A memory automatic parking method and a parking system

Abstract

The present invention provides a memory-based automatic parking method and a parking system. The method includes: during the automatic parking process, when the parking path planning fails according to the searched parking space and the position of the vehicle itself, searching for a matching historical parking trajectory in the SLAM positioning feature map; extracting the key points of the parking path of the historical parking trajectory; re-planning a new parking path based on the extracted key points of the parking path; and parking the vehicle into the parking space based on the new parking path. In view of the parking space scenario where the path planning of the automatic parking system fails, the automatic parking system extracts the key path points of the driver's parking trajectory into this parking space by memorizing the historical trajectory of the driver parking into this parking space, re-plans and fits the automatic parking trajectory, so as to realize automatic parking into the parking space. It solves the problem that the automatic parking cannot be automatically parked into the parking space due to the failure of the automatic parking path planning.

Description

Technical Field

[0001] The present invention relates to the field of parking control, and more particularly, to a memory automatic parking method and a parking system. Background Art

[0002] With the development of intelligent driving, the demand of drivers for automatic parking systems is increasing, and more and more vehicle models are equipped with automatic parking systems. However, in the face of complex and changeable parking environments, the automatic parking system is limited by the strict requirements of the path planning algorithm for space and distance, resulting in the failure of path planning for 30% or more of the parking spaces, thus leading to parking failure and great complaints from customers. In order to improve the customer experience and reduce customer complaints, it is very important to study how to solve the problem of path planning failure of the automatic parking system for some parking spaces. Summary of the Invention

[0003] The present invention provides a memory automatic parking method and a parking system for solving the problem of path planning failure in the prior art.

[0004] According to a first aspect of the present invention, there is provided a memory automatic parking method, including:

[0005] During automatic parking, when the parking path planning fails according to the searched parking space and the position of the vehicle itself, find a matching historical parking trajectory in the SLAM positioning feature map;

[0006] Extract the parking path key points of the historical parking trajectory;

[0007] Based on the extracted parking path key points, re-plan a new parking path;

[0008] Park the vehicle into the parking space based on the new parking path.

[0009] According to a second aspect of the present invention, there is provided a memory automatic parking system, including:

[0010] A path planning module for, during automatic parking, when the parking path planning fails according to the searched parking space and the position of the vehicle itself, finding a matching historical parking trajectory in the SLAM positioning feature map; extracting the parking path key points of the historical parking trajectory; and re-planning a new parking path based on the extracted parking path key points;

[0011] A motion control module for parking the vehicle into the parking space based on the new parking path.

[0012] A memory-based automatic parking method and a parking system provided by the present invention are directed to a parking space scenario where the path planning of the automatic parking system fails. The automatic parking system extracts key path points of the driver's parking trajectory into this parking space through the memorized historical trajectory of the driver parking into this parking space, and re-plans and fits the automatic parking trajectory, thereby realizing automatic parking into the parking space. This solves the problem that the automatic parking fails to park automatically due to the failure of the automatic parking path planning. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 FIG. is a flowchart of a memory-based automatic parking method provided by the present invention;

[0014] Figure 2 FIG. is a schematic diagram of a conventional automatic parking path planning;

[0015] Figure 3 FIG. is a schematic diagram of a scenario where the automatic parking path planning fails;

[0016] Figure 4 FIG. is a schematic diagram of the process of path planning failure and extraction of key parking path points;

[0017] Figure 5 FIG. is a schematic diagram of the process of path re-planning and automatic parking into the parking space;

[0018] Figure 6 FIG. is a schematic diagram of tracking key points of the parking path;

[0019] Figure 7 FIG. is a schematic structural diagram of a memory-based automatic parking system provided by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0020] The following further describes in detail the specific embodiments of the present invention with reference to the drawings and embodiments. The following embodiments are used to illustrate the present invention but are not used to limit the scope of the present invention.

[0021] Embodiment 1

[0022] A memory-based automatic parking method, referring to Figure 1 , the memory-based automatic parking method includes:

[0023] S1. During the automatic parking process, when the parking path planning fails according to the searched parking space and the position of the vehicle itself, search for a matching historical parking trajectory in the SLAM positioning feature map.

