples, and those skilled in the art can easily understand other advantages and effects of the present application from the contents disclosed in this specification. The present application can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present application. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0047] The existing automatic parking system is mainly composed of the following modules: sensor module, camera, ultrasonic sensor, radar sensor, parking space detection module and vehicle control module.

[0048] Among them, the sensor module is used to detect parking spaces and obstacles. Common sensors include: Camera: Use the front and rear cameras to capture images of the environment around the vehicle for visual perception. Ultrasonic sensors are used to measure the distance between the vehicle and obstacles. Radar sensors, such as millimeter-wave radars, can identify parking space boundaries and obstacle locations by emitting and receiving electromagnetic waves. The parking space detection module can identify the location and boundaries of the parking space based on the data sent back by the sensor. Existing automatic parking systems generally determine the corner coordinates of the parking space by the contours of objects or parking lines around the parking space, and use this to plan the parking path of the vehicle.

[0049] The vehicle control module is responsible for controlling the vehicle's steering, acceleration and braking to ensure that the vehicle is safely parked in the parking space along the planned path. The vehicle control module usually calculates a suitable driving route by combining the vehicle's real-time location information, speed and steering wheel angle.

[0050] The working principle of the existing automatic parking system is as follows:

[0051] Parking space detection stage: After the vehicle starts the automatic parking function, the sensor module begins to scan the surrounding environment. The parking space detection module extracts the boundary information of potential parking spaces from the environmental information, generates a geometric model of the parking space, and determines the parking path of the vehicle through a path planning algorithm.

[0052] Parking stage: The vehicle travels along the established route according to path planning, and the vehicle control module dynamically adjusts the vehicle speed and direction, and finally parks the vehicle in the target parking space.

[0053] Although the current automatic parking system can provide a certain degree of automatic parking function, its reliability and accuracy in complex environments still need to be improved, especially in terms of parking space detection accuracy, time synchronization, confidence determination, and vehicle control response.

[0054] In view of the time delay, inaccurate corner point recognition, untimely vehicle position update and other problems faced by the automatic parking system during parking detection and parking, the embodiments of the present application provide a parking position update method, storage medium and electronic device in automatic parking. The parking position update method, storage medium and electronic device provided in the embodiments of the present application are an improved solution integrating visual detection, time synchronization, corner point sorting and confidence verification. Through these technical means, the reliability and accuracy of automatic parking can be improved.

[0055] The main purpose of this embodiment is to solve the problems existing in the existing automatic parking system, such as insufficient parking space detection accuracy, reduced parking accuracy due to time delay, unreliable parking space corner confidence, and delayed vehicle control response, so as to improve the overall reliability and accuracy of automatic parking. Specifically, this embodiment can achieve the following goals:

[0056] 1) Improve parking space detection accuracy: Through multi-sensor fusion technology, the recognition ability of parking space corners is enhanced to solve the problem of inaccurate detection in complex environments such as lighting changes and obstacle occlusion.

[0057] 2) Solve the time synchronization problem: By designing a time synchronization mechanism, ensure that the vehicle’s current position is aligned with the detected parking space information in real time to avoid parking failures caused by system delays.

[0058] 3) Enhanced parking corner point sorting and confidence: Through an improved corner point sorting algorithm and multi-condition confidence verification, the stability and reliability of parking corner point information are improved and misjudgment is reduced.

[0059] 4) Optimize the smoothness of vehicle control: Use continuous detection confirmation and sliding window filtering strategies to ensure a smooth path during vehicle parking and avoid frequent adjustments and trajectory deviations.

[0060] The following will be combined with the attached examples of the present application Figure 1 To Attachment Figure 8 , the technical solution in the embodiment of the present application is described in detail, so that those skilled in the art can understand and implement the parking position updating method in the automatic parking of the present embodiment without creative work.

[0061] Figure 1 The flowchart of the method for updating the parking position in automatic parking in the embodiment of the present application is shown. Figure 1 As shown, the parking position updating method in automatic parking provided in the embodiment of the present application includes the following steps S100 to S400.

[0062] Step S100, obtaining the coordinates of the four corner points of the parking space, and arranging the four corner points of the parking space in sequence;

[0063] Step S200, in response to receiving the automatic parking instruction, establishing a coordinate system of the vehicle itself, and converting the coordinates of the four corner points of the parking space into the coordinate system of the vehicle itself in the order of arrangement, to form initial parking space corner point position coordinates;

[0064] Step S300, obtaining the real-time position of the vehicle, and updating the vehicle position coordinates in the vehicle's own coordinate system in real time based on the real-time position of the vehicle;

[0065] Step S400, obtain the real-time parking corner point position and the corresponding parking corner point position timestamp, and update the initial parking corner point position coordinates in the vehicle's own coordinate system based on the real-time parking corner point position, the parking corner point position timestamp and a vehicle spatiotemporal array that pre-stores the vehicle position and the vehicle position timestamp to form updated parking corner point position coordinates.

[0066] The above steps S100 to S400 of the parking position updating method in automatic parking in this embodiment are described in detail below.

[0067] Step S100, obtaining the coordinates of the four corner points of the parking space, and arranging the four corner points of the parking space in sequence.

[0068] In this embodiment, the coordinates of the four corner points of the parking space are obtained in the parking space detection stage, and the four corner points of the parking space are arranged in sequence. Specifically, a visual sensor is used to obtain the coordinates of the four corner points of the parking space, wherein the visual sensor preferably uses a monocular camera, a stereo camera or a laser radar. In addition, an ultrasonic sensor, a millimeter wave radar or an infrared sensor can also be used. These sensors can detect the boundaries of the parking space in different ways and generate the coordinates of the corner points. On this basis, the accuracy and robustness of the parking space detection can also be further improved by combining the advantages of different sensors through multi-sensor fusion technology.

[0069] In this embodiment, the order of the four corner points of the parking space is arranged in a clockwise or counterclockwise order according to the opening direction of the parking space. This embodiment adopts a multi-sensor fusion parking space corner point detection and sorting method, and solves the problem of inaccurate detection of traditional single sensors in complex environments by combining visual sensors with inertial navigation modules.

[0070] The automatic parking stage executes steps S200 to S400.

[0071] Step S200, in response to receiving the automatic parking instruction, establish the vehicle's own coordinate system, and convert the coordinates of the four corner points of the parking space into the vehicle's own coordinate system in the order of arrangement to form initial parking space corner point position coordinates.

