Point model boundary constraint construction method, path planning method and corresponding device

Abstract

The invention discloses a point model boundary constraint construction method, a path planning method and a corresponding device. The method comprises the following steps: constructing a lane local map based on lane information and obstacle information, determining an optimal passing topological space according to an occupied area and a passable area in the lane local map, and obtaining boundary point sequence pairs distributed on two sides in the optimal passing topological space; then, dividing the boundary point sequence pair into at least one first-class boundary interval and at least one second-class boundary interval, performing interpolation processing on the first-class boundary interval and the second-class boundary interval which are obtained through division, and obtaining a first-class boundary interval and a second-class boundary interval on the basis of a sequence pair obtained through interpolation processing and a first boundary point sequence pair and a tail boundary point sequence pair of each boundary interval; and constructing a point model boundary constraint. Therefore, boundary anomalies caused by obstacles are reflected by dividing different types of boundary intervals, and accurate, safe and reliable boundary constraints can be determined for path planning under the condition of not influencing computational power expenditure.

Description

Technical Field

[0001] The present application relates to the field of computer technology, specifically to technical fields such as intelligent transportation and autonomous driving, and in particular to a method for constructing point model boundary constraints, a path planning method and corresponding devices. Background Art

[0002] In the field of autonomous driving technology, autonomous driving planning is generally divided into lateral path planning and longitudinal speed planning. Among them, lateral path planning is responsible for generating a spatial path that does not contain time information to avoid and detour static obstacles; longitudinal speed planning considers time information on the basis of the spatial path and generates a space-time trajectory that can avoid dynamic obstacles.

[0003] In the existing lateral path planning scheme, the key to ensure accurate path planning is to build reliable boundary constraints to guide the construction of the optimized path. At present, the lateral path planning problem is usually constrained by a point model in the Frenet coordinate system; the s axis is sampled in the vertical direction to obtain a discrete sequence of path point sets s to be optimized, the coordinates of the path point set s sequence are determined, and the left and right l boundaries corresponding to each path point are obtained according to prior information, which are used as the optimization boundary constraints for graph search or optimization.

[0004] However, when constructing boundary constraints, the unreasonable sampling of the path point set often increases the computing power cost (intensive sampling) or affects the accuracy of the boundary constraints (sparse sampling). Summary of the invention

[0005] The present application provides a method for constructing a point model boundary constraint, a path planning method and a corresponding device to solve the problem of unreasonable sampling of a path point set, which increases computing power expenditure or affects the accuracy of boundary constraints. By constructing boundary constraints through a point model, points with abnormal changes in the boundary of the passage space are used as boundary constraint conditions. This can determine more accurate, safe and reliable boundary constraints for path planning without affecting computing power expenditure.

[0006] The technical solution is as follows:

[0007] In a first aspect, a method for constructing a point model boundary constraint is provided, comprising:

[0008] Constructing a lane local map based on the lane information and the obstacle information, wherein the area where the obstacle is located in the lane local map is the occupied area, and the other areas are the passable areas;

[0009] Determine an optimal traffic topology space according to the occupied area and the passable area in the lane local map, and obtain boundary point sequence pairs distributed on both sides of the optimal traffic topology space;

[0010] Dividing the boundary point sequence pairs into at least one first-class boundary interval and at least one second-class boundary interval; wherein the first-class boundary interval contains boundary point sequence pairs corresponding to obstacles, and the second-class boundary interval only contains boundary point sequence pairs corresponding to lanes;

[0011] Interpolation processing is performed on the first-category boundary interval and the second-category boundary interval obtained by the division, respectively, and point model boundary constraints are constructed based on the sequence pairs obtained by the interpolation processing and the first and last boundary point sequence pairs of each boundary interval.

[0012] In a possible implementation, a lane local map is constructed based on lane information and obstacle information, specifically including:

[0013] Based on the map and perception, the lane information and obstacle information within the current planned driving range are obtained respectively;

[0014] Preprocessing the obstacles contained in the obstacle information according to a preset processing rule;

[0015] Determining the position information of the obstacle obtained through preprocessing based on the boundary coordinates in the obstacle information;

[0016] A lane local map is constructed based on the road boundary, the current planned driving range, and the location information of obstacles in the lane information.

[0017] In a possible implementation, dividing the boundary point sequence pair into at least one first-type boundary interval and at least one second-type boundary interval specifically includes:

[0018] Performing boundary shrinkage processing on the boundary point sequence pairs based on a set shrinkage strategy, and truncating sequence pairs with shrinkage abnormalities;

[0019] Dividing the boundary sequence pairs retained after the shrinkage process into at least one first-class boundary interval and at least one second-class boundary interval according to the boundary type;

[0020] For each first-category boundary interval, search for N boundary point sequence pairs with abnormal changes in the boundary of the passage space, and divide the first-category boundary interval into N+1 first-category boundary sub-intervals according to the N boundary point sequence pairs; and / or; for each second-category boundary interval, search for M boundary point sequence pairs with abnormal changes in the boundary of the passage space, and divide the first-category boundary interval into M+1 first-category boundary sub-intervals according to the M boundary point sequence pairs; wherein the N boundary point sequences are not the first and last boundary sequence pairs of the first-category boundary interval, and the M boundary point sequence pairs are not the first and last boundary sequence pairs of the second-category boundary interval; and both N and M are positive integers.

[0021] In a possible implementation, for each first-class boundary interval, N boundary point sequence pairs with abnormal changes in the boundary of the passage space are found, and the first-class boundary interval is divided into N+1 first-class boundary sub-intervals according to the N boundary point sequence pairs, specifically including:

[0022] For each first-class boundary interval, execute:

[0023] Traversing each boundary point sequence pair in the first type boundary interval;

[0024] Based on the coordinate information of the boundary point sequence pair, it is determined in turn whether each boundary point sequence pair satisfies at least one of the following first-category constraints: the lane width corresponding to the current boundary point sequence pair is less than a first threshold, and the absolute value of the difference between the lane width corresponding to the current boundary point sequence pair and the lane width corresponding to the previous boundary point sequence pair is less than a second threshold; the left boundary and / or the right boundary of the lane corresponding to the current boundary point sequence pair is discontinuous;

[0025] If there is a boundary point sequence pair that satisfies at least one of the first-type constraint conditions, it is determined that the boundary point sequence pair can cause an abnormal change in the boundary of the passage space, and N boundary point sequence pairs that satisfy the constraint conditions are counted; based on the first and last boundary point sequence pairs of the first-type boundary interval and the N boundary point sequence pairs, the first-type boundary interval is divided into N+1 first-type boundary sub-intervals;

[0026] If none of the boundary point sequence pairs satisfy the at least one first-type constraint condition, no processing is performed.

[0027] In a possible implementation, for each second-class boundary interval, M boundary point sequence pairs with abnormal changes in the boundary of the passage space are found, and the second-class boundary interval is divided into M+1 second-class boundary sub-intervals according to the M boundary point sequence pairs, specifically including:

[0028] For each second-class boundary interval, execute:

[0029] Determine whether the number of boundary point sequence pairs included in the second type boundary interval is less than a third threshold;

[0030] If the result of the judgment is that the number of boundary point sequence pairs is less than the third threshold, no processing is performed;

[0031] If the judgment result is that the number of boundary point sequence pairs is not less than the third threshold, then traverse each boundary point sequence pair in the second-type boundary interval, and based on the coordinate information of the boundary point sequence pair, judge whether each boundary point sequence pair satisfies at least one of the following second-type constraint conditions: the absolute value of the difference between the lane width corresponding to the current boundary point sequence pair and the lane width corresponding to the previous boundary point sequence pair is less than the fourth threshold; the left boundary and / or the right boundary of the lane corresponding to the current boundary point sequence pair are discontinuous;

[0032] If there is a boundary point sequence pair that satisfies the at least one second-type constraint condition, it is determined that the boundary point sequence pair can cause an abnormal change in the boundary of the passage space, and M boundary point sequence pairs that satisfy the constraint condition are counted; based on the first and last boundary point sequence pairs of the second-type boundary interval and the M boundary point sequence pairs, the second-type boundary interval is divided into M+1 second-type boundary sub-intervals;

[0033] If none of the boundary point sequence pairs satisfy the at least one second type constraint condition, no processing is performed.

[0034] In a possible implementation, interpolation processing is performed on the first-category boundary interval and the second-category boundary interval obtained by division, specifically including:

[0035] For a plurality of first-class boundary subintervals obtained by dividing the first-class boundary interval and the first-class boundary interval that has not been divided, interpolation processing is performed using a first interpolation method;

[0036] The second interpolation method is used to perform interpolation processing on a plurality of second-type boundary sub-intervals obtained by dividing the second-type boundary interval and the second-type boundary interval that has not been divided.

[0037] In a second aspect, a path planning method is provided, comprising:

[0038] Acquire lane information and obstacle information, and determine boundary constraints; wherein the boundary constraints are determined by the method for constructing point model boundary constraints described in the first aspect;

[0039] Based on the boundary constraints, a graph structure is constructed, and a preliminary path is determined by using a graph search method;

[0040] Based on the boundary constraints and the preliminary path, a precise path is optimized.