[0024] As an embodiment, determine the success or failure result of the parking path planning from the vehicle position to the parking space according to the searched parking space information, vehicle position information, and obstacle information around the parking space.

[0025] It is understandable that when the vehicle drives into the parking lot and searches for a parking space, automatic parking path planning is performed. Figure 2 It is a schematic diagram of conventional automatic parking path planning. Due to the obstruction of the front wall, the distance between points P0 and P1 is insufficient, resulting in the failure of parking path planning (point P0 is the center point of the rear axle of the vehicle).

[0026] Among them, multiple sensing modules can be installed on the vehicle. For example, multiple surround-view cameras and multiple ultrasonic radars can be installed. When the vehicle enters the parking lot, they are used to sense the information of the searched parking space and the obstacle information around the parking space. Based on the environmental information such as obstacles around the parking space, parking path planning is carried out. If the parking path planning fails, refer to Figure 3 , which is a schematic diagram of the failure scenario of automatic parking path planning. Figure 3 In this scenario, after the automatic parking system searches for parking space No. 002, due to the limitation of the front wall, the front passing distance is insufficient, resulting in the failure of parking path planning and the vehicle cannot be parked into the parking space.

[0027] At this time, the historical parking trajectory can be referred to for parking. Among them, for the historical parking process, refer to Figure 4 , when the automatic parking path planning is successful, the vehicle is automatically parked into the parking space; when the automatic parking path planning fails, the driver parks the vehicle by himself / herself, and the automatic parking system will memorize the trajectory of the driver parking the vehicle into the parking space, which is called the historical parking trajectory. Extract the path key points on the historical parking trajectory, turn on SLAM vision, establish positioning feature map data, and record the absolute position information of the high-precision positioned parking space obtained after parking is completed. The automatic parking system background turns on visual SLAM to establish a positioning feature map, and each parking trajectory is recorded in the positioning feature map.

[0028] When, according to the current position of the vehicle and the position of the parking space, the parking path planning fails, the matching historical parking trajectory can be searched in the positioning feature map according to the current position of the vehicle and the parking space information.

[0029] Among them, as an embodiment, before searching for the matching historical parking trajectory in the SLAM positioning feature map, it further includes: when the position of the vehicle itself is within a preset distance range from the position of the parking space, turn on the SLAM positioning feature map; search for the historical parking trajectory from the position of the vehicle itself to the position of the parking space in the SLAM positioning feature map, and the historical parking trajectory is recorded in the SLAM positioning map.

[0030] It can be understood that when searching for a matching historical parking trajectory in the positioning feature map, in order to save network resources and reduce the pressure on the server, the SLAM positioning feature map is only enabled when the vehicle position is within a certain distance range from the parking space position. For example, when the vehicle arrives within 500m of the parking space, the SLAM positioning feature map is enabled to search for a matching historical parking trajectory on the positioning feature map.

[0031] S2. Extract the key points of the parking path of the historical parking trajectory.

[0032] As an embodiment, the extraction of the key points of the parking path of the historical parking trajectory includes: extracting the starting point, ending point, gear shift point of the historical parking trajectory, and the entry point when the vehicle first enters the wireframe as the first set of feature points to be extracted; calculating the integral value of the curvature change between two adjacent first feature points in the first set of feature points. When the integral value is greater than the set integral threshold, for each set integral threshold between the two adjacent first feature points, record a second feature point; traverse each pair of adjacent first feature points to record the obtained second set of feature points; merge the first set of feature points and the second set of feature points to form the key points of the parking path.