[0072] In a specific implementation of this embodiment, in the coordinate system of the vehicle itself, the X-axis direction is along the front of the vehicle, and the Y-axis direction is along the left side of the vehicle.

[0073] Specifically, at the beginning of automatic parking, the vehicle's own coordinate system 01 is established, and the coordinates of the corner points of the parking space are converted to the vehicle's own coordinate system 01. The X-axis direction of the vehicle is the direction of the front of the vehicle, and the Y-axis direction is the left side of the vehicle. When the vehicle's own coordinate system 01 is established, the coordinate system is initialized and updated through the vehicle-mounted inertial navigation module or other positioning device (such as GPS or vehicle-mounted radar). At the same time, the coordinates of the four corner points of the parking space arranged in sequence are converted to the coordinate system 01 through coordinate transformation to form the initial parking space corner point position coordinates, for example, recorded as I0.

[0074] In addition, it should be noted that in addition to the inertial navigation module, the vehicle can also establish its own coordinate system through geomagnetic sensors or lane line recognition technology. These technologies can also provide the vehicle with the current direction and position, and are more applicable in certain environments, such as underground parking lots with weak GPS signals.

[0075] In this embodiment, by converting the coordinates of the four corner points of the parking space into the coordinate system of the vehicle itself in the order of arrangement, the problem of aligning the vehicle with the parking space can be solved, ensuring that the vehicle can be correctly guided into the parking space during the current parking operation.

[0076] Step S300, obtaining the real-time position of the vehicle, and updating the vehicle position coordinates in the vehicle's own coordinate system in real time based on the real-time position of the vehicle.

[0077] In a specific implementation of this embodiment, the real-time position of the vehicle is obtained based on at least two of vehicle navigation data, image acquisition data, vehicle networking data, positioning data, and vehicle body data. In addition, the parking lot may be equipped with a dedicated positioning beacon or induction coil, and the vehicle can update its own coordinates through these external devices.

[0078] During the automatic parking process, the real-time position and orientation of the vehicle will be continuously updated as the vehicle moves. In this embodiment, the position and orientation of the vehicle in the coordinate system 01 are updated in real time through visual detection, inertial navigation module, and vehicle body data (such as vehicle speed, steering wheel angle, etc.). Compared with the existing parking system, this embodiment obtains the real-time position of the vehicle by multi-sensor data fusion, which can ensure the accuracy of vehicle position update.

[0079] Specifically, in this embodiment, a specific implementation method of obtaining the real-time position of the vehicle includes: obtaining body data; establishing a vehicle motion equation based on the body data; obtaining the vehicle's current speed, wheel status, and sensor measurement data, and fusing the vehicle's current speed, the wheel status, the sensor measurement data, and the vehicle motion equation according to a filtering algorithm to obtain vehicle state parameters; wherein, in obtaining the wheel status, each wheel speed is mapped to a virtual speed relative to the center of the vehicle's rear axle, and the wheel status is determined based on the difference between each of the virtual speeds; and the vehicle's real-time position is obtained based on the vehicle state parameters.

[0080] Among them, in this embodiment, the vehicle body data includes various data required to construct the vehicle motion equation, and the vehicle body data includes but is not limited to tire sensor pulse and direction, steering wheel angle, rear axle center angular velocity output by ESC, vehicle wheelbase and vehicle width, etc.

[0081] In a specific implementation of this embodiment, the vehicle motion equation includes a state equation and an observation equation of a vehicle model.

[0082] In this embodiment, the state vector X=[x, y, θ, β, v, ω] is selected by considering the kinematic characteristics of the vehicle. T , and establish the state equation f(X k ,u k ):

[0083]

[0084] in:

[0085] x represents the horizontal coordinate of the center of the rear axle of the vehicle in the initial coordinate system;

[0086] y represents the ordinate of the center of the rear axle of the vehicle in the initial coordinate system;

[0087] θ represents the heading angle of the rear axle center of the vehicle in the initial coordinate system;

[0088] β represents the side slip angle of the vehicle’s rear axle center;

[0089] v represents the centerline speed of the vehicle's rear axle;

[0090] ω represents the angular velocity of the rear axle of the vehicle.

[0091] u β 、u v 、u ω is the process noise.

[0092] Select observation vector in:

[0093] Represents the observed value of the centerline speed of the vehicle's rear axle, output by the body sensor;

[0094] Represents the observed angular velocity of the rear axle center of the vehicle, output by the body sensor;

[0095] Represents the observed side slip angle of the rear axle of the vehicle. It can be expressed as: Among them, ξ is the front wheel steering angle, k1 is the proportional coefficient, and the value varies depending on the vehicle model.

[0096] The observation equation can be expressed as in To measure noise, h(X) can be expressed as:

[0097]

[0098] Among them, k2 is the proportional coefficient, and its value varies according to different vehicle models.

[0099] In this embodiment, the vehicle current speed, the wheel state, the sensor measurement data and the vehicle motion equation are fused according to the filtering algorithm, including: predicting the vehicle state and state covariance matrix at the next moment according to the state equation of the vehicle model; calculating the observation residual and the Kalman gain according to the sensor measurement data and the observation equation; obtaining the observation update of the vehicle state according to the observation residual, the Kalman gain and the predicted vehicle state at the next moment; obtaining the observation update of the state covariance matrix according to the predicted state covariance matrix at the next moment, the Kalman gain and the Jacobian matrix of the observation equation; and dynamically adjusting the parameters of the Kalman gain according to the vehicle current speed and the wheel state.

[0100] Specifically, the specific process of fusing the current speed of the vehicle, the wheel state, the sensor measurement data and the vehicle motion equation according to the filtering algorithm is as follows:

[0101] 1) State prediction: At each time step k, the state X at the next moment is predicted based on the state equation of the vehicle model in step a). k+1 , and predict the state covariance matrix:

[0102] (A): In predicting the vehicle state at the next moment according to the state equation of the vehicle model, the state X at the next moment is calculated based on the following formula: k+1 Prediction:

[0103] X k+1|k =f(X k ,u k )

[0104] Among them, f(X k ,u k ) is the state equation, X k+1|k is the vehicle state at the next moment, where u k is the process noise, that is, u in each time step k mentioned above β 、u v 、u ω .

