Vehicle travelable area boundary generation method and device, storage medium and vehicle

Abstract

The invention discloses a vehicle drivable area boundary generation method and device, a storage medium and a vehicle, and relates to the technical field of auxiliary driving, and the method comprises the steps: obtaining lane data and obstacle target data of a current vehicle in real time; sampling lane data to obtain a lane discrete point set; sampling the obstacle target data to obtain an obstacle target trace point set; and constructing a current occupation grid map according to the lane discrete point set and the obstacle target trace point set, and generating a drivable area boundary according to the current occupation grid map. According to the invention, sampling processing is carried out on the obtained lane data and obstacle target data to construct the current occupied grid map, and then the drivable area boundary of the current vehicle is generated, so that different perception data which can be obtained by the vehicle can be processed in a unified manner; therefore, the problem of abnormal vehicle driving track control caused by excessive dependence on lane line data at the present stage is solved in different scenes, especially in a scene with difficulty in lane line recognition.

Description

Technical Field

[0001] The present application relates to the field of assisted driving technology, and in particular to a method, device, storage medium and vehicle for generating the boundary of a vehicle drivable area. Background Art

[0002] With the development of intelligent assistance technology for automobiles, L2-level assisted driving and even higher-level assisted driving functions have been basically popularized in new energy vehicles, and users' requirements and acceptance of assisted driving are also increasing. Vehicles achieve assisted driving based on the Advanced Driver Assistance System (ADAS). During the assisted driving process, ADAS controls the vehicle laterally by defining the safe areas where the vehicle can drive to the left and right.

[0003] However, most vehicles' ADAS rely entirely on lane line data when controlling the vehicle using Lane Center Control (LCC) technology. Therefore, when faced with difficult lane line recognition scenarios such as blurred lane markings or lane lines blocked by stationary vehicles, ADAS has difficulty accurately knowing the boundary line of the vehicle's drivable area, which in turn leads to abnormal control of the vehicle's driving trajectory. Summary of the invention

[0004] The main purpose of the present application is to provide a method, device, storage medium and vehicle for generating the boundary of a vehicle drivable area, aiming to solve the technical problem that the existing vehicle driving trajectory control is completely dependent on lane line data, which easily causes abnormal vehicle trajectory control in scenarios where lane line recognition is difficult.

[0005] To achieve the above purpose, the present application proposes a method for generating a boundary of a vehicle drivable area, the method comprising:

[0006] Obtain the lane data and obstacle target data of the current vehicle in real time;

[0007] Sampling the lane data to obtain a lane discrete point set;

[0008] Sampling the obstacle target data to obtain an obstacle target point trace set;

[0009] A current occupied grid map is constructed according to the lane discrete point set and the obstacle target point trace set, and a drivable area boundary is generated according to the current occupied grid map.

[0010] In one embodiment, the step of sampling the lane data to obtain a lane discrete point set includes:

[0011] Based on the preset discrete sampling rules, determine the lane sampling step length and the number of lane sampling points;

[0012] The lane data is sampled according to the lane sampling step length and the number of lane sampling points to obtain a lane discrete point set.

[0013] In one embodiment, the step of sampling the obstacle target data to obtain an obstacle target point trace set includes:

[0014] Determine the target position and target size of the obstacle target according to the obstacle target data;

[0015] Based on a preset contour sampling rule, determining an obstacle target sampling interval according to the target position and the target size;

[0016] The obstacle target is sampled according to the obstacle target sampling interval to obtain an obstacle target point trace set.

[0017] In one embodiment, after the step of determining the obstacle target sampling interval according to the target position and the target size based on the preset contour sampling rule, the following steps are included:

[0018] Determine the nearest identifiable edge of each obstacle target according to the target position, and judge whether the nearest identifiable edge is the lateral nearest edge;

[0019] If yes, a nearest edge conversion operation is performed on the nearest identified edge, and the target position is updated according to the conversion result, wherein the nearest edge conversion operation is used to convert the nearest identified edge into a rear-preserved nearest edge;

[0020] If not, when the nearest identified edge is the nearest edge of the front guard and the obstacle target is a lateral target, a nearest edge conversion operation is performed on the nearest identified edge, and each target position is updated according to the conversion result;

[0021] Accordingly, the step of determining the obstacle target sampling interval according to the target position and the target size based on the preset contour sampling rule includes:

[0022] Based on the preset contour sampling rule, the obstacle target sampling interval is determined according to the updated target position and the target size.

[0023] In one embodiment, the step of constructing a current occupancy grid map according to the lane discrete point set and the obstacle target point trace set, and generating a drivable area boundary according to the current occupancy grid map, further includes:

[0024] Numerically filling an initial grid map with the lane discrete point set and the obstacle target point trace set according to a preset filling rule to obtain a current occupied grid map, wherein the initial grid map is generated based on a pre-configured map size;

[0025] Extracting the center point position of each occupied grid in the current occupied grid map, and generating a set of boundary points of the driving area at the current moment according to the center point position;

[0026] The driving area boundary point set at the previous moment is updated according to the driving area boundary point set at the current moment, and the drivable area boundary at the current moment is obtained according to the update result.