[0041] In a third aspect, a device for constructing a point model boundary constraint is provided, comprising:

[0042] A first construction module is used to construct a lane local map based on the lane information and the obstacle information, wherein the area where the obstacle is located in the lane local map is an occupied area, and other areas are passable areas;

[0043] A determination module, used to determine an optimal pass topological space according to the occupied area and the passable area in the lane local map, and obtain boundary point sequence pairs distributed on both sides of the optimal pass topological space;

[0044] A division module, used to divide the boundary point sequence pairs into at least one first-type boundary interval and at least one second-type boundary interval; wherein the first-type boundary interval contains boundary point sequence pairs corresponding to obstacles, and the second-type boundary interval only contains boundary point sequence pairs corresponding to lanes;

[0045] The second construction module is used to perform interpolation processing on the first type boundary interval and the second type boundary interval obtained by division, and construct point model boundary constraints based on the sequence pairs obtained by the interpolation processing and the first and last boundary point sequence pairs of each boundary interval.

[0046] In a fourth aspect, a path planning device is provided, comprising:

[0047] An acquisition module, used to acquire lane information and obstacle information, and determine boundary constraints; wherein the boundary constraints are determined by the construction method of the point model boundary constraints described in the first aspect;

[0048] A determination module, configured to construct a graph structure based on the boundary constraints and determine a preliminary path using a graph search method;

[0049] The optimization module is used to optimize and obtain a precise path based on the boundary constraints and the preliminary path.

[0050] In a fifth aspect, a computer-readable storage medium is provided, wherein the storage medium stores at least one instruction, and the at least one instruction is loaded and executed by a processor to implement the method of the above-mentioned aspects and any possible implementation manner.

[0051] In a sixth aspect, an electronic device is provided, including:

[0052] at least one processor; and

[0053] a memory communicatively connected to the at least one processor; wherein,

[0054] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method of any possible implementation manner and the aspects described above.

[0055] In a seventh aspect, an autonomous driving vehicle is provided, comprising the electronic device as described above.

[0056] The beneficial effects of the technical solution provided by this application include at least:

[0057] It can be seen from the above technical solution that a lane local map is constructed based on lane information and obstacle information, and the optimal traffic topological space is determined according to the occupied area and passable area in the constructed lane local map, and the boundary point sequence pairs distributed on both sides in the optimal traffic topological space are obtained; then, the boundary point sequence pairs are divided into at least one first-class boundary interval and at least one second-class boundary interval, and the first-class boundary interval and the second-class boundary interval obtained by the division are interpolated respectively, and the point model boundary constraints are constructed based on the sequence pairs obtained by the interpolation processing and the first and last boundary point sequence pairs of each boundary interval. Since the first-class boundary interval contains the boundary point sequence pairs corresponding to the obstacles, and the second-class boundary interval only contains the boundary point sequence pairs corresponding to the lanes, in this way, the boundary anomalies caused by the obstacles can be reflected by dividing the boundary intervals of different categories, and the first and last boundary point sequence pairs of the boundary interval are extracted, and the point model boundary constraints are constructed by combining the sequence pairs obtained by the interpolation processing, so that the key point sequence pairs of the abnormal changes in the boundary of the traffic space can be screened out to ensure the accuracy and reliability of the boundary constraints, and will not affect the computing power expenditure of the subsequent path planning.

[0058] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0060] Figure 1 A schematic diagram of the steps of a method for constructing a point model boundary constraint provided in this application;

[0061] Figure 2a Schematic diagrams of several optimal transit topological spaces provided for this application;

[0062] Figure 2b A schematic diagram of the optimal travel topological space for dividing the first type of boundary interval and the second type of boundary interval in this application;

[0063] Figure 3 A schematic diagram of the steps of a path planning method provided in this application;

[0064] Figure 4 A schematic flow chart of a method for constructing point model boundary constraints provided in another embodiment of the present application;

[0065] Figure 5 It is a structural block diagram of a device for constructing point model boundary constraints provided by an embodiment of the present application;

[0066] Figure 6 is a structural block diagram of a path planning device provided by an embodiment of the present application;

[0067] Figure 7 It is a block diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0068] The following is a description of exemplary embodiments of the present application in conjunction with the accompanying drawings, including various details of the embodiments of the present application to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present application. Similarly, for the sake of clarity and conciseness, the description of well-known functions and structures is omitted in the following description.

[0069] Obviously, the described embodiments are only part of the embodiments of the present application, but not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present application.

[0070] It should be noted that the terminal devices involved in the embodiments of the present application may include but are not limited to mobile phones, personal digital assistants (PDAs), wireless handheld devices, tablet computers and other smart devices; display devices may include but are not limited to personal computers, televisions and other devices with display functions.

[0071] In addition, the term "and / or" in this article is only a description of the association relationship between the associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0072] In the path planning scheme, especially for lateral path planning, the unreasonable sampling of the path point set affects the path planning. For example, if the sampling is too dense, it will increase the computing power in the path planning process; if the sampling is too sparse, the path planning will be affected due to inaccurate boundary constraints. It can be seen that it is very important to construct accurate, reasonable and safe boundary constraints.

[0073] To this end, the present application proposes a boundary constraint construction scheme based on a point model, and the inventive concept is: according to the occupied area and the passable area in the constructed lane local map, the optimal pass topological space is determined, and the boundary point sequence pairs distributed on both sides in the optimal pass topological space are obtained; then, the boundary point sequence pairs are divided into at least one first-class boundary interval and at least one second-class boundary interval, and the first-class boundary interval and the second-class boundary interval obtained by the division are interpolated respectively, and the point model boundary constraint is constructed based on the sequence pairs obtained by the interpolation processing and the first and last boundary point sequence pairs of each boundary interval. Since the first-class boundary interval contains the boundary point sequence pairs corresponding to the obstacle, and the second-class boundary interval only contains the boundary point sequence pairs corresponding to the lane, in this way, the boundary anomaly caused by the obstacle can be reflected by dividing the boundary intervals of different categories, and the first and last boundary point sequence pairs of the boundary interval are extracted, and the point model boundary constraint is constructed by combining the sequence pairs obtained by the interpolation processing, so that the key point sequence pairs of the abnormal changes in the boundary of the pass space can be screened out to ensure the accuracy and reliability of the boundary constraint conditions, and will not affect the computing power expenditure of the subsequent path planning. In short, more reasonable and safe boundary constraints can be constructed to facilitate the efficient implementation of path planning.

[0074] Reference Figure 1 As shown, it is a schematic diagram of the steps of a method for constructing a point model boundary constraint provided in the present application. The execution subject of the construction method is a construction device for the point model boundary constraint. In the present application, the construction device for the point model boundary constraint can be a software module with computing and data processing capabilities, or an electronic device integrated with a similar software module. Optionally, the execution subject of the construction device for the point model boundary constraint in the present application can be an autonomous driving system server, an autonomous driving vehicle, or a related processing module integrated in the autonomous driving vehicle.

[0075] like Figure 1 As shown, the method for constructing the point model boundary constraint may include the following steps:

[0076] Step 102: construct a lane local map based on the lane information and the obstacle information, wherein the area where the obstacle is located in the lane local map is the occupied area, and the other areas are the passable areas.

[0077] In this application, when constructing a lane local map based on lane information and obstacle information, the following steps may be specifically included:

[0078] In the first step, the lane information and obstacle information within the current planned driving range are obtained based on the map and perception.

[0079] In the present application, lane information can be obtained from map data, and obstacle information can be obtained through perception. For example, map data contains spatial distribution location data, graphic features and geographic attribute data of multiple types of map elements, among which multiple types of map elements may include: lanes, residential areas, hydrology, vegetation, etc. According to the business needs of this application, the lane information of the driving scene where the autonomous driving vehicle is located is screened and extracted. Usually, autonomous driving vehicles are equipped with various sensors to obtain environmental information, such as the most common lidar and cameras, to sense whether there are vehicles, pedestrians, or other obstacles around. In addition, corresponding sensors will be configured on the roadside or directly above the lane in the driving scene to obtain driving scene information and the driving condition of the driving vehicle. Therefore, lane information can be obtained through map data, and obstacle information can be obtained through perception data.

[0080] Step 2: pre-process the obstacles contained in the obstacle information according to preset processing rules.

[0081] However, considering that the range of obstacles covered in the perception data is relatively large, some obstacles may not be required this time. Therefore, in order to reduce processing costs and time and ensure the accuracy of building a local lane map, the acquired obstacle information can be preprocessed to obtain accurate obstacle information.

[0082] In fact, when obtaining lane information, it is necessary to select according to the lane where the autonomous driving vehicle is located. Specifically, you can choose the lane where the autonomous driving vehicle is located or a nearby lane to roughly determine the lane boundary. At the same time, you can also set a longitudinal driving planning range according to the driving plan.

[0083] In this application, the acquired obstacle information is preprocessed, and the preprocessing rules can be flexibly set according to business needs. The preprocessing rules may include: ignoring obstacles outside the lane boundary, ignoring obstacles outside the planning range, ignoring obstacles with special custom attributes, ignoring special decision obstacles, ignoring virtual obstacles, and special processing of obstacle polygons (densification, cutting, expansion, etc.).