[0105] (B): The state covariance matrix of the next moment is predicted according to the state equation of the vehicle model, and the state covariance matrix of the next moment is predicted based on the following formula:

[0106]

[0107] Among them, P k+1|k is the state covariance matrix at the next moment, P k is the current state covariance matrix, F k is the Jacobian matrix of the state equation, Q k is the process noise covariance matrix.

[0108] 2) Observation update: execution state X k+1 Observation update with state covariance matrix

[0109] The observation residual is calculated based on the sensor measurement data and the observation equation. Specifically, at time step k+1, the sensor observation Z is obtained k+1 , calculate the observed residual y k+1 :

[0110] y k+1 =Z k+1 -h(X k+1|k )

[0111] Among them, h(X k+1|k ) is the observation equation, and the predicted observations are based on the state estimation.

[0112] In a specific implementation of this embodiment, the Kalman gain is calculated based on the predicted state covariance matrix of the next moment, the Jacobian matrix of the observation equation, and the observation noise covariance matrix. Specifically, the Kalman gain K is calculated. k+1 The method is:

[0113] Among them, H k+1 is the Jacobian matrix of the observation equation, is the observation noise covariance matrix.

[0114] The observation update of the vehicle state is obtained according to the observation residual, the Kalman gain and the predicted vehicle state at the next moment, specifically: k+1 =X k+1|k +K k+1 y k+1

[0115] The observation update of the state covariance matrix is ​​obtained according to the predicted state covariance matrix of the next moment, the Kalman gain and the Jacobian matrix of the observation equation, specifically: k+1 =(IK k+1 H k+1 ) k+1|k .

[0116] In this embodiment, the covariance matrix is ​​dynamically adjusted according to the current speed of the vehicle and the abnormal state of the tire to improve the robustness of the filtering algorithm to the system state. Specifically, the observation noise covariance matrix in the Kalman gain is dynamically adjusted according to the current speed of the vehicle and the wheel state.

[0117] In a specific implementation of this embodiment, the dynamically adjusting the observation noise covariance matrix in the Kalman gain according to the current speed of the vehicle and the wheel state includes: dynamically constructing a diagonal matrix according to the current speed of the vehicle and the wheel state; using the diagonal matrix as the coefficient of the observation noise covariance matrix to adjust the observation noise covariance matrix.

[0118] When a tire is in an abnormal state, the confidence of the tire will be reduced. If the confidence data sent by the sensor is used directly, the vehicle position information obtained by subsequent calculations will be inaccurate. Therefore, in this embodiment, the observation noise covariance matrix is ​​dynamically adjusted according to the current vehicle speed and the abnormal state of the tire to improve the robustness of the filtering algorithm to the system state. R k+1 is a constant, λ is a diagonal matrix. In a specific implementation of this embodiment, the diagonal matrix is:

[0119]

[0120] Where S is the wheel state, and v is the current speed of the vehicle. (1:4)=S means that the elements of the 1st to 4th rows of the diagonal matrix are S, (5)=1 means that the elements of the 5th row of the diagonal matrix are 1, if the current speed of the vehicle is <0.1m / s, then the elements of the 6th row of the diagonal matrix are 0, otherwise, the elements of the 6th row of the diagonal matrix are 1, (7)=1 means that the elements of the 7th row of the diagonal matrix are 1.

[0121] This embodiment combines the extended Kalman filter (EKF) algorithm to perform vehicle positioning and dynamically adjusts the filter parameters, which can improve the accuracy and robustness of vehicle positioning.

[0122] In a specific implementation of the present embodiment, mapping the speed of each wheel to a virtual speed relative to the center of the rear axle of the vehicle includes: acquiring the wheel speed, the vehicle wheelbase and the vehicle width from the vehicle body data, and mapping the steering wheel angle to a turning radius according to a preset mapping table; mapping the speed of each rear wheel of the vehicle to a virtual speed relative to the center of the rear axle of the vehicle according to the wheel speed, the vehicle width and the turning radius; mapping the speed of each front wheel of the vehicle to a virtual speed relative to the center of the rear axle of the vehicle according to the wheel speed, the vehicle width, the vehicle wheelbase and the turning radius.

[0123] Among them, the vehicle four-wheel speed can be expressed as: v = [v lf ,v rf ,v lr ,v rr ], where v lf ,v rf ,v lr ,v rr Represent the speeds of the left front wheel, right front wheel, left rear wheel, and right rear wheel respectively.

[0124] In this embodiment, filtering technology is applied to smooth the wheel speed signal and reduce the impact of noise, and the vehicle kinematic model is used to calculate the virtual speed of each wheel mapped to the center of the vehicle's rear axle. Finally, a statistical method is used to analyze the four wheel speed signals to identify possible abnormal conditions (such as passing over speed bumps, potholes, and other road surfaces).

[0125] Specifically, since the output frequency of the vehicle's four-wheel speed is relatively high, this embodiment uses a filtering algorithm to smooth the wheel speed signal and reduce the impact of noise. In this embodiment, taking sliding window filtering as an example, the filtered four-wheel speed at time k is It is expressed as: Where N is the sliding window size, The sliding window filtering algorithm is used to reduce the noise in the wheel speed signal and improve the reliability of the wheel speed data.

[0126] Then, the vehicle kinematic model is used to calculate the virtual speed of each wheel mapped to the center of the vehicle's rear axle. The details are as follows:

[0127] Convert the steering wheel angle to the theoretical turning radius R through the corresponding table TT =H(δ), where δ is the steering wheel angle. In this embodiment, the turning radius calculated in the data preprocessing in step S100 can be directly called.

[0128] The virtual speed of the left rear wheel mapped to the center of the vehicle's rear axle can be expressed as:

[0129] The virtual speed of the right rear wheel mapped to the center of the vehicle's rear axle can be expressed as:

[0130] The virtual speed of the left front wheel mapped to the center of the vehicle's rear axle can be expressed as:

[0131] The virtual speed of the right front wheel mapped to the center of the vehicle's rear axle can be expressed as:

[0132] Where: W represents the width of the vehicle, L represents the wheelbase of the vehicle, R TT is the turning radius.

[0133] This embodiment converts the speed of each wheel into a virtual speed relative to the center of the rear axle of the vehicle through a vehicle kinematic model, forming a unified reference system to facilitate wheel abnormality detection.