[0033] It can be understood that when the automatic parking path planning fails, refer to Figure 5 , and according to the current vehicle position information and the parking space information, search for a matching historical parking trajectory in the positioning feature map data. When a matching historical parking trajectory is found, extract the key points of the parking path on the historical parking trajectory. The strategy for extracting the key points of the parking path is: record the historical parking trajectory, first extract the starting point, ending point, gear shift point, and the entry point into the basket when the vehicle first enters the wireframe of the historical parking trajectory as the first set of feature points to be extracted. Based on the first feature points in the first set of feature points, calculate the integral value of the curvature change between two adjacent first feature points in sequence. When the integral value is greater than the set integral threshold, extract a second feature point every set integral threshold. For example, when the integral value of the curvature change between two adjacent first feature points is 80 degrees and is greater than 15 degrees, record a second feature point every 15 degrees between the two first feature points, and a total of 5 feature points are recorded, and then the integral value is cleared. Calculate the integral value of the curvature change between the subsequent two adjacent first feature points again, and traverse each pair of adjacent first feature points in this way to record multiple second feature points to form a second set of feature points. Among them, the first set of feature points and the second set of feature points are merged to jointly constitute the key points of the extracted parking trajectory.

[0034] Among them, each key point of the parking trajectory will record the abscissa X, ordinate Y, and heading angle of the vehicle in the coordinate system of the parking space when the vehicle passes through the key point during the parking process to form the position information

[0035] S3. Re-plan a new parking path based on the extracted key points of the parking path.

[0036] As an embodiment, the re-planning of a new parking path based on the extracted key points of the parking path includes: based on the position information of the key points of the parking path, determining a starting point to be traced according to the real-time position information of the vehicle, and then tracing in the order of the key points of the parking path to form an approximation trajectory; generating a new parking path based on the approximation trajectory.

[0037] It can be understood that after extracting the key points of the parking trajectory from the historical parking trajectory, the parking path is re-planned according to the extracted key points of the parking trajectory to obtain a new parking path.

[0038] Specifically, according to the real-time position information of the vehicle, a starting point to be traced is determined among the key points of the parking path, and subsequent tracing points can be traced in the order of the original key points to form an approximation trajectory, and a new parking path is generated based on the approximation trajectory.

[0039] Among them, refer to Figure 6 , the position information of the key points of the parking path, determining a starting point to be traced according to the real-time position information of the vehicle, and then tracing in the order of the key points of the parking path to form an approximation trajectory, includes: making an extension line along the front-facing direction of the vehicle's current position; making perpendicular lines to the extension line at each key point; calculating the distance values from each perpendicular point to the extension line, and retaining the key points with distance values greater than a set distance value. For example, retain the key points with distance values greater than 0.5 m; calculate the characteristic distance P from each retained key point to the vehicle n ; determine minP n The corresponding retained key point is used as the starting point for tracing, and subsequent tracing is carried out in the order of the retained key points to form an approximation trajectory.

[0040] Among them, the calculation of the characteristic distance P from each retained key point to the vehicle n includes: calculating the perpendicular line length of each retained key point and recording it as Z n , where n is the retained key point number, and calculating for each retained key point Calculating the distance from the perpendicular point of each retained key point to the vehicle and recording it as J n , and calculating for each retained key point Calculating the heading angle of each retained key point and the difference from the heading angle of the vehicle's real-time position and calculating for each retained key point Calculating the characteristic distance P from each retained key point to the vehicle n =a*M n +b*K n+c*L n where a, b, and c are proportionality coefficients, which are calibrated based on the acceleration and steering performance of the vehicle during actual vehicle tests.

[0041] As an embodiment, generating a new parking path based on the approximated trajectory includes: extracting the starting point, shifting points, and ending point of the approximated trajectory; generating corresponding trajectory segments by using B-spline curve interpolation between the starting point and the first shifting point, between every two adjacent shifting points, and between the last shifting point and the ending point; and combining the multiple trajectory segments to form a new parking path.

[0042] It can be understood that after obtaining the approximated trajectory, extract the starting point, each shifting point, and the ending point of the approximated trajectory. Generate corresponding trajectory segments by using B-spline curve interpolation between the starting point and the first shifting point, between every two adjacent shifting points, and between the last shifting point and the ending point; and combine the multiple trajectory segments to form a new parking path. If there is no shifting point, generate a new parking path based on B-spline curve interpolation between the starting point and the ending point.

[0043] S4. Park the vehicle into the parking space based on the new parking path.