[0084] Ignoring obstacles outside the lane boundary means not considering irrelevant obstacles outside the lane boundary. This is because there may be multiple obstacles sensed, both inside and outside the lane. Obviously, for an autonomous vehicle, obstacles outside the lane boundary generally do not affect its driving plan, so obstacles outside the lane boundary can be removed by preprocessing.

[0085] Ignoring obstacles outside the planning range means not considering irrelevant obstacles outside the planning range. The planning range is the driving range from the autonomous driving vehicle, that is, the position of the vehicle, to the set distance in the forward direction. Therefore, obstacles behind the autonomous driving vehicle and outside the set distance in the forward direction will not affect its driving plan. Therefore, obstacles outside the planning range can be removed by preprocessing.

[0086] Ignoring obstacles with special custom attributes refers to pre-processing rules that are customized according to the current business needs or user needs to eliminate irrelevant obstacles. This application does not limit this special custom attribute and can be set flexibly. Ignoring special decision obstacles can be: for example, before the boundary constraint is constructed, a related algorithm has made a decision on whether to avoid a certain type of obstacle. The decision can be marked for use in subsequent obstacle identification. Ignoring virtual obstacles refers to obstacles that may not exist in the lane but will hinder the passage of autonomous driving vehicles, such as traffic lights, or road surface signs for lane changes, etc. Such obstacles can be ignored.

[0087] In addition, it is also necessary to consider that the obstacle information obtained by perception is incomplete. Therefore, according to the category of the obstacle or the needs of the user, the obstacle can also be subjected to special polygon processing, such as at least one of densification processing, cutting processing, and expansion processing. Among them, the densification processing is mainly aimed at larger obstacles, such as trucks, etc., and the details of the truck can be completed based on the incomplete truck information obtained. For example, the truck can be quadrilateralized and the details of the upper left, lower left, upper right and lower right of the truck can be completed, so that the truck information is more accurate and true. The cutting processing can cut a polygon into two or more polygons for easy processing. The expansion processing can be used for smaller vehicles, but such vehicles may be loaded with cargo that exceeds their volume. At this time, for safety reasons, the vehicle can be expanded so that its expanded size is larger than the actual size.

[0088] After the above series of preprocessing, the obstacle information after preprocessing can be saved. Specifically, the present application can save the preprocessed obstacles (including obstacle information) to an obstacle container, so that it can be conveniently called from the obstacle container for subsequent use. It should be understood that the present application does not limit the method or carrier for saving obstacle information. Saving obstacles in an obstacle container is only a specific example, and other saving methods are also possible.

[0089] Step 3: Determine the position information of the obstacle obtained through preprocessing based on the boundary coordinates in the obstacle information.

[0090] Specifically, the aabox of the obstacle can be obtained, where the aabox can be an axis-aligned box that specifies the center position and size of the collision volume, which can be regarded as the size of the obstacle. Then, the horizontal and vertical coordinates of the boundary of the obstacle are determined according to the size information of the obstacle, and then the position occupied by the obstacle is determined.

[0091] It should be understood that in step 3 here, when determining the position information of the obstacle, all obstacles in the obstacle container may be traversed to determine the occupied position of each obstacle in turn.

[0092] Step 4: construct a lane local map based on the road boundary, the current planned driving range and the location information of obstacles in the lane information.

[0093] In specific implementation, a frenet coordinate system can be established, and a lane local map can be constructed based on the road boundary, the current planned driving range, and the position occupied by obstacles. In the frenet coordinate system, obstacles are represented by aabox. Based on the aabox outer contour of the obstacle, the lane local map can be divided into grids of different sizes. The grid where the obstacle is located is the occupied area, and the non-obstacle grid is the passable area.

[0094] In fact, in the present application scheme, the above four steps may not be used to construct a lane local map. Instead, the lane information and obstacle information within the current planned driving range may be obtained through mapping and perception respectively, and then the obstacle position information is determined based on the obstacle information. The lane local map is constructed based on the obstacle position information and the road boundary and the current planned driving range in the lane information.

[0095] Step 104: Determine an optimal traffic topology space according to the occupied area and the passable area in the lane local map, and obtain boundary point sequence pairs distributed on both sides of the optimal traffic topology space.

[0096] In the present application, the topological space can be selected based on the lane local map; the specific selection method may include:

[0097] Method 1: Design the cost and weight according to the size and relative relationship of the graph structure, and use the graph search algorithm to obtain the optimal pass topological space;

[0098] Method 2: Based on the forward deduction of the vehicle's kinematics, the motion cost is designed, and the space where the deduced optimal path is located is used as the optimal travel topological space.

[0099] The grid obtained by the above division can be used as a graph structure to select the optimal pass topological space.

[0100] Reference Figure 2aAs shown in FIG. 1 , it is a schematic diagram of several optimal pass topological spaces provided by the present application. Among them, the shaded part is the occupied area where the obstacle is located, and the other areas are passable areas. As shown by the dotted arrow, it is the topological space path that the vehicle can pass. According to method 1 or method 2, the optimal pass topological space can be determined, as shown by the solid arrow.

[0101] It should be understood that in addition to determining the optimal transit topological space by the above-mentioned method 1 or method 2, other feasible methods may also be used, which are not listed here one by one.

[0102] In the scheme of the present application, when obtaining the boundary point sequence pairs distributed on both sides of the optimal pass topological space, sampling can be performed vertically according to the grid based on the optimal pass topological space to collect boundary point sequence pairs distributed in the optimal pass topological space, that is, the boundary point Pair sequence.

[0103] Each boundary point sequence pair stores the l coordinates of the corresponding left and right boundary points, and marks whether the left and right boundary points are road boundary points or lane boundary points, and whether there are obstacles on the left or right side of the left and right boundary points and the types of obstacles. Moreover, the boundary point sequence pairs are collected in sequence, generally speaking, they are sampled in the longitudinal direction in the order from the vehicle position forward.

[0104] Step 106: Divide the boundary point sequence pairs into at least one first-category boundary interval and at least one second-category boundary interval; wherein the first-category boundary interval contains boundary point sequence pairs corresponding to obstacles, and the second-category boundary interval only contains boundary point sequence pairs corresponding to lanes.

[0105] It should be noted that the boundaries involved in this application refer to the left and right boundaries of the road or lane, not the front and back boundaries.

[0106] Considering the optimal traffic topology space obtained, although it is all lane space, due to the existence of obstacles, some lane sections in the optimal traffic topology space correspond to obstacles (there are obstacles on the left or right side of the lane boundary), and some lane sections have no obstacles. Therefore, it can be regarded as an abnormal change in the traffic space boundary, that is, the lane boundary changes from the first-class boundary corresponding to the obstacle to the second-class boundary without obstacles, or from the second-class boundary without obstacles to the first-class boundary corresponding to obstacles.

[0107] It should be understood that in the present application, the number of the first-class boundary intervals and the second-class boundary intervals obtained by division is not limited. In a driving scene where obstacles are determined to exist, there is at least one first-class boundary interval and at least one second-class boundary interval. Figure 2bAs shown, it is a schematic diagram of the optimal traffic topological space for dividing the first type of boundary interval and the second type of boundary interval in this application. If there is an obstacle in the driving scene and it is in front of the autonomous driving vehicle, then, in general, one first type of boundary interval and two second type of boundary intervals are divided. If there are two obstacles in the driving scene that are a certain distance apart and in front of the autonomous driving vehicle, then, generally, two first type of boundary intervals and three second type of boundary intervals can be divided. The number of the first type of boundary intervals and the second type of boundary intervals here is only for example and is not limited to this.

[0108] Step 108: interpolate the first-category boundary interval and the second-category boundary interval obtained by the division respectively, and construct a point model boundary constraint based on the sequence pairs obtained by the interpolation process and the first and last boundary point sequence pairs of each boundary interval.

[0109] After the first-class boundary interval and the second-class boundary interval are divided, it is equivalent to determining the key point where the boundary of the passage space has abnormal changes, that is, the connection point between the first-class boundary interval and the second-class boundary interval, which can be the minimum or maximum value of the first-class boundary interval, or the minimum or maximum value of the second-class boundary interval, which is not limited here. In short, the first and last boundary point sequence pairs of each boundary interval can be extracted by interpolating the first and second boundary intervals respectively. Among them, the interpolation processing can be based on business needs or user-defined distances (the distance here can be understood as the number of boundary point sequence pairs), and then a point is interpolated every custom distance in the first and second boundary intervals, and the interpolated points are extracted as interpolated sequence pairs. After processing according to the interpolation method, multiple interpolated sequence pairs are obtained, and together with the first and last boundary point sequence pairs of each boundary interval, the point model boundary constraints are constructed.

[0110] In the present application, the distance of the first type of boundary interval interpolation and the distance of the second type of boundary interval interpolation can be the same or different. For example, the distance of the first type of boundary interval interpolation is 5 boundary points, and the distance of the second type of boundary interval interpolation is 5 boundary points; for another example, the distance of the first type of boundary interval interpolation is 5 boundary points, and the distance of the second type of boundary interval interpolation is 10 boundary points.