[0134] In a specific implementation of this embodiment, determining the wheel state based on the difference between the virtual speeds includes: obtaining the mean and / or standard deviation of the difference between the virtual speed of any wheel mapped to the center of the rear axle of the vehicle and the virtual speeds of other wheels mapped to the center of the rear axle of the vehicle; when the absolute value of the mean is greater than the mean threshold and / or the absolute value of the standard deviation is greater than the standard deviation threshold, confirming that the wheel state is abnormal.

[0135] This embodiment uses a statistical method to analyze the four wheel speed signals of the vehicle to identify possible abnormal conditions, such as passing over speed bumps, potholes, and other road surfaces.

[0136] On a flat road, the four-wheel speeds corresponding to the virtual speed of the vehicle's rear axle center through the kinematic model should be the same. Considering factors such as sensor accuracy and data output frequency, this embodiment performs statistical analysis on the wheel speeds over a period of time to identify abnormal conditions and locate the wheels with abnormal conditions. The wheel state can be expressed as S = [S lf ,S rf ,S lr ,S rr ],S lf is the state of the left front wheel, S rf is the state of the right front wheel, S lr is the status of the left rear wheel, S rr is the state of the right rear wheel, each S lf ,S rf ,S lr ,S rr A value of 1 indicates abnormality and a value of 0 indicates normality.

[0137] Taking the left rear wheel as an example, calculate the mean and standard deviation of the virtual speeds of the right rear wheel, left front wheel, right front wheel mapped to the center of the vehicle's rear axle and the virtual speed of the left rear wheel mapped to the center of the vehicle's rear axle. When the absolute values ​​of the mean and standard deviation are greater than their respective corresponding thresholds, it means that the speed of the left rear wheel is unreliable. lr Set to 1, otherwise 0. The same applies to the right rear wheel, left front wheel, and right front wheel, which will not be repeated here.

[0138] This embodiment effectively identifies and locates the abnormal state of the wheels by statistically analyzing the virtual speed differences of the four wheels, thereby improving the accuracy of wheel abnormality detection.

[0139] In addition, in other implementations, the four-wheel speed data can also be pattern recognized through machine learning or deep learning algorithms to determine the wheel status, thereby further improving the accuracy and response speed of abnormality detection.

[0140] After the real-time position of the vehicle is acquired in the above manner, the vehicle position coordinates in the vehicle's own coordinate system are updated in real time based on the real-time position of the vehicle, and step S400 is continued.

[0141] Step S400, obtain the real-time parking corner point position and the corresponding parking corner point position timestamp, and update the initial parking corner point position coordinates in the vehicle's own coordinate system based on the real-time parking corner point position, the parking corner point position timestamp and a vehicle spatiotemporal array that pre-stores the vehicle position and the vehicle position timestamp to form updated parking corner point position coordinates.

[0142] In order to solve the time delay problem between visual detection and the device receiving the detection results, this embodiment establishes a vehicle time-space array and compares it with the timestamp information of visual detection, and selects the vehicle position information closest to the time point for time synchronization. Compared with the prior art, this embodiment innovatively introduces a time synchronization mechanism to ensure the real-time nature of vehicle detection and parking process, and guarantees the consistency of updating the coordinates of the parking space corners through time synchronization.

[0143] Figure 2 The flowchart of updating the parking space corner point coordinates in the parking position updating method in automatic parking according to one embodiment of the present application is shown. Figure 2 As shown, in a specific implementation of the present embodiment, the initial parking corner point position coordinates in the vehicle's own coordinate system are updated based on the real-time parking corner point position, the parking corner point position timestamp, and a vehicle spatiotemporal array that pre-stores the vehicle position and the vehicle position timestamp to form updated parking corner point position coordinates, including the following steps S410 to S430.

[0144] Step S410, searching the vehicle position timestamp having the smallest time difference with the parking space corner point position timestamp from the vehicle spatiotemporal array based on the parking space corner point position timestamp;

[0145] Step S420, obtaining a corresponding vehicle position based on the vehicle position timestamp with the smallest time difference, and obtaining a corresponding relationship between the vehicle position and the current vehicle position;

[0146] Step S430: acquiring the parking space corner point position corresponding to the real-time parking space corner point position based on the corresponding relationship, and converting the parking space corner point position into the parking space corner point position coordinates in the coordinate system of the vehicle itself.

[0147] Since there is a time delay from when the camera takes a picture to when the visual detection result is received, time synchronization is a key step to ensure parking accuracy. In the real-time position change of the vehicle, this embodiment pre-records the vehicle position P and the corresponding timestamp for each change, and stores the vehicle position P and the corresponding timestamp recorded multiple times in the vehicle spatiotemporal array [(P1, t1), (P2, t2), ..., (Pn, tn)], where P1, P2, ..., Pn are vehicle positions, and t1, t2, ..., tn are vehicle position timestamps corresponding to the vehicle positions. This embodiment pre-configures the capacity time span of the spatiotemporal array, for example, the capacity time span is 3 seconds, that is, the time difference between t1 and tn is 3 seconds. After receiving the parking space corner point and the parking space corner point position timestamp, record it as (Ik, tk), compare the parking space corner point position timestamp with the vehicle position timestamp in the vehicle time-space array, select the vehicle position timestamp tk that is closest to the shooting time (i.e., the time difference is the smallest) with the parking space corner point position timestamp, and then obtain the vehicle position Pk corresponding to the vehicle position timestamp tk for synchronization correction. That is, find the vehicle position Pk corresponding to the vehicle position timestamp tk in the time-space array, calculate the corresponding relationship between the current vehicle position Pn and the vehicle position Pk by coordinate conversion, and obtain the corresponding parking space corner point position In through the corresponding relationship and the real-time parking space corner point position Ik, and convert the parking space corner point position In to its own coordinate system 01, and record it as the parking space corner point position coordinate In'. In order to ensure the accuracy of synchronization, when the time difference between the vehicle position timestamp tk in the vehicle time-space array and the parking space corner point position timestamp is less than the preset time threshold, it is considered to be synchronized successfully, otherwise it is considered to be failed. For example, the preset time threshold can be set according to actual needs, such as 0.1 seconds.

[0148] This embodiment aligns the vehicle spatiotemporal array with the timestamp to reduce the problem of inaccurate parking space information caused by visual delay and ensure the real-time nature of the parking process.