[0044] It can be understood that after successful path planning, the automatic parking controller automatically controls the vehicle to park into the parking space according to the replanned parking path.

[0045] Among them, this automatic parking process includes the following steps:

[0046] Step 1: Turn on the automatic parking system.

[0047] The driver turns on the automatic parking system and selects to park. Then the automatic parking system starts to search for a parking space. At the same time, the visual SLAM is started in the background of the automatic parking system, and a positioning feature map is established.

[0048] Step 2: Find a parking space.

[0049] The automatic parking system finds a parking space and displays the parking space information on the display screen of the in-vehicle infotainment system. The driver selects the parking space and clicks "Start Parking".

[0050] Step 3: Automatic parking path planning.

[0051] After the driver selects the parking space and clicks "Start Parking", the automatic parking system performs path planning. If the path planning is successful, the system executes automatic parking and deletes the positioning feature map established by this SLAM positioning. If the path planning fails, proceed to the next step.

[0052] Step 4: Extract key parking path points.

[0053] The driver parks the vehicle into the parking space by himself / herself, and the automatic parking system memorizes the parking trajectory of the driver, and extracts the key points of the parking path according to the trajectory of the driver parking into the parking space.

[0054] Step 5: When it is detected that the vehicle enters the vicinity of the memorized position, the background enables SLAM positioning.

[0055] Obtain the longitude and latitude information of high-precision positioning. When it is detected that the vehicle enters within 500 m of the memorized longitude and latitude position (x0, y0, z0), that is, when the absolute position information of the current vehicle is within the range of (x0 ± 500 m) && (y0 ± 500 m) && (z0 ± 50 m), the positioning module in the automatic parking controller enables visual SLAM positioning in the background, and the SLAM positioning module matches the surrounding environment of the vehicle.

[0056] Step 6: Enable the APA system to search for the previous parking space.

[0057] The driver enables the APA system and selects the parking mode, and the automatic parking system starts to search for the parking space. After the automatic parking system searches for the previously memorized parking space and the SLAM positioning module successfully matches the parking space, the driver clicks "Start Parking" to enter the next step.

[0058] Step 7: The automatic parking performs path replanning.

[0059] Among the key feature points of the extracted user trajectory, select the starting point of the trajectory to be traced according to the real-time position of the vehicle, and then sequentially trace the subsequent feature points to form an approximation trajectory.

[0060] Step 8: Generate a new parking path according to the approximation trajectory.

[0061] Step 9: Automatically control the vehicle to park into the parking space according to the new parking path.

[0062] Embodiment 2

[0063] A memory automatic parking system, see Figure 7 , the memory automatic parking system includes a path planning module 71 and a motion control module 72, wherein:

[0064] The path planning module 71 is configured to, during automatic parking, when the parking path planning fails according to the searched parking space and the position of the vehicle itself, search for a matching historical parking trajectory in the SLAM positioning feature map; extract the key points of the parking path of the historical parking trajectory; and, based on the extracted key points of the parking path, replan a new parking path;

[0065] The motion control module 72 is configured to park the vehicle into the parking space based on the new parking path.

[0066] It is understandable that multiple sensing modules can be installed on a vehicle. For example, surround-view cameras and ultrasonic radars. In the embodiments of the present invention, 4 surround-view cameras and 12 ultrasonic radars are installed on the vehicle. Among them, the 4 surround-view cameras are respectively arranged on the left outside rearview mirror, the right outside rearview mirror, the front bumper, and the rear bumper. For the 12 ultrasonic radars, 4 are arranged at intervals on the front bumper, 4 are arranged at intervals on the rear bumper, and 4 on the sides are respectively arranged at the left front, right front, left rear, and right rear of the vehicle body.

[0067] When the vehicle enters the parking lot, the surround-view cameras and ultrasonic radars are turned on to sense the parking space information and the surrounding environmental obstacle information. Based on the parking space information and the surrounding environmental obstacle information, it is determined whether the automatic parking path planning can succeed. If the automatic path planning fails, it is necessary to refer to the historical parking trajectory to re-plan a new parking path, and the vehicle is parked in the parking space according to the new parking path.