[0111] Therefore, through the above scheme, the optimal traffic topology space is determined in the constructed lane local map, and then the boundary point sequence pairs in the optimal traffic topology space are obtained. The present application does not directly construct boundary constraints based on the boundary point sequence pairs, but divides the boundary point sequence pairs into at least one first-class boundary interval and at least one second-class boundary interval. Thus, by dividing different boundary intervals, the key points reflecting the abnormal boundary changes (the lane boundary changes from the corresponding no obstacle to the corresponding obstacle, or the lane boundary changes from the corresponding obstacle to the corresponding no obstacle) are screened out. In this way, the key information and overall information of the lane can be better reflected, so as to facilitate the construction of real and reliable boundary constraints. At the same time, it can also avoid excessive sampling, save example overhead and processing time.

[0112] Optionally, in the solution of the present application, when dividing the boundary point sequence pair into at least one first-type boundary interval and at least one second-type boundary interval, the following steps may be specifically included:

[0113] In the first step, the boundary point sequence pairs are subjected to boundary shrinkage processing based on the set shrinkage strategy, and the sequence pairs with shrinkage abnormalities are truncated.

[0114] In the present application scheme, after obtaining the boundary point sequence pair, it is equivalent to determining the left and right boundaries. When considering that the autonomous driving vehicle passes through these lanes, the driving safety may be affected by the different boundary types. Therefore, the corresponding contraction strategy can be appropriately adopted according to the characteristics of the boundary, and the optimal passage space can be contracted according to the corresponding contraction strategy. For example, for the curb, the curb boundary can be appropriately contracted according to visual perception to avoid the autonomous driving vehicle from colliding with the curb. For another example, some boundaries can be appropriately contracted by half the width of the autonomous driving vehicle, so that the autonomous driving vehicle can pass safely. For another example, some lanes have different curvatures and may also need to be contracted by different distances, which needs to be set according to the curvature requirements of the lanes. For another example, some lanes need to be classified and discussed according to different types of obstacles. For example, when the obstacle is a vehicle, it can be contracted by 20 centimeters, and when the obstacle is a person, it can be contracted by 5 centimeters, or more. This requires flexible settings according to the type of obstacle and the type of autonomous driving vehicle.

[0115] It should be understood that the contraction here refers to contraction towards the boundary and the inside of the lane. The contraction distance of each type of contraction strategy can be flexibly set to ensure that the autonomous driving vehicle can stay away from obstacles and boundaries, thereby improving driving safety.

[0116] In the second step, the boundary sequence pairs retained after the shrinkage process are divided into at least one first-category boundary interval and at least one second-category boundary interval according to the boundary type.

[0117] The specific division method of this step can refer to the above step 106 and will not be described in detail here.

[0118] The third step is to perform the same or different processing on the boundary intervals obtained by division according to different categories.

[0119] Specifically, it can be divided into three processing methods:

[0120] The first processing method: for each first-class boundary interval, find N boundary point sequence pairs with abnormal changes in the boundary of the passage space, and divide the first-class boundary interval into N+1 first-class boundary sub-intervals according to the N boundary point sequence pairs; the second-class boundary interval is not processed.

[0121] The second processing method: for each second-class boundary interval, find M boundary point sequence pairs with abnormal changes in the boundary of the passage space, and divide the first-class boundary interval into M+1 first-class boundary sub-intervals according to the M boundary point sequence pairs; the first-class boundary interval is not processed.

[0122] The third processing method: for each first-class boundary interval, find N boundary point sequence pairs with abnormal changes in the boundary of the passage space, and divide the first-class boundary interval into N+1 first-class boundary sub-intervals according to the N boundary point sequence pairs; for each second-class boundary interval, find M boundary point sequence pairs with abnormal changes in the boundary of the passage space, and divide the first-class boundary interval into M+1 first-class boundary sub-intervals according to the M boundary point sequence pairs.

[0123] Among them, the N boundary point sequences are not the first and last boundary sequence pairs of the first type of boundary interval, and the M boundary point sequence pairs are not the first and last boundary sequence pairs of the second type of boundary interval; and N and M are both positive integers.

[0124] The reason for further dividing the first type of boundary interval and / or the second type of boundary interval is that there may be abnormal boundary changes within each boundary interval, for example, the lane width becomes narrower and undergoes a large mutation, or the boundary on one side of the lane is discontinuous, or other abnormal boundary conditions. In order to accurately collect boundary point sequence pairs that can reflect boundary characteristics without excessive sampling, sequence pairs of key points with abnormal boundary changes can be further extracted within the boundary interval.

[0125] Optionally, in the solution of the present application, for each first-class boundary interval, N boundary point sequence pairs with abnormal changes in the boundary of the passage space are searched, and the first-class boundary interval is divided into N+1 first-class boundary sub-intervals according to the N boundary point sequence pairs, specifically including: for each first-class boundary interval, respectively execute: traverse each boundary point sequence pair in the first-class boundary interval; based on the coordinate information of the boundary point sequence pair, determine in turn whether each boundary point sequence pair satisfies at least one of the following first-class constraint conditions: the lane width corresponding to the current boundary point sequence pair is less than the first threshold, and the lane width corresponding to the current boundary point sequence pair is the same as the previous one The absolute value of the difference between lane widths corresponding to a boundary point sequence pair is less than a second threshold value; the left lane boundary and / or the right lane boundary corresponding to the current boundary point sequence pair are discontinuous; if there is a boundary point sequence pair that satisfies the at least one first-category constraint condition, it is determined that the boundary point sequence pair can cause abnormal changes in the boundary of the passage space, and N boundary point sequence pairs that satisfy the constraint conditions are counted; based on the first and last boundary point sequence pairs of the first-category boundary interval and the N boundary point sequence pairs, the first-category boundary interval is divided into N+1 first-category boundary sub-intervals; if none of the boundary point sequence pairs satisfy the at least one first-category constraint condition, no processing is performed.

[0126] For example, if the current driving scene is divided into the first type boundary interval and the second type boundary interval, assuming that there are 100 pairs of boundary point sequences, that is, the numbers are marked from 1 to 100. Among them, the first type boundary intervals are [26-50] and [86-100], and the second type boundary intervals are [1-25] and [51-85].

[0127] When processing the first type of boundary interval [26-50], start from the boundary point sequence pair numbered 26 and traverse each boundary point sequence pair. If a boundary point sequence pair, such as 29, corresponds to a boundary point sequence pair, determine whether the sequence pair meets one or all of the following conditions:

[0128] Condition 1: The lane width of the boundary point sequence pair is less than the first threshold. The first threshold here can be set according to actual business needs. This application takes half the vehicle width as an example, that is, the lane width corresponding to the boundary point sequence pair (that is, the distance between the two boundary point sequence pairs (left and right boundaries)) is less than half the vehicle width of the autonomous driving vehicle. Moreover, the absolute value of the difference between the lane width of the boundary point sequence pair with sequence number 29 and the lane width of the boundary point sequence pair corresponding to sequence number 28 is less than the second threshold. The second threshold here can be flexibly set according to user and business needs.

[0129] Condition 2: The value of the left boundary of the lane of the boundary point sequence pair corresponding to sequence number 29 is greater than the value of the right boundary of the lane of the boundary point sequence pair corresponding to sequence number 28, or the value of the right boundary of the lane of the boundary point sequence pair corresponding to sequence number 29 is less than the value of the left boundary of the lane of the boundary point sequence pair corresponding to sequence number 28, or the value of the left boundary of the lane of the boundary point sequence pair corresponding to sequence number 29 is greater than the value of the right boundary of the lane of the boundary point sequence pair corresponding to sequence number 28, and the value of the right boundary of the lane of the boundary point sequence pair corresponding to sequence number 29 is less than the value of the left boundary of the lane of the boundary point sequence pair corresponding to sequence number 28. In general, the value of the left boundary of the lane is less than the value of the right boundary, and they are at a safe passable distance. However, if the left boundary of the current boundary point sequence pair is located on the right side of the previous boundary point sequence pair, or the right boundary of the current boundary point sequence pair is located on the left side of the previous boundary point sequence pair, it means that the boundary corresponding to sequence number 29 is not continuous with the boundary corresponding to 28.

[0130] If the boundary point sequence pair corresponding to sequence number 29 meets any of the above conditions, it is determined that the boundary corresponding to sequence number 29 has an abnormal change. In this way, all boundary point sequence pairs in the first-class boundary interval [26-50] are traversed to determine that sequence number 29 and sequence number 45 meet the above conditions. Then, with 29 and 45, the first-class boundary interval [26-50] is divided into three first-class boundary sub-intervals [26-29], [27-45], and [46-50].

[0131] Similarly, according to a similar processing method, the first-class boundary interval [86-100] is divided into two first-class boundary sub-intervals [86-89] and [90-100].