[0149] Figure 3The flowchart is shown in the method for updating the parking position in automatic parking according to one embodiment of the present application, which ensures the consistency of the updated parking space corner point coordinates by sorting the corner points. Figure 3 As shown, in a specific implementation of this embodiment, the following steps S510 to S540 are also included.

[0150] Step S510, respectively obtaining the initial parking space corner point position coordinates and the updated parking space corner point position coordinates of the four corner points of the parking space;

[0151] Step S520, detect whether the corner points in the updated parking space corner point position coordinates match the corner points in the initial parking space corner point position coordinates according to a preset matching algorithm: if so, continue to execute step S530: maintain the current arrangement order of the corner points corresponding to the updated parking space corner point position coordinates; if not, continue to execute step S540: adjust the arrangement order of the corner points corresponding to the updated parking space corner point position coordinates according to the matching results.

[0152] In this embodiment, a preset matching algorithm is used to sort the parking space corner points detected by vision, and the consistency of the updated parking space corner point coordinates is ensured by sorting the corner points. If an error is found in the corner point correspondence during the sorting process, the sorting is deemed to have failed. This embodiment improves the accuracy problem in the parking space corner point matching process and ensures the validity of the parking space information.

[0153] In a specific implementation of this embodiment, the two sets of parking space corner point coordinates (I0In', i.e., the initial parking space corner point position coordinates and the updated parking space corner point position coordinates) are calculated using a preset matching algorithm, i.e., corresponding matching is performed for each corner point of the parking space. If it is found that the corresponding positions of the same corner point in the initial parking space corner point position coordinates and the updated parking space corner point position coordinates are different, the entire corner point sequence in the updated parking space corner point position coordinates In' is readjusted according to the matching results and the corner point sequence confirmed in the initial parking space corner point position coordinates I0.

[0154] The preset matching algorithm matches the corner points according to the rule of minimum Euclidean distance, for example. In addition to the Euclidean distance matching algorithm, a matching algorithm based on shape similarity, such as Hausdorff distance or curvature matching algorithm, can also be used. The preset matching algorithm has better robustness and noise resistance in complex environments, and an abnormality detection module can also be added. If a corner point sorting failure is detected, re-detection or other strategies can be adopted to improve parking safety.

[0155] Figure 4 The flowchart of the method for updating the parking position in automatic parking according to one embodiment of the present application is shown as follows: Figure 4As shown, in a specific implementation method of this embodiment, it also includes: verifying whether the updated parking space corner point position coordinates simultaneously meet at least two preset confidence conditions; if so, confirming that the updated parking space corner point position coordinates are credible; if not, confirming that the updated parking space corner point position coordinates are uncredible, and re-acquiring the initial parking space corner point position coordinates and the updated parking space corner point position coordinates; wherein the confidence conditions include: whether any parking space corner point is within the field of view of the rear, left or right camera; whether two parking space corner points that are relatively close to the vehicle are located outside the camera stitching area; whether the vehicle steering wheel angle is within a threshold range.

[0156] For example, if any corner point is within the field of view of the rear, left, or right camera, it is considered a reliable number; if the two corner points closer to the vehicle are outside the camera stitching area, they are considered reliable data; if the vehicle steering wheel angle is within a small range and the vehicle's tilt angle is small, the output corner point is considered a reliable number. The current output corner point coordinates must meet the above confidence conditions at the same time to ensure that the output corner point coordinates are accurate and reliable, otherwise the output corner point coordinates are considered unreliable.

[0157] This embodiment uses multiple confidence conditions, such as the corner point is within the camera range, the vehicle inclination angle is small, etc., to verify the confidence of the corner point output after the consistency detection of the updated parking space corner point coordinates is ensured by corner point sorting, and further ensures the consistency of the updated parking space corner point coordinates through confidence conditions. The confidence detection judgment of this embodiment is different from the confidence judgment of a single condition in the existing system, which enhances the reliability of the consistency of the updated parking space corner point coordinates.

[0158] In addition, in this embodiment, a confidence strategy based on machine learning can also be added. By training a neural network or other learning algorithm, the parking system can adaptively adjust the confidence standard according to historical parking data and vehicle status. For example, when it is detected that the vehicle has performed parking operations under similar conditions many times, the system can automatically lower the threshold of certain confidence conditions to improve parking efficiency.

[0159] In order to improve the stability of the corner coordinates of the parking space, this embodiment adopts the sliding window mean filtering technology. Through the sliding window mean filtering, the corner coordinates of the vehicle when it stays inside the parking space are stabilized, reducing the detection error caused by slight shaking of the vehicle. Compared with the traditional system, this embodiment adopts a more flexible filtering strategy, which can adapt to parking space detection in a dynamic environment.

[0160] Figure 5 The flowchart is shown in the method for updating the parking position in automatic parking according to one embodiment of the present application, which ensures the consistency of the updated parking space corner coordinates through filtering. Figure 5As shown, in a specific implementation of this embodiment, the following steps S610 to S640 are also included.

[0161] Step S610, constructing a quadrilateral region based on each corner point corresponding to the updated parking space corner point position coordinates;

[0162] Step S620, detect whether the vehicle is properly located inside the quadrilateral area according to the vehicle position coordinates: if so, continue to execute step S630: use a fixed time window to filter the updated parking space corner point position coordinates, and output the updated parking space corner point position coordinates after filtering; if not, continue to execute step S640: directly output the updated parking space corner point position coordinates.

[0163] Specifically, a quadrilateral area is constructed based on the currently output corner point coordinates of the parking space corner points and a judgment is made. If the vehicle is inside the quadrilateral, a fixed time window is enabled to perform mean filtering on each corner point to eliminate noise and ensure the stability of the parking space boundary; if it is not inside the quadrilateral, the updated parking space corner point position coordinates In' are directly output. Among them, whether the vehicle is inside the quadrilateral is confirmed by comparing the corner point coordinates corresponding to the parking space corner points and the vehicle position.

[0164] In order to ensure the smoothness of vehicle control, this embodiment does not update and use the parking angle output after time synchronization, corner point sorting, confidence condition screening, and filtering processing every time it is detected. For example, the corner point update operation is performed only after the consistent parking corner point information is detected for multiple consecutive times, otherwise the historical corner point is continued to be used for parking to avoid unnecessary updates caused by visual detection jitter or error, ensuring a smooth parking process.