[0027] In one embodiment, the step of updating the set of boundary points of the driving area at the previous moment according to the set of boundary points of the driving area at the current moment, and obtaining the boundary of the drivable area at the current moment according to the update result, comprises:

[0028] Acquire a set of boundary points of the driving area at a previous moment, and update the set of boundary points of the driving area at a previous moment according to the driving displacement of the current vehicle from the previous moment to the current moment, so as to obtain a set of boundary points of a candidate drivable area;

[0029] Performing linear interpolation on the candidate drivable area boundary point set according to the drivable area boundary point set at the current moment to obtain a drivable area boundary point set;

[0030] The set of boundary points of the drivable area is filtered, and the boundary of the drivable area at the current moment is generated according to the filtering process.

[0031] In one embodiment, the lane data and the obstacle target data are collected by a forward-looking sensor and a millimeter-wave radar installed on the current vehicle, the lane data includes: lane line data and curb data, and the obstacle target data includes obstacle data and traffic target data;

[0032] Correspondingly, the lane discrete point set includes: a lane line discrete point set and a roadside point trace set; the obstacle target point trace set includes: an obstacle contour point trace set and a target contour point trace set.

[0033] In addition, to achieve the above purpose, the present application also proposes a device for generating a boundary of a vehicle drivable area, the device comprising:

[0034] Data acquisition module, used to obtain the lane data of the current vehicle and obstacle target data in real time;

[0035] A data processing module, used to sample the lane data to obtain a lane discrete point set; and to sample the obstacle target data to obtain an obstacle target point trace set;

[0036] The boundary generation module is used to construct a current occupied grid map according to the lane discrete point set and the obstacle target point trace set, and generate a drivable area boundary according to the current occupied grid map.

[0037] In addition, to achieve the above-mentioned purpose, the present application also proposes a storage medium, on which a vehicle drivable area boundary generation program is stored. When the vehicle drivable area boundary generation program is executed by a processor, the steps of the vehicle drivable area boundary generation method as described above are implemented.

[0038] In addition, to achieve the above-mentioned purpose, the present application also proposes a vehicle, on which an advanced driving assistance system is installed, and the advanced driving assistance system includes: a memory, a processor, and a vehicle drivable area boundary generation program stored in the memory and executable on the processor, and the vehicle drivable area boundary generation program is configured to implement the steps of the vehicle drivable area boundary generation method as described above.

[0039] The present application provides a method for generating the boundary of a drivable area of ​​a vehicle, including: acquiring the lane data and obstacle target data of the current vehicle in real time; sampling the lane data to obtain a lane discrete point set; sampling the obstacle target data to obtain an obstacle target point trace set; constructing a current occupied grid map based on the lane discrete point set and the obstacle target point trace set, and generating the boundary of the drivable area based on the current occupied grid map. The present application samples and processes the obtained lane data and obstacle target data to construct a current occupied grid map, and then generates the boundary of the drivable area of ​​the current vehicle, so as to uniformly process the different perception data available to the vehicle, thereby compensating for the abnormal vehicle driving trajectory control problem caused by over-reliance on lane line data in different scenarios, especially in scenarios where lane line recognition is difficult. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0041] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0042] Figure 1 This is a flow chart of the first embodiment of the method for generating the boundary of the vehicle drivable area of ​​the present application;

[0043] Figure 2 This is a schematic diagram of discrete sampling of lane line data;

[0044] Figure 3 This is the overall flow chart of the method for generating the boundary of the vehicle drivable area of ​​the present application;

[0045] Figure 4 This is a flow chart of a second embodiment of the method for generating a boundary of a vehicle drivable area of ​​the present application;

[0046] Figure 5 It is the logic diagram of transformation rules for the most recently identified edge;

[0047] Figure 6 This is a schematic diagram of the sampling of traffic target data contours;

[0048] Figure 7 This is a flow chart of a third embodiment of the method for generating a boundary of a vehicle drivable area of ​​the present application;

[0049] Figure 8 This is an example of the current occupancy grid map;

[0050] Fig. 9 Update the flow chart for the set of driving area boundary points;

[0051] Fig.10 The diagram is a schematic diagram of the calculation method of the set of boundary points of the drivable area;

[0052] Fig.11 is the transformation map of the drivable area boundary;

[0053] Fig.12 A schematic diagram showing the comparison between the boundary generation results of the vehicle drivable area boundary generation method and the traditional method;

[0054] Fig.13 This is a schematic diagram of the module structure of the first embodiment of the vehicle drivable area boundary generation device of the present application.