[0132] Optionally, in the scheme of the present application, for each second-class boundary interval, M boundary point sequence pairs with abnormal changes in the boundary of the passage space are searched, and the second-class boundary interval is divided into M+1 second-class boundary sub-intervals according to the M boundary point sequence pairs, specifically including: for each second-class boundary interval, respectively execute: determine whether the number of boundary point sequence pairs contained in the second-class boundary interval is less than a third threshold; if the judgment result is that the number of boundary point sequence pairs is less than the third threshold, no processing is performed; if the judgment result is that the number of boundary point sequence pairs is not less than the third threshold, traverse each boundary point sequence pair in the second-class boundary interval, and based on the coordinate information of the boundary point sequence pair, determine whether each boundary point sequence pair satisfies the following At least one second-category constraint condition: the absolute value of the difference between the lane width corresponding to the current boundary point sequence pair and the lane width corresponding to the previous boundary point sequence pair is less than a fourth threshold; the left lane boundary and / or the right lane boundary corresponding to the current boundary point sequence pair are discontinuous; if a boundary point sequence pair satisfies the at least one second-category constraint condition, it is determined that the boundary point sequence pair can cause abnormal changes in the boundary of the passage space, and M boundary point sequence pairs that satisfy the constraint conditions are counted; based on the first and last boundary point sequence pairs of the second-category boundary interval and the M boundary point sequence pairs, the second-category boundary interval is divided into M+1 second-category boundary sub-intervals; if none of the boundary point sequence pairs satisfies the at least one second-category constraint condition, no processing is performed.

[0133] When processing the second type of boundary interval [1-25], start from the boundary point sequence pair numbered 1 and traverse each boundary point sequence pair. If a boundary point sequence pair, such as 10, corresponds to a boundary point sequence pair, determine whether the sequence pair meets one or all of the following conditions:

[0134] Condition a: The absolute value of the difference between the lane width corresponding to the boundary point sequence pair of sequence number 10 and the lane width corresponding to the boundary point sequence pair of sequence number 9 is less than the fourth threshold, which can be flexibly set according to user and business requirements. An empirical value can be taken here, indicating that there is a sudden change in lane width near the boundary point.

[0135] Condition b: The value of the left boundary of the lane of the boundary point sequence pair corresponding to sequence number 10 is greater than the value of the right boundary of the lane of the boundary point sequence pair corresponding to sequence number 9, or the value of the right boundary of the lane of the boundary point sequence pair corresponding to sequence number 10 is less than the value of the left boundary of the lane of the boundary point sequence pair corresponding to sequence number 9, or the value of the left boundary of the lane of the boundary point sequence pair corresponding to sequence number 10 is greater than the value of the right boundary of the lane of the boundary point sequence pair corresponding to sequence number 9, and the value of the right boundary of the lane of the boundary point sequence pair corresponding to sequence number 10 is less than the value of the left boundary of the lane of the boundary point sequence pair corresponding to sequence number 9. In general, the value of the left boundary of the lane is less than the value of the right boundary, and they are at a safe passable distance. However, if the left boundary of the current boundary point sequence pair is located on the right side of the previous boundary point sequence pair, or the right boundary of the current boundary point sequence pair is located on the left side of the previous boundary point sequence pair, it means that the boundary corresponding to sequence number 10 is not continuous with the boundary corresponding to 9.

[0136] If the boundary point sequence pair corresponding to sequence number 10 meets any of the above conditions, it is determined that the boundary corresponding to sequence number 10 has an abnormal change. In this way, all boundary point sequence pairs in the second-class boundary interval [1-25] are traversed to determine that sequence number 10 meets the above conditions. Then, with 10, the first-class boundary interval [1-25] is again divided into two second-class boundary sub-intervals [1-10] and [11-25].

[0137] Similarly, according to a similar processing method, the second-class boundary interval [51-85] is divided into three second-class boundary sub-intervals [51-63], [64-85], and [86-100].

[0138] It should be noted that before dividing the second-class boundary interval, the number of boundary point sequence pairs contained in each second-class boundary interval can also be determined. If it is less than the third threshold (for example, the third threshold is 3, in fact, it can also be other values, which can be reasonably determined according to the number of boundary points), the second-class boundary interval can be exempted from division. This is because the number of sequence pairs contained in the boundary interval itself is small, and only taking the first and last boundary point sequence pairs can well reflect the actual situation of the lane boundary.

[0139] It should be understood that the above examples show that all first-class boundary intervals or all second-class boundary intervals are divided and processed. In fact, in the actual processing process, it may happen that only some first-class boundary intervals are divided and processed, while other first-class boundary intervals are not divided and processed. Alternatively, only some second-class boundary intervals are divided and processed, while other second-class boundary intervals are not divided and processed. Alternatively, only some first-class boundary intervals are divided and processed, and only some second-class boundary intervals are divided and processed. Alternatively, only some first-class boundary intervals are divided and processed, and all second-class boundary intervals are divided and processed, etc. Therefore, after the above division processing, the obtained boundary intervals may include: part or all of the first-class boundary intervals, the first-class boundary sub-intervals, the second-class boundary intervals, and the second-class boundary sub-intervals.

[0140] Further, when interpolation processing is performed on the first-category boundary interval and the second-category boundary interval obtained by division, it can specifically include: using a first interpolation method to perform interpolation processing on multiple first-category boundary sub-intervals obtained by dividing the first-category boundary interval and the first-category boundary interval that has not been divided; using a second interpolation method to perform interpolation processing on multiple second-category boundary sub-intervals obtained by dividing the second-category boundary interval and the second-category boundary interval that has not been divided.

[0141] It should be noted that the first interpolation method here is similar to the interpolation method for the first type of boundary interval, that is, the number of intervals can be different from or the same as the second type of boundary interval and the second type of boundary subinterval. The second interpolation method is similar to the interpolation method for the second type of boundary interval. The above interpolation processing description can be referred to and will not be repeated here.

[0142] Reference Figure 3 As shown, it is a schematic diagram of the steps of a path planning method provided by the present application. The execution subject of the path planning method is a path planning method device. In the present application, the path planning method device can be a software module with computing and data processing capabilities, or an electronic device integrated with a similar software module. Optionally, the execution subject of the path planning method device in the present application can be an autonomous driving system server, an autonomous driving vehicle, or a related processing module integrated in an autonomous driving vehicle.

[0143] The path planning method may include the following steps:

[0144] Step 302: Acquire lane information and obstacle information, and determine boundary constraints; wherein the boundary constraints are determined by the above-mentioned point model boundary constraint construction method.

[0145] This part can refer to the above Figure 1 and detailed content.

[0146] Step 304: Based on the boundary constraints, a graph structure is constructed, and a preliminary path is determined using a graph search method.

[0147] It should be understood that when constructing a graph structure, a preliminary path can be determined based on a graph search algorithm such as depth-first search, breadth-first search, Dijkstra, etc.

[0148] Step 306: Based on the boundary constraints and the preliminary path, optimize to obtain a precise path.

[0149] After that, the preliminary path can be further optimized using boundary constraints to obtain a precise path. Since the boundary constraints used are of reasonable scale and can truly reflect lane boundary information, they can more accurately assist in the precise planning and layout of the path and ensure the safe and reliable driving of the autonomous driving vehicle.

[0150] It should be noted that part or all of the executor of the above steps may be an application located in the local terminal, or may also be a functional unit such as a plug-in or software development kit (SDK) set in the application located in the local terminal, or may also be a processing engine located in a network-side server, or may also be a distributed system located on the network side, for example, a processing engine or distributed system in an autonomous driving platform on the network side, etc. This embodiment does not specifically limit this.

[0151] It is understandable that the application may be a local program (nativeApp) installed on the local terminal, or may be a webpage program (webApp) of a browser on the local terminal, which is not limited in this embodiment.

[0152] In this way, according to the occupied area and passable area in the constructed lane local map, the optimal pass topological space is determined, and the boundary point sequence pairs distributed on both sides of the optimal pass topological space are obtained; then, the boundary point sequence pairs are divided into at least one first-class boundary interval and at least one second-class boundary interval, and the first-class boundary interval and the second-class boundary interval obtained by the division are interpolated respectively. Based on the sequence pairs obtained by the interpolation processing and the first and last boundary point sequence pairs of each boundary interval, the point model boundary constraints are constructed. Since the first-class boundary interval contains the boundary point sequence pairs corresponding to the obstacles, and the second-class boundary interval only contains the boundary point sequence pairs corresponding to the lanes, the boundary anomalies caused by the obstacles can be reflected by dividing the boundary intervals of different categories, and the first and last boundary point sequence pairs of the boundary interval are extracted. The point model boundary constraints are constructed by combining the sequence pairs obtained by the interpolation processing. Therefore, the key point sequence pairs of the abnormal changes in the boundary of the pass space can be screened out to ensure the accuracy and reliability of the boundary constraints, and will not affect the computing power expenditure of the subsequent path planning. In short, the constructed boundary constraints are of reasonable scale and can truly reflect the lane boundary information, so as to facilitate the efficient implementation of path planning.

[0153] It should be noted that the specific implementation process provided in this implementation can be combined with the various specific implementation processes provided in the aforementioned implementation to implement the method for constructing the point model boundary constraint of this embodiment. For a detailed description, please refer to the relevant content in the aforementioned implementation, which will not be repeated here.

[0154] In order to better understand the method of the embodiment of the present application, the method of the embodiment of the present application is described below in conjunction with the accompanying drawings and specific application scenarios.

[0155] Reference Figure 4 , which is a flow chart of a method for constructing point model boundary constraints provided in another embodiment of the present application.

[0156] Step 402: Acquire lane data and obstacle data.

[0157] Specifically, you can obtain lane data from the map, obstacle data from perception, and at the same time, design the planning range.

[0158] Step 404: obstacle preprocessing.