[0165] In addition, it should be noted that in addition to sliding window mean filtering, more complex filtering techniques such as Kalman filtering or particle filtering can also be used. These filtering methods can better handle the problem of multi-source sensor data fusion and more accurately estimate the relative position of the vehicle and the parking space in a dynamic environment.

[0166] This embodiment verifies the confidence of the corner point coordinates through multiple conditions and introduces a sliding window mean filtering algorithm to improve the stability and accuracy of the parking process.

[0167] Figure 6 The flowchart of whether to use the updated parking space corner point in the parking position update method in automatic parking in one embodiment of the present application is shown. Figure 6 As shown, in a specific implementation of this embodiment, the following steps S710 to S730 are also included.

[0168] Step S710, accumulate the number of times the currently updated parking space corner point position coordinates are consistent with the last updated parking space corner point position; the consistency of the updated parking space corner point position coordinates with the last updated parking space corner point position refers to the result of processing through the above-mentioned time synchronization, corner point sorting, confidence condition screening, filtering processing and other steps.

[0169] Step S720, in response to the number of consistency being greater than or equal to the preset number of consistency, using the currently updated parking space corner point position coordinates for parking;

[0170] Step S730: In response to the number of consistency being less than a preset number, parking is performed using the updated parking space corner point position coordinates updated historically.

[0171] This embodiment determines whether the updated parking space corner point position coordinates can be used for parking by continuous detection. For example, the parking space corner point is updated only when the parking space corner point is detected five times in a row to ensure the smoothness of vehicle control and avoid parking track jitter caused by frequent updates.

[0172] Determine whether to use the parking corner points output after the consistency test of the above-mentioned time synchronization, corner point sorting, confidence condition screening, and filtering processing for parking. If the conditions are met, accept this group of parking corner points for parking, otherwise use the stored historical parking corner points for parking.

[0173] This embodiment updates the corner point information after multiple consecutive consistency checks, thereby ensuring smoothness of vehicle control and avoiding overly frequent adjustments.

[0174] In the update strategy, the update frequency of the corner points can be dynamically adjusted according to parameters such as vehicle speed and steering wheel angle through an adaptive update frequency algorithm. When the vehicle approaches a parking space, the update frequency increases; when it moves away from a parking space, the update frequency decreases.

[0175] Figure 7 The figure shows the overall implementation process of the method for updating the parking position in automatic parking according to one embodiment of the present application. Figure 7 As shown, the specific implementation process of this embodiment is as follows:

[0176] 1. Parking space detection stage:

[0177] A visual sensor (which can be a monocular or binocular camera or a laser radar, etc.) is used to obtain the coordinates of the four corner points of the parking space, and the order of the corner points is memorized according to the direction of the parking space opening. Compared with the prior art, this embodiment introduces a combination of multiple sensors, which can adapt to different types of parking lot environments and improve the accuracy of parking space detection.

[0178] 2. Automatic parking stage:

[0179] Coordinate system establishment: At the beginning of automatic parking, the vehicle's own coordinate system 01 is established, and the coordinates of the corner points of the parking space are converted to this coordinate system. The X-axis direction of the vehicle is the direction of the front of the vehicle, and the Y-axis direction is the left side of the vehicle. Among them, this embodiment uses the vehicle's inertial navigation module, GPS or radar and other equipment to solve the problem of aligning the vehicle with the parking space.

[0180] Vehicle position update: The position and orientation of the vehicle in coordinate system 01 are updated in real time through visual detection, inertial navigation module, and vehicle body data (such as vehicle speed, steering wheel angle, etc.). Compared with the existing parking system, this embodiment adds multi-sensor data fusion to ensure the accuracy of vehicle position update.

[0181] Time synchronization: To solve the time delay problem between visual detection and the device receiving the detection results, this embodiment establishes a vehicle time-space array and compares it with the timestamp information of the visual detection, and selects the vehicle position information closest to the time point for time synchronization. Compared with the prior art, this embodiment innovatively introduces a time synchronization mechanism to ensure the real-time nature of vehicle detection and parking process.

[0182] Corner point sorting: The Euclidean distance algorithm is used to sort the corner points of parking spaces detected by vision. If the corner point correspondence error is found during the sorting process, the sorting is considered to have failed. This technology improves the accuracy of the parking space corner point matching process and ensures the validity of the parking space information.

[0183] Confidence strategy: Through multiple conditions, such as the corner point is within the camera range, the vehicle tilt angle is small, etc., the confidence of the corner points output after sorting is verified. This strategy is different from the confidence judgment of a single condition in the existing system, which increases the reliability of the system.

[0184] Filtering strategy: Through sliding window mean filtering, the corner coordinates of the vehicle when it is parked inside the parking space are stabilized to reduce the detection error caused by slight shaking of the vehicle. Compared with the traditional parking system, this embodiment adopts a more flexible filtering strategy that can adapt to parking space detection in a dynamic environment.

[0185] Update strategy: This embodiment designs a continuous detection mechanism. The parking corner point is updated only when the parking corner point is detected for multiple consecutive times to ensure the smoothness of vehicle control and avoid parking trajectory jitter caused by frequent updates.

[0186] This embodiment effectively improves the accuracy and reliability of the parking system, and at the same time, through the combination of multiple sensors and algorithms, provides more guarantees for parking space detection and vehicle control during automatic parking, overcoming many limitations in the prior art. The parking position update method in automatic parking described in the embodiment of the present application has significant technical advantages and positive effects compared to the prior art, which are mainly reflected in the following aspects:

[0187] 1. Improve the accuracy and robustness of parking space detection

[0188] The parking position update method in automatic parking described in the embodiment of the present application adopts multi-sensor fusion technology (such as camera, radar, inertial navigation module) to enable the parking system to accurately identify the corner points of the parking space in different environments, and accurately sort the corner points according to the geometric relationship between the actual position of the vehicle and the corner points of the parking space. This technical means greatly improves the accuracy of parking space detection, especially under complex conditions such as insufficient lighting and obstruction by obstacles. The sorting and calibration algorithm of the parking space corner points ensures the consistency of the parking space information and avoids the common parking space corner point recognition errors in the prior art.