[0068] When the parking path planning fails, the path planning module 71 searches for a matching historical parking trajectory in the SLAM positioning feature map; extracts the parking path key points of the historical parking trajectory; and based on the extracted parking path key points, re-plans a new parking path.

[0069] The motion control module 72 performs lateral control and longitudinal control on the vehicle according to the re-planned new parking path, and parks the vehicle in the parking space.

[0070] It is understandable that a memory automatic parking system provided by the present invention corresponds to the memory automatic parking methods provided in the foregoing embodiments. The relevant technical features of the memory automatic parking system can refer to the relevant technical features of the memory automatic parking method, which will not be elaborated here.

[0071] A memory automatic parking method and parking system provided by the embodiments of the present invention. After the automatic parking system searches for a specified parking space, if the path planning of the automatic parking system fails, then after the driver parks the vehicle into the parking space by himself / herself, the automatic parking system memorizes the driver's parking path and extracts the key path points for the driver to automatically park into the parking space. The next time the automatic parking system is used, through the absolute position information of high-precision positioning and combined with visual SLAM relative positioning to the specified parking position, and when this specified parking space is searched, after the driver requests to activate the automatic parking system, the automatic parking system will re-plan and fit the automatic parking trajectory according to the key path points of the historical parking trajectory of the driver parking into the parking space, so as to realize automatic parking into the parking space.

[0072] Compared with traditional path planning algorithms, by memorizing the historical trajectories of the driver parking in the corresponding parking spaces, extracting the key path points of the driver's parking trajectory, and re-planning and fitting the automatic parking trajectory, the adaptability of the automatic parking system to the environment is greatly improved, the problem that the automatic parking system cannot cope with complex and changeable parking environments is solved, and customer satisfaction is enhanced.

[0073] It should be noted that in the above embodiments, the descriptions of the various embodiments have their own emphases. For parts not described in detail in a certain embodiment, reference may be made to the relevant descriptions of other embodiments.

[0074] Those skilled in the art should understand that the embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0075] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram, and the combination of flows and / or blocks in the flowchart and / or block diagram, can be realized by computer program instructions. These computer program instructions can be provided to the processors of general-purpose computers, special-purpose computers, embedded computers, or other programmable data processing devices to generate a machine, such that the instructions executed by the processors of the computer or other programmable data processing devices generate means for realizing the functions specified in Figure 1 one or more flows and / or blocks Figure 1 one or more blocks.

[0076] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including instruction means that realizes the functions specified in Figure 1 one or more flows and / or blocks Figure 1 one or more blocks.

[0077] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, so that the instructions executed on the computer or other programmable device provide means for realizing the functions specified in Figure 1 one or more flows and / or blocksFigure 1 Steps of the functions specified in one or more boxes.

[0078] Although the preferred embodiments of the present invention have been described, those skilled in the art can make additional changes and modifications to these embodiments once they learn the basic inventive concept. Therefore, the appended claims are intended to be construed to include the preferred embodiments as well as all changes and modifications that fall within the scope of the present invention.

[0079] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these modifications and variations.

Claims

1. A memory-based automatic parking method, characterized in that, it includes: During the automatic parking process, when the parking path planning fails according to the searched parking space and the position of the vehicle itself, search for a matching historical parking trajectory in the SLAM positioning feature map; Extract the key points of the parking path of the historical parking trajectory; Based on the extracted key points of the parking path, re-plan a new parking path; Based on the new parking path, park the vehicle into the parking space; The extraction of the key points of the parking path of the historical parking trajectory includes: Extract the starting point of the trajectory, the ending point of the trajectory, the gear shift point, and the entry point when the vehicle first enters the wireframe in the historical parking trajectory as the first set of extracted feature points; Calculate the integral value of the curvature change between two adjacent first feature points in the first set of feature points. When the integral value is greater than the set integral threshold, record a second feature point for each set integral threshold between the two adjacent first feature points; Traverse every two adjacent first feature points and record the obtained second set of feature points; Combine the first set of feature points and the second set of feature points to form the key points of the parking path.