[0159] Specifically, it is possible to: ignore obstacles outside the lane boundary; ignore obstacles outside the planning scope; ignore obstacles with special custom attributes; ignore special decision obstacles; ignore virtual obstacles; perform special processing on obstacle polygons: densification, cutting, and expansion; and save the processed obstacles to the obstacle container to be considered.

[0160] Step 406: Construct a local map.

[0161] Specifically, all obstacles in the obstacle container to be considered can be traversed to obtain their aaboxes, and the occupied position of the obstacle can be determined according to the horizontal and vertical coordinates of the obstacle boundary; a local map determined by the road boundary, planning scope, and obstacle occupied position under the Frenet system is constructed. The obstacle is represented by an aabox, and the local map is divided into grids of different sizes based on the outer contour of the obstacle aabox. The grid where the obstacle is located is the occupied area, and the non-obstacle grid is the passable area.

[0162] Step 408: Select a topological space and construct a pair of boundary point sequences.

[0163] Specifically, we can select the topological space based on the local map and treat the grid as a graph structure. Method 1: Design the cost and weight according to the size and relative relationship of the grid, and use graph search algorithms such as Dijkstra to obtain the optimal topological space; Method 2: Design the motion cost according to the forward deduction of the vehicle kinematics, and use the space where the optimal deduction path is located as the optimal topological space; Provide interfaces for other methods.

[0164] According to the grid, based on the optimal topological space, it is sampled in the longitudinal direction to obtain the boundary point pair sequence of the optimal topological space. Among them, each boundary point pair will store: the l coordinates of the left and right boundary points (calculated above); the l coordinates of the left and right lane boundary points and the left and right road boundary points (obtained from the map); the left and right obstacle types (indicating whether there are obstacles and the types of obstacles); each boundary point pair is sampled in the longitudinal direction and arranged from the vehicle position forward.

[0165] Step 410: Customize the relevant shrinkage strategy for the boundary point sequence and perform appropriate shrinkage processing.

[0166] Specifically, the boundary point sequence pairs can be shrunk according to the following parameter attributes: visual curb; half vehicle width; curvature; obstacle attribute classification discussion; development of other interfaces, etc. And the truncated boundary point pair sequence after shrinkage can be processed.

[0167] Step 412: Classify and process the contracted boundary point pair sequence to obtain the ROAD boundary and the OBS boundary.

[0168] Separate the boundaries formed by obstacles (OBS boundaries, corresponding to the first type of boundary interval) and the boundaries formed by lane boundaries (ROAD boundaries, corresponding to the second type of boundary interval) and sample them separately. First, distinguish between OBS and ROAD (assuming that there are 100 pairs of boundary point sequences in total, and the following numbers represent the sequence numbers of boundary point pairs). For example: 1-25 is ROAD boundary, 26-50 is OBS boundary, 51-85 is ROAD boundary, and 86-100 is OBS boundary.

[0169] Step 414: Sampling is performed again on the ROAD boundary and the OBS boundary respectively.

[0170] For example, 21-25 and 51-85 perform ROAD boundary sampling, and 26-50 and 86-100 perform OBS boundary sampling. It should be noted that the following OBS boundary and ROAD boundary processing are examples and do not limit their implementation methods. Without changing the principle, the examples that can be implemented by any programming language and any programming rules are all within the scope of protection of this application.

[0171]

OBS Boundary

[0172] First, find the interval of OBS boundary pairs from the shrunk boundary point pair sequence, for example: 26-50 and 86-100. Record the boundary point pair numbers in the OBS interval into the container need_sample_index and perform duplicate removal; for example: 26,27,...,49,50,86,87,...,99,100

[0173] Define a container obstacle_s_pairs, whose elements store intervals consisting of pair numbers of boundary points (element form [pair_1 number, pair_2 number], representing the interval from pair_1 number to pair_2 number). When obstacle_s_pairs is initialized, it has only one element, and the pair_1 number and pair_2 number of this element are both the first element of the container need_sample_index.

[0174] For example: initial container: 26-26; after update (detailed process later): 26-29, 30-45, 46-50, 86-89, 90-100.

[0175] Details of the update process of the container obstacle_s_pairs:

[0176] First, traverse need_sample_index; get the boundary point pair number bounds_index corresponding to each traversal index; record the l coordinates of the pair corresponding to bounds_index and the pair corresponding to bounds_index-1, and get the horizontal width (laterally passable space) of the two boundary point pairs: left_l, right_l, last_left_l, last_right_l, width, last_width.

[0177] Define is_extend. If the serial number of pair_2 of the last element of the container obstacle_s_pairs is equal to bounds_index - 1, is_extend is true; otherwise, it is false.

[0178] Define is_narrow. If width is less than the user-defined threshold (half of the vehicle width in this invention), is_narrow is true; otherwise, it is false.

[0179] Define is_mutation. If the absolute value of the difference between width and last_width is less than the user-defined threshold (0.3 meters in this invention), is_mutation is true; otherwise, it is false.

[0180] Define is_disjoint. If right_l is greater than last_left_l or left_l < last_right_l, is_disjoint is true; otherwise, it is false.

[0181] If the judgment logic is_extend &&!(is_narrow && is_mutation) &&!is_disjoint is true: the serial number of pair_2 of the last element of the obstacle_s_pairs container is updated to bounds_index; if it is false: a new element is added to the end of the obstacle_s_pairs container, and the start and end serial numbers of the element are both bounds_index.

[0182]

ROAD Boundary

[0183] Define an add_cut function that takes the start and end pair serial numbers as input and outputs a container road_s_pairs (similar to the container obstacle_s_pairs); create a new container cut_indexs with elements being pair serial numbers; if the number between the input start and end pair serial numbers is less than 3, directly store them in cut_indexs, otherwise traverse each pair serial number between the start and end pair serial numbers. Record the left and right boundaries of the two pairs before and after the traversed pair (similar to left_l, right_l, last_left_l, last_right_l, width, last_width above). If the absolute value of the difference between width and last_width is less than the user-defined threshold, add the traversed pair serial number to cut_indexs.

[0184] Sort cut_indexes by their longitudinal positions, traverse cut_indexes every 2 elements, extract 2 elements and construct an interval as [pair_1 number, pair_2 number] and add them to road_s_pairs. Define the start and end pair numbers of the contracted boundary point pair sequence as start_index and end_index, and classify the ROAD boundaries according to obstacle_s_pairs.

[0185] obstacle_s_pairs is empty, NO_OBS, a pure ROAD boundary without obstacles; call the add_cut function, and pass in the start and end pair numbers: start_index and end_index.

[0186] obstacle_s_pairs is not empty. If the pair_1 index of the first element of obstacle_s_pairs is greater than start_index, ONLY_BEFORE_OBS, the ROAD boundary before all OBS boundaries; call the add_cut function, pass in the starting pair index as start_index, and pass in the last pair index as the pair_1 index of the first element of obstacle_s_pairs - 1.

[0187] obstacle_s_pairs is not empty. If the pair_2 number of the last element of obstacle_s_pairs is less than end_index, ONLY_AFTER_OBS means the ROAD boundary after all OBS boundaries. Call the add_cut function and pass in the starting pair number as the pair number of the last element of obstacle_s_pairs + 1, and the last pair number as end_index.

[0188] In other cases, traverse obstacle_s_pairs. If the value of the traversed element pair_1 index - 1 is greater than the index of the last traversed element pair_2, ELSE, the ROAD boundary between the OBS boundary and the OBS boundary; call the add_cut function, pass in the starting pair index of the last traversed element pair_2 index + 1, and pass in the last pair index of the traversed element pair_1 index - 1.

[0189] Finally, sort the road_s_pairs according to their longitudinal front-to-back positions.

[0190] Step 416: perform interpolation processing on obstacle_s_pairs and road_s_pairs respectively to construct boundary constraints.

[0191] The elements in obstacle_s_pairs interpolate one point every custom distance (5 points in the present invention); the elements in road_s_pairs interpolate one point every custom distance (10 points in the present invention). Add the points originally stored in obstacle_s_pairs and road_s_pairs, and the output is sample_intervals.

[0192] Step 418: Perform path planning based on the constructed boundary constraints.

[0193] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0194] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0195] Figure 5 FIG. 1 shows a structural block diagram of a device for constructing a point model boundary constraint provided by an embodiment of the present application. Figure 5 As shown. The construction device 500 of the point model boundary constraint of this embodiment may include a first construction module 501, a determination module 502, a division module 503, and a second construction module 504. The first construction module 501 is used to construct a lane local map based on lane information and obstacle information, wherein the area where the obstacle is located in the lane local map is the occupied area, and the other areas are the passable areas; the determination module 502 is used to determine the optimal pass topological space according to the occupied area and the passable area in the lane local map, and obtain the boundary point sequence pairs distributed on both sides in the optimal pass topological space; the division module 503 is used to divide the boundary point sequence pairs into at least one first-class boundary interval and at least one second-class boundary interval; wherein the first-class boundary interval contains the boundary point sequence pairs corresponding to the obstacle, and the second-class boundary interval only contains the boundary point sequence pairs corresponding to the lane; the second construction module 504 is used to perform interpolation processing on the first-class boundary interval and the second-class boundary interval obtained by the division, respectively, and construct the point model boundary constraint based on the sequence pairs obtained by the interpolation processing and the first and last boundary point sequence pairs of each boundary interval.