[0189] 2. Solve the problem of inaccurate parking caused by time delay

[0190] The parking position update method in automatic parking described in the embodiment of the present application uses a time synchronization mechanism to accurately align the current position information of the vehicle with the parking space corner point data detected by vision, thereby solving the time delay problem from image capture to data processing. Specifically, the historical position information is recorded using the vehicle spatiotemporal array and compared with the timestamp of visual detection, and the position information with the smallest time difference is selected to dynamically update the coordinates of the parking space corner point. This method effectively avoids the phenomenon of the vehicle position and parking space information being out of sync during the parking process, and ensures the accuracy of the vehicle parking path.

[0191] 3. Improve the reliability of parking corner point sorting and confidence verification

[0192] The parking space corner point sorting method proposed in the automatic parking position updating method described in the embodiment of the present application ensures the correctness of the corner points by calculating the Euclidean distance between two groups of parking space corner points, classifying and minimizing them. At the same time, confidence verification is performed in combination with multiple conditions (such as camera viewing angle, vehicle tilt angle, etc.) to ensure the credibility of the parking space corner point data. Compared with the confidence algorithm of single condition judgment in the prior art, the parking space position updating method described in the embodiment of the present application greatly improves the stability and accuracy of parking space detection through multi-condition constraints, and reduces misjudgment.

[0193] 4. Filtering strategy effectively improves the stability of parking space corners

[0194] The method for updating the parking position in automatic parking described in the embodiment of the present application adopts a sliding window mean filtering technique to filter the corner points of the vehicle coordinates located in the parking space, ensuring the smoothness and stability of the parking space corner point data during the parking process. Through the filtering strategy, the instability of the corner point information caused by slight vehicle shaking or environmental noise interference is effectively reduced. This technical means ensures that the path of the vehicle is smoother and more accurate during the parking process, especially during fine-tuning.

[0195] 5. Improve the smoothness of vehicle control

[0196] The parking position update method in the automatic parking described in the embodiment of the present application ensures the accuracy of the parking space corner point update by designing a continuous 5-time detection and confirmation mechanism, thereby avoiding frequent adjustments and jitters of the vehicle during the parking process. This strategy ensures that the system will not make adjustments immediately after detecting the parking space corner point, but ensures the stability of the parking space information through multiple confirmations, thereby improving the smoothness of vehicle control and avoiding trajectory deviation or vibration caused by excessive correction during parking.

[0197] 6. Flexibility to adapt to different environments and complex conditions

[0198] The multi-sensor fusion technology of the automatic parking position update method described in the embodiment of the present application can adapt to a variety of different parking environments, including open-air parking lots, underground garages, etc., and can cope with different types of parking spaces (parallel parking spaces, vertical parking spaces, etc.). At the same time, through time synchronization and confidence verification, the system can dynamically adapt to the influence of lighting, obstacles, etc. in complex environments to ensure the reliability and accuracy of parking.

[0199] 7. Real-time dynamic update of vehicle and parking space information

[0200] The parking position update method in the automatic parking described in the embodiment of the present application can ensure the real-time synchronization update of the vehicle and the parking space corner point information through the time synchronization mechanism, the confidence strategy and the dynamic update strategy, and ensure the real-time control of the parking space boundary by the vehicle control system during the parking process. This mechanism solves the problem of the asynchrony between the vehicle and the parking space information in the prior art, and avoids parking failures caused by delays.

[0201] Through experimental testing, the parking position update method in automatic parking described in the embodiment of the present application shows a higher success rate and stability in actual parking scenarios compared with the existing technology, especially in narrow and poorly lit parking environments, the system can accurately identify the corner points of the parking space and successfully park. The introduction of the filtering strategy significantly reduces the deviation and vibration of the vehicle trajectory during parking, ensuring the smoothness of the parking path. Compared with the existing parking system, the parking position update method in automatic parking described in the embodiment of the present application shows higher accuracy and reliability in all aspects of parking space detection, path planning and vehicle control.

[0202] In summary, the parking position update method in automatic parking described in the embodiment of the present application has higher accuracy, stability and reliability through multiple technical improvements to the existing automatic parking system, especially in terms of parking space detection accuracy, time synchronization, confidence verification and vehicle control smoothness, and significantly solves many problems in the prior art. Through multi-sensor data fusion, time synchronization and dynamic update mechanism, the parking position update method in automatic parking described in the embodiment of the present application effectively improves the adaptability and performance of the automatic parking system in complex environments.

[0203] The protection scope of the parking position update method in automatic parking described in the embodiment of the present application is not limited to the execution order of the steps listed in the present embodiment. All solutions implemented by adding, reducing or replacing steps in the prior art based on the principles of the present application are included in the protection scope of the present application.

[0204] An embodiment of the present application further provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for updating the parking position in automatic parking provided in any embodiment of the present application is implemented.

[0205] In the embodiment of the present application, any combination of one or more storage media can be used. The storage medium can be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium can be, for example, - but not limited to - an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples (non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a RAM, a ROM, an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this document, a computer-readable storage medium can be any tangible medium containing or storing a program, which can be used by an instruction execution system, a device or a device or used in combination with it.

[0206] An embodiment of the present application also provides an electronic device. Figure 8 Shown is a schematic diagram of the structure of the electronic device 100 provided in the embodiment of the present application. In some embodiments, the electronic device may be a mobile phone, a tablet computer, a wearable device, a vehicle-mounted device, an augmented reality (AR) / virtual reality (VR) device, a laptop computer, an ultra-mobile personal computer (UMPC), a netbook, a personal digital assistant (PDA) and other terminal devices. In addition, the parking position update method in automatic parking provided in the present application can also be applied to databases, servers and service response systems based on terminal artificial intelligence. The embodiment of the present application does not impose any restrictions on the specific application scenarios of the parking position update method in automatic parking.

[0207] like Figure 8 As shown, the electronic device 100 provided in the embodiment of the present application includes a memory 101 and a processor 102 .

[0208] The memory 101 is used to store computer programs; preferably, the memory 101 includes: ROM, RAM, disk, USB flash drive, memory card or CD-ROM and other media that can store program codes.

[0209] Specifically, the memory 101 may include a computer system readable medium in the form of a volatile memory, such as a random access memory (RAM) and / or a cache memory. The electronic device 100 may further include other removable / non-removable, volatile / non-volatile computer system storage media. The memory 101 may include at least one program product having a set (e.g., at least one) of program modules, which are configured to perform the functions of the various embodiments of the present application.

[0210] The processor 102 is connected to the memory 101 and is used to execute the computer program stored in the memory 101 so that the electronic device 100 executes the parking position updating method in automatic parking provided in any embodiment of the present application.