2. The memory-based automatic parking method according to claim 1, characterized in that, Determine the success or failure result of the parking path planning from the position of the vehicle itself to the parking space according to the searched parking space information, the vehicle position information, and the obstacle information around the parking space.

3. The memory-based automatic parking method according to claim 1, characterized in that, Before searching for a matching historical parking trajectory in the SLAM positioning feature map, it further includes: When the distance between the vehicle position and the parking space position is within a preset distance range, turn on the SLAM positioning feature map; Search for the historical parking trajectory from the vehicle position to the parking space position in the SLAM positioning feature map, and the historical parking trajectory is recorded in the SLAM positioning feature map.

4. The memory-based automatic parking method according to claim 1, characterized in that, Record the position information of the key points of the parking path, and the position information includes the abscissa, ordinate, and heading angle of each key point of the parking path.

5. The memory-based automatic parking method according to claim 4, characterized in that, The re-planning of the new parking path based on the extracted key points of the parking path includes: Based on the position information of the key points of the parking path, determine the starting point to be tracked according to the real-time position information of the vehicle, and then follow the order of the key points of the parking path for tracking to form an approximation trajectory; Generate a new parking path based on the approximation trajectory.

6. The memory-based automatic parking method according to claim 5, characterized in that, The determination of the starting point to be tracked according to the real-time position information of the vehicle based on the position information of the key points of the parking path, and then following the order of the key points of the parking path for tracking to form an approximation trajectory includes: Make an extension line along the head orientation direction of the current vehicle position; Make perpendicular lines to the extension line at each key point; Calculate the distance value from each perpendicular point to the extension line, and retain the key points with a distance value greater than the set distance value; Calculate the characteristic distance P from each reserved key point to the vehicle n ; Determine minP n The corresponding reserved key points are used as the starting point for tracking, and subsequent tracking is carried out in the order of the reserved key points to form an approximation trajectory.

7. The memory-based automatic parking method according to claim 6, characterized in that, Calculating the characteristic distance P from each retained key point to the vehicle n , including: Record the perpendicular lengths of each remaining key point as Z n , where n is the remaining key point number, and calculate the Calculate the distance from the foot of the perpendicular of each reserved key point to the vehicle and denote it as J n and calculate for each reserved key point Calculate the heading angles of each remaining key point and the heading angle of the real-time position of the vehicle difference and calculate the Calculate the characteristic distance P from each remaining key point to the vehicle n = a * M n + b * K n + c * L n , where a, b, and c are proportionality coefficients, which are calibrated based on the acceleration and steering performance of the vehicle in actual vehicles.

8. The memory-based automatic parking method according to claim 5 or 6 or 7, It is characterized in that generating a new parking path based on the approximation trajectory, including: extracting the trajectory start point, shift points, and trajectory end point in the approximation trajectory; adopting B-spline curve interpolation to generate corresponding trajectory segments between the trajectory start point and the first shift point, between every two adjacent shift points, and between the last shift point and the trajectory end point; combining multiple trajectory segments to form a new parking path.

9. A memory automatic parking system It is characterized in that including: a path planning module, configured to, during automatic parking, when the parking path planning fails according to the searched parking space and the position of the vehicle itself, search for a matching historical parking trajectory in the SLAM positioning feature map; extract the parking path key points of the historical parking trajectory; and re-plan a new parking path based on the extracted parking path key points; a motion control module, configured to park the vehicle into the parking space based on the new parking path; the extracting the parking path key points of the historical parking trajectory includes: extracting the trajectory start point, trajectory end point, shift points, and the frame entry point when the vehicle first enters the wireframe in the historical parking trajectory as the first set of feature points to be extracted; calculating the integral value of the curvature change between every two adjacent first feature points in the first set of feature points, and when the integral value is greater than the set integral threshold, recording a second feature point for every set integral threshold between the two adjacent first feature points; traversing every two adjacent first feature points and recording the obtained second set of feature points; combining the first set of feature points and the second set of feature points to form the parking path key points.

Images