[0196] It should be noted that part or all of the point model boundary constraint construction device of the present embodiment may be an application located in the local terminal, or may also be a functional unit such as a plug-in or software development kit (SDK) set in the application located in the local terminal, or may also be a processing engine located in a network-side server, or may also be a distributed system located on the network side, for example, a processing engine or distributed system in an autonomous driving platform on the network side, etc. This embodiment does not specifically limit this.

[0197] It is understandable that the application may be a local program (nativeApp) installed on the local terminal, or may be a webpage program (webApp) of a browser on the local terminal, which is not limited in this embodiment.

[0198] Optionally, in a possible implementation of this embodiment, when the first construction module 501 constructs a lane local map based on lane information and obstacle information, it is specifically used to obtain lane information and obstacle information within the current planned driving range based on the map and perception respectively; pre-process the obstacles contained in the obstacle information according to preset processing rules; determine the position information of the obstacles obtained after pre-processing based on the boundary coordinates in the obstacle information; and construct a lane local map based on the road boundary in the lane information, the current planned driving range and the position information of the obstacles.

[0199] Optionally, in a possible implementation of the present embodiment, when the division module 503 divides the boundary point sequence pairs into at least one first-class boundary interval and at least one second-class boundary interval, it is specifically used to perform boundary shrinkage processing on the boundary point sequence pairs based on the set shrinkage strategy, and truncate the sequence pairs with shrinkage abnormalities; the boundary sequence pairs retained after the shrinkage processing are divided into at least one first-class boundary interval and at least one second-class boundary interval according to the boundary type; for each first-class boundary interval, N boundary point sequence pairs with abnormal changes in the boundary of the passage space are searched, and the first-class boundary interval is divided into N+1 first-class boundary sub-intervals according to the N boundary point sequence pairs; and / or; for each second-class boundary interval, M boundary point sequence pairs with abnormal changes in the boundary of the passage space are searched, and the first-class boundary interval is divided into M+1 first-class boundary sub-intervals according to the M boundary point sequence pairs; wherein the N boundary point sequences are not the first and last boundary sequence pairs of the first-class boundary interval, and the M boundary point sequence pairs are not the first and last boundary sequence pairs of the second-class boundary interval; and N and M are both positive integers.

[0200] Optionally, in a possible implementation of the present embodiment, the division module 503 searches for N boundary point sequence pairs in which abnormal changes exist in the boundary of the passage space for each first-class boundary interval, and divides the first-class boundary interval into N+1 first-class boundary sub-intervals according to the N boundary point sequence pairs, and is specifically used to perform for each first-class boundary interval: traverse each boundary point sequence pair in the first-class boundary interval; based on the coordinate information of the boundary point sequence pairs, determine in turn whether each boundary point sequence pair satisfies at least one of the following first-class constraint conditions: the lane width corresponding to the current boundary point sequence pair is less than the first threshold, and the lane width corresponding to the current boundary point sequence pair is less than the first threshold, and the lane width corresponding to the current boundary point sequence pair is less than the first threshold. The absolute value of the difference between the corresponding lane width and the lane width corresponding to the previous boundary point sequence pair is less than the second threshold; the left lane boundary and / or the right lane boundary corresponding to the current boundary point sequence pair are discontinuous; if there is a boundary point sequence pair that satisfies the at least one first-category constraint condition, it is determined that the boundary point sequence pair can cause abnormal changes in the boundary of the passage space, and N boundary point sequence pairs that satisfy the constraint conditions are counted; based on the first and last boundary point sequence pairs of the first-category boundary interval and the N boundary point sequence pairs, the first-category boundary interval is divided into N+1 first-category boundary sub-intervals; if none of the boundary point sequence pairs satisfy the at least one first-category constraint condition, no processing is performed.

[0201] Optionally, in a possible implementation of the present embodiment, the division module 503 searches for M boundary point sequence pairs with abnormal changes in the boundary of the passage space for each second-class boundary interval, and divides the second-class boundary interval into M+1 second-class boundary sub-intervals according to the M boundary point sequence pairs. Specifically, for each second-class boundary interval, the following operations are performed: determining whether the number of boundary point sequence pairs contained in the second-class boundary interval is less than a third threshold value; if the determination result is that the number of boundary point sequence pairs is less than the third threshold value, no processing is performed; if the determination result is that the number of boundary point sequence pairs is not less than the third threshold value, traversing each boundary point sequence pair in the second-class boundary interval, and determining each boundary point sequence pair based on the coordinate information of the boundary point sequence pair. Whether the sequence pair satisfies at least one of the following second-category constraints: the absolute value of the difference between the lane width corresponding to the current boundary point sequence pair and the lane width corresponding to the previous boundary point sequence pair is less than a fourth threshold; the left lane boundary and / or the right lane boundary corresponding to the current boundary point sequence pair are discontinuous; if a boundary point sequence pair satisfies the at least one second-category constraint, it is determined that the boundary point sequence pair can cause abnormal changes in the boundary of the passage space, and M boundary point sequence pairs that satisfy the constraint conditions are counted; based on the first and last boundary point sequence pairs of the second-category boundary interval and the M boundary point sequence pairs, the second-category boundary interval is divided into M+1 second-category boundary sub-intervals; if none of the boundary point sequence pairs satisfies the at least one second-category constraint, no processing is performed.

[0202] Optionally, in a possible implementation of the present embodiment, when the second construction module 504 performs interpolation processing on the first-class boundary interval and the second-class boundary interval obtained by division, it is specifically used to use the first interpolation method to perform interpolation processing on the multiple first-class boundary sub-intervals obtained by dividing the first-class boundary interval and the first-class boundary interval that has not been divided; and use the second interpolation method to perform interpolation processing on the multiple second-class boundary sub-intervals obtained by dividing the second-class boundary interval and the second-class boundary interval that has not been divided.

[0203] Figure 6 FIG. 1 shows a structural block diagram of a path planning device provided by an embodiment of the present application. Figure 6 The path planning device 600 of this embodiment may include an acquisition module 601, a determination module 602 and an optimization module 603; wherein the acquisition module 601 is used to acquire lane information and obstacle information and determine boundary constraints; wherein the boundary constraints are obtained by Figure 1 The method for constructing the boundary constraints of the point model is determined; a determination module 602 is used to construct a graph structure based on the boundary constraints and determine a preliminary path using a graph search method; an optimization module 603 is used to optimize the precise path based on the boundary constraints and the preliminary path.

[0204] Through the technical solution of the present application, the optimal traffic topological space is determined according to the occupied area and the passable area in the constructed lane local map, and the boundary point sequence pairs distributed on both sides in the optimal traffic topological space are obtained; then, the boundary point sequence pairs are divided into at least one first-class boundary interval and at least one second-class boundary interval, and the first-class boundary interval and the second-class boundary interval obtained by the division are interpolated respectively, and the point model boundary constraints are constructed based on the sequence pairs obtained by the interpolation processing and the first and last boundary point sequence pairs of each boundary interval. Since the first-class boundary interval contains the boundary point sequence pairs corresponding to the obstacles, and the second-class boundary interval only contains the boundary point sequence pairs corresponding to the lanes, in this way, the boundary anomalies caused by the obstacles can be reflected by dividing the boundary intervals of different categories, and the first and last boundary point sequence pairs of the boundary interval are extracted, and the point model boundary constraints are constructed by combining the sequence pairs obtained by the interpolation processing, so that the key point sequence pairs of the abnormal changes in the boundary of the traffic space can be screened out to ensure the accuracy and reliability of the boundary constraints, and will not affect the computing power expenditure of the subsequent path planning. In short, the constructed boundary constraints are of reasonable scale and can truly reflect the lane boundary information, so as to facilitate the efficient implementation of path planning.

[0205] An embodiment of the present application provides a computer-readable storage medium, wherein the storage medium stores at least one instruction, and the at least one instruction is loaded and executed by a processor to implement the method for constructing a point model boundary constraint as described above.

[0206] An embodiment of the present application provides an electronic device, which includes a processor and a memory, wherein the memory stores at least one instruction, and the instruction is loaded and executed by the processor to implement the method for constructing point model boundary constraints as described above.

[0207] An embodiment of the present application provides an autonomous driving vehicle, including the electronic device as described above. Specifically, the autonomous driving vehicle can be a vehicle of level L2 or above.

[0208] In the technical solution of this application, the collection, storage, use, processing, transmission, provision and disclosure of user personal information involved are in compliance with the relevant laws and regulations and do not violate public order and good morals.

[0209] Figure 7 A schematic block diagram of an example electronic device 700 that can be used to implement an embodiment of the present application is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workbenches, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present application described and / or required herein.

[0210] like Figure 7 As shown, the electronic device 700 includes a computing unit 701, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 702 or a computer program loaded from a storage unit 708 into a random access memory (RAM) 703. In the RAM 703, various programs and data required for the operation of the electronic device 700 can also be stored. The computing unit 701, the ROM 702, and the RAM 703 are connected to each other via a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.