[0211] Optionally, the processor 102 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0212] Optionally, the electronic device 100 in this embodiment may further include a display 103. The display 103 is communicatively connected with the memory 101 and the processor 102, and is used to display a GUI interaction interface related to the parking position update method in automatic parking.

[0213] In summary, the present application accurately sorts corner points according to the geometric relationship between the actual position of the vehicle and the corner points of the parking space, can accurately identify the corner points of the parking space in different environments, greatly improves the accuracy of parking space detection, and shows a higher success rate and stability in actual parking scenarios compared to the existing technology, especially in narrow and poorly lit parking environments, can accurately identify the corner points of the parking space, and has higher accuracy and reliability in parking space detection, path planning and vehicle control, so it can effectively improve the reliability and accuracy of automatic parking. Therefore, the present application effectively overcomes the various shortcomings of the existing technology and has a high industrial utilization value.

[0214] The above embodiments are merely illustrative of the principles and effects of the present application and are not intended to limit the present application. Anyone familiar with the technology may modify or change the above embodiments without violating the spirit and scope of the present application. Therefore, all equivalent modifications or changes made by a person of ordinary skill in the art without departing from the spirit and technical ideas disclosed in the present application shall still be covered by the claims of the present application.

Claims

1. A parking position updating method in automatic parking, characterized in that: include: Get the coordinates of the four corner points of the parking space and arrange the four corner points of the parking space in sequence; In response to receiving the automatic parking instruction, establishing the vehicle's own coordinate system, and converting the coordinates of the four corner points of the parking space into the vehicle's own coordinate system in the order of arrangement, to form initial parking space corner point position coordinates; Acquire the real-time position of the vehicle, and update the vehicle position coordinates in the vehicle's own coordinate system in real time based on the real-time position of the vehicle; The real-time parking corner point position and the corresponding parking corner point position timestamp are obtained, and based on the real-time parking corner point position, the parking corner point position timestamp and a vehicle spatiotemporal array pre-stored with vehicle positions and vehicle position timestamps, the initial parking corner point position coordinates in the vehicle's own coordinate system are updated to form updated parking corner point position coordinates.

2. The parking position updating method in automatic parking according to claim 1, characterized in that: The updating of the initial parking space corner position coordinates in the vehicle's own coordinate system based on the real-time parking space corner position, the parking space corner position timestamp, and a vehicle spatiotemporal array pre-stored with vehicle positions and vehicle position timestamps to form updated parking space corner position coordinates includes: Based on the parking space corner point position timestamp, searching from the vehicle spatiotemporal array for the vehicle position timestamp with the smallest time difference from the parking space corner point position timestamp; Acquire the corresponding vehicle position based on the vehicle position timestamp with the smallest time difference, and acquire the corresponding relationship between the vehicle position and the current vehicle position; Based on the corresponding relationship, a parking space corner point position corresponding to the real-time parking space corner point position is acquired, and the parking space corner point position is converted into the parking space corner point position coordinates in the coordinate system of the vehicle itself.

3. The parking position updating method in automatic parking according to claim 1 or 2, characterized in that: Also includes: Respectively obtaining the initial parking space corner point position coordinates and the updated parking space corner point position coordinates of the four corner points of the parking space; According to a preset matching algorithm, it is detected whether each corner point in the updated parking space corner point position coordinates corresponds to each corner point in the initial parking space corner point position coordinates: If yes, keep the current arrangement order of the corner points corresponding to the updated parking space corner point position coordinates; If not, the arrangement order of the corner points corresponding to the updated parking space corner point position coordinates is adjusted according to the matching result.

4. The parking position updating method in automatic parking according to claim 3, characterized in that: Also includes: Verify whether the updated parking space corner point position coordinates satisfy at least two preset confidence conditions at the same time; if so, confirm that the updated parking space corner point position coordinates are credible; if not, confirm that the updated parking space corner point position coordinates are uncredible, and re-acquire the initial parking space corner point position coordinates and the updated parking space corner point position coordinates; The confidence condition includes: Whether any corner of the parking space is within the field of view of the rear, left or right camera; Whether the two parking space corner points that are relatively close to the vehicle are outside the camera stitching area; Whether the vehicle's steering wheel angle is within the threshold range.

5. The parking position updating method in automatic parking according to claim 4, characterized in that: Also includes: Constructing a quadrilateral region based on each corner point corresponding to the updated parking space corner point position coordinates; Detect whether the vehicle is properly located within the quadrilateral area according to the vehicle position coordinates: If yes, filtering the updated parking space corner point position coordinates using a fixed time window, and outputting the updated parking space corner point position coordinates after filtering; If not, the updated parking space corner point position coordinates are directly output.

6. The parking position updating method in automatic parking according to claim 1, characterized in that: Also includes: Accumulating the number of times the coordinates of the currently updated parking space corner point position are consistent with the last updated parking space corner point position; In response to the number of consistency being greater than or equal to a preset number of times, using the currently updated parking space corner point position coordinates for parking; In response to the number of consistency being less than a preset number, parking is performed using the updated parking space corner point position coordinates updated historically.

7. The parking position updating method in automatic parking according to claim 1, characterized in that: The obtaining of the real-time position of the vehicle comprises: Get vehicle body data; Establishing a vehicle motion equation based on the vehicle body data; Acquire the current speed of the vehicle, the wheel state and the sensor measurement data, and fuse the current speed of the vehicle, the wheel state, the sensor measurement data and the vehicle motion equation according to a filtering algorithm to obtain a vehicle state parameter; wherein, in acquiring the wheel state, each wheel speed is mapped to a virtual speed relative to the center of the rear axle of the vehicle, and the wheel state is determined according to the difference between each virtual speed; The real-time position of the vehicle is obtained according to the vehicle status parameters.

8. The parking position updating method in automatic parking according to claim 1, characterized in that: The real-time position of the vehicle is obtained based on at least two of vehicle navigation data, image acquisition data, vehicle networking data, positioning data and vehicle body parameters.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the parking position updating method in automatic parking according to any one of claims 1 to 8 is implemented.

10. An electronic device, characterized in that: The electronic device comprises: Processor and memory; The memory stores program instructions; The processor is used to run the program instructions to execute the parking position updating method in automatic parking according to any one of claims 1 to 8.

Images