[0211] Multiple components in the electronic device 700 are connected to the I / O interface 705, including: an input unit 706, such as a keyboard, a mouse, etc.; an output unit 707, such as various types of displays, speakers, etc.; a storage unit 708, such as a disk, an optical disk, etc.; and a communication unit 709, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 709 allows the electronic device 700 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0212] The computing unit 701 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 701 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 701 performs the various methods and processes described above, such as the method for constructing the point model boundary constraints. For example, in some embodiments, the method for constructing the point model boundary constraints may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 708. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 700 via the ROM 702 and / or the communication unit 709. When the computer program is loaded into the RAM 703 and executed by the computing unit 701, one or more steps of the method for constructing the point model boundary constraints described above may be executed. Alternatively, in other embodiments, the computing unit 701 may be configured to execute the method for constructing the point model boundary constraints in any other appropriate manner (for example, by means of firmware).

[0213] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0214] The program code for implementing the method of the present application can be written in any combination of one or more programming languages. These program codes can be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, implements the functions / operations specified in the flow chart and / or block diagram. The program code can be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0215] In the context of the present application, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (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 foregoing.

[0216] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0217] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0218] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship of client and server is generated by computer programs running on respective computers and having a client-server relationship with each other. The server may be a cloud server, a server of a distributed system, or a server combined with a blockchain.

[0219] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution disclosed in this application can be achieved, and this document is not limited here.

[0220] The above specific implementations do not constitute a limitation on the protection scope of this application. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of this application should be included in the protection scope of this application.

Claims

1. A method for constructing a point model boundary constraint, characterized in that: include: Constructing a lane local map based on the lane information and the obstacle information, wherein the area where the obstacle is located in the lane local map is the occupied area, and the other areas are the passable areas; Determine an optimal traffic topology space according to the occupied area and the passable area in the lane local map, and obtain boundary point sequence pairs distributed on both sides of the optimal traffic topology space; Dividing the boundary point sequence pairs into at least one first-class boundary interval and at least one second-class boundary interval; wherein the first-class boundary interval contains boundary point sequence pairs corresponding to obstacles, and the second-class boundary interval only contains boundary point sequence pairs corresponding to lanes; Interpolation processing is performed on the first-category boundary interval and the second-category boundary interval obtained by the division, respectively, and point model boundary constraints are constructed based on the sequence pairs obtained by the interpolation processing and the first and last boundary point sequence pairs of each boundary interval.

2. The method according to claim 1, characterized in that Construct a lane local map based on lane information and obstacle information, including: Based on the map and perception, the lane information and obstacle information within the current planned driving range are obtained respectively; Preprocessing the obstacles contained in the obstacle information according to a preset processing rule; Determining the position information of the obstacle obtained through preprocessing based on the boundary coordinates in the obstacle information; A lane local map is constructed based on the road boundary, the current planned driving range, and the location information of obstacles in the lane information.

3. The method according to claim 1, characterized in that Dividing the boundary point sequence pair into at least one first-type boundary interval and at least one second-type boundary interval specifically includes: Performing boundary shrinkage processing on the boundary point sequence pairs based on a set shrinkage strategy, and truncating sequence pairs with shrinkage abnormalities; Dividing the boundary sequence pairs retained after the shrinkage process into at least one first-class boundary interval and at least one second-class boundary interval according to the boundary type; For each first-category boundary interval, search for N boundary point sequence pairs with abnormal changes in the boundary of the passage space, and divide the first-category boundary interval into N+1 first-category boundary sub-intervals according to the N boundary point sequence pairs; and / or; for each second-category boundary interval, search for M boundary point sequence pairs with abnormal changes in the boundary of the passage space, and divide the first-category boundary interval into M+1 first-category boundary sub-intervals according to the M boundary point sequence pairs; wherein the N boundary point sequences are not the first and last boundary sequence pairs of the first-category boundary interval, and the M boundary point sequence pairs are not the first and last boundary sequence pairs of the second-category boundary interval; and both N and M are positive integers.

4. The method according to claim 3, characterized in that For each first-class boundary interval, find N boundary point sequence pairs with abnormal changes in the boundary of the passage space, and divide the first-class boundary interval into N+1 first-class boundary sub-intervals according to the N boundary point sequence pairs, specifically including: For each first-class boundary interval, execute: Traversing each boundary point sequence pair in the first type boundary interval; Based on the coordinate information of the boundary point sequence pair, it is determined in turn whether each boundary point sequence pair satisfies at least one of the following first-category constraints: the lane width corresponding to the current boundary point sequence pair is less than a first threshold, and the absolute value of the difference between the lane width corresponding to the current boundary point sequence pair and the lane width corresponding to the previous boundary point sequence pair is less than a second threshold; the left boundary and / or the right boundary of the lane corresponding to the current boundary point sequence pair is discontinuous; If there is a boundary point sequence pair that satisfies at least one of the first-type constraint conditions, it is determined that the boundary point sequence pair can cause an abnormal change in the boundary of the passage space, and N boundary point sequence pairs that satisfy the constraint conditions are counted; based on the first and last boundary point sequence pairs of the first-type boundary interval and the N boundary point sequence pairs, the first-type boundary interval is divided into N+1 first-type boundary sub-intervals; If none of the boundary point sequence pairs satisfy the at least one first-type constraint condition, no processing is performed.

5. The method according to claim 3, characterized in that For each second-class boundary interval, find M boundary point sequence pairs with abnormal changes in the boundary of the passage space, and divide the second-class boundary interval into M+1 second-class boundary sub-intervals according to the M boundary point sequence pairs, specifically including: For each second-class boundary interval, execute: Determine whether the number of boundary point sequence pairs included in the second type boundary interval is less than a third threshold; If the result of the judgment is that the number of boundary point sequence pairs is less than the third threshold, no processing is performed; If the judgment result is that the number of boundary point sequence pairs is not less than the third threshold, then traverse each boundary point sequence pair in the second-type boundary interval, and based on the coordinate information of the boundary point sequence pair, judge whether each boundary point sequence pair satisfies at least one of the following second-type constraint conditions: the absolute value of the difference between the lane width corresponding to the current boundary point sequence pair and the lane width corresponding to the previous boundary point sequence pair is less than the fourth threshold; the left boundary and / or the right boundary of the lane corresponding to the current boundary point sequence pair are discontinuous; If there is a boundary point sequence pair that satisfies the at least one second-type constraint condition, it is determined that the boundary point sequence pair can cause an abnormal change in the boundary of the passage space, and M boundary point sequence pairs that satisfy the constraint condition are counted; based on the first and last boundary point sequence pairs of the second-type boundary interval and the M boundary point sequence pairs, the second-type boundary interval is divided into M+1 second-type boundary sub-intervals; If none of the boundary point sequence pairs satisfy the at least one second type constraint condition, no processing is performed.

6. The method according to any one of claims 4 or 5, characterized in that The first-class boundary interval and the second-class boundary interval obtained by division are interpolated respectively, including: For a plurality of first-class boundary subintervals obtained by dividing the first-class boundary interval and the first-class boundary interval that has not been divided, interpolation processing is performed using a first interpolation method; The second interpolation method is used to perform interpolation processing on a plurality of second-type boundary sub-intervals obtained by dividing the second-type boundary interval and the second-type boundary interval that has not been divided.

7. A path planning method, characterized in that: include: Acquire lane information and obstacle information, and determine boundary constraints; wherein the boundary constraints are determined by the point model boundary constraint construction method according to any one of claims 1 to 6; Based on the boundary constraints, a graph structure is constructed, and a preliminary path is determined by using a graph search method; Based on the boundary constraints and the preliminary path, a precise path is optimized.

8. A device for constructing a point model boundary constraint, characterized in that: include: A first construction module is used to construct a lane local map based on the lane information and the obstacle information, wherein the area where the obstacle is located in the lane local map is an occupied area, and other areas are passable areas; A determination module, used to determine an optimal pass topological space according to the occupied area and the passable area in the lane local map, and obtain boundary point sequence pairs distributed on both sides of the optimal pass topological space; A division module, used to divide the boundary point sequence pairs into at least one first-type boundary interval and at least one second-type boundary interval; wherein the first-type boundary interval contains boundary point sequence pairs corresponding to obstacles, and the second-type boundary interval only contains boundary point sequence pairs corresponding to lanes; The second construction module is used to perform interpolation processing on the first type boundary interval and the second type boundary interval obtained by division, and construct point model boundary constraints based on the sequence pairs obtained by the interpolation processing and the first and last boundary point sequence pairs of each boundary interval.

9. A path planning device, characterized in that: include: An acquisition module, used to acquire lane information and obstacle information, and determine boundary constraints; wherein the boundary constraints are determined by the point model boundary constraint construction method according to any one of claims 1 to 6; A determination module, configured to construct a graph structure based on the boundary constraints and determine a preliminary path using a graph search method; The optimization module is used to optimize and obtain a precise path based on the boundary constraints and the preliminary path.

10. An electronic device comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method for constructing point model boundary constraints according to any one of claims 1-6, or the path planning method according to claim 7.

11. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to enable the computer to execute the method for constructing point model boundary constraints according to any one of claims 1 to 6, or the path planning method according to claim 7.

12. A computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the method for constructing a point model boundary constraint according to any one of claims 1 to 6, or the path planning method according to claim 7 is implemented.

13. An autonomous driving vehicle comprising the electronic device as claimed in claim 10.

Images