[0055] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0056] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0057] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0058] The present application embodiment provides a method for generating a boundary of a vehicle drivable area, referring to Figure 1 , Figure 1 This is a flow chart of a first embodiment of a method for generating a boundary of a vehicle drivable area of ​​the present application. In this embodiment, the method includes: Steps S10 to S40:

[0059] Step S10: Acquire the lane data and obstacle target data of the current vehicle in real time.

[0060] It should be noted that the method of this embodiment can be applied in the scenario where the vehicle realizes assisted driving through ADAS, and the execution subject of the method of this embodiment can be a vehicle drivable area boundary generation device with data processing and program running functions, and can also be other control units in the vehicle for lateral control of the vehicle based on LCC technology. Here, the above-mentioned vehicle drivable area boundary generation device (referred to as "boundary generation device") is taken as an example to illustrate the various embodiments of this application.

[0061] It is understandable that the current vehicle may be equipped with a forward-looking sensor and a millimeter-wave radar, which complement each other in the vehicle's ADAS to generate fused perception data. Based on the fused perception data, the lane data and obstacle target data of the current vehicle can be obtained in real time.

[0062] It should be understood that the lane data may include lane line data and curb data; the obstacle target data may include obstacle data and traffic target data. Obstacles and traffic targets may be objects in the perception and recognition whitelist. Obstacles may include: stationary non-traffic participants such as cones and water barriers; traffic targets may include traffic participants such as cars, trucks, buses, two-wheeled vehicles, pedestrians, etc. Traffic participants have basic information such as location, speed, orientation angle, type, length, width, height, etc.

[0063] It should also be noted that before processing the acquired lane data and obstacle target data, the map size of the current vehicle can be set first, so as to facilitate the subsequent unified and standardized data sampling of different lane data and obstacle target data. The map size setting process is as follows:

[0064] First, you can set the length and width of the map: set the vertical length of the Range_X surface map to Xm, and the horizontal length of the Range_Y surface map to Ym; then set the number of vertical grids in the map to Key_X, the number of horizontal grids to Key_Y, and calculate the vertical and horizontal grid specifications: the length of a single grid is Width is rectangle.

[0065] It should be understood that, based on the above map size setting process, an initial grid map can be obtained, and the coordinate system of the initial grid map can be a yx coordinate system.

[0066] Step S20: sampling the lane data to obtain a lane discrete point set.

[0067] It should be understood that the boundary generating device may first determine the lane sampling step and the number of lane sampling points based on a preset discrete sampling rule; and then sample the lane data according to the lane sampling step and the number of lane sampling points to obtain a lane discrete point set.

[0068] It should be noted that the lane line data can be used as an example to illustrate the sampling and processing process of the lane line data. Based on the coordinate system of the above initial grid map, the lane line data can be output as a cubic polynomial equation y=C0+C1x+C2x 2 +C3x 3 , and recorded as f(x) Lane .

[0069] Reference Figure 2 The discrete sampling process is explained. Figure 2 This is a schematic diagram of the discrete sampling of lane line data. According to the fixed longitudinal sampling step x step Lane line equation f(x) Lane Sampling is performed, sampling N points to obtain the lane line discrete point set X Lane Among them, the number of sampling points N can be set to 30, and the sampling step size is set

[0070] The lane line discrete point set X Lane It can be expressed as:

[0071] X Lane ={(x1,y1)(x2,y2)...(x i ,y i )},i=1,2,...,N

[0072] x i =0+(i-1)x step

[0073] y i =f(x i ) Lane

[0074] Correspondingly, the roadside data and lane line data are both cubic polynomial equations, which can be expressed as f(x) RoadEdge , the roadside data f(x) is sampled in the same way as the lane line data sampling process above RoadEdgeBy performing discrete sampling, we can obtain the roadside point set X RoadEdge .

[0075] In a specific implementation, the boundary generation device performs discrete sampling on the lane line data and the roadside data to obtain the corresponding lane line discrete point set X Lane And the roadside point set X RoadEdge .

[0076] Step S30: sampling the obstacle target data to obtain an obstacle target point trace set.

[0077] It should be noted that the sampling process of obstacle target data can be described here by taking traffic target data as an example. Based on the coordinate system of the initial grid map, the traffic target data can be output as follows: the midpoint position of the nearest edge of the traffic target to the current vehicle (Pos x ,Pos y ), the length Length and width Width of the traffic target.

[0078] It should be understood that based on the preset lateral sampling step length x onj_step and the longitudinal sampling step length y obj_step , the midpoint position of the above traffic target (Pos x ,Pos y ) as the benchmark, the horizontal profile sampling and the vertical profile sampling of the traffic target are performed respectively to obtain the target contour point set X Object .

[0079] Accordingly, the obstacle contour point set X can be obtained by sampling the obstacle contour data in the same way as the above traffic target data sampling process. Obstacles .

[0080] In the specific implementation, the boundary generation device performs contour sampling on the traffic target data and the obstacle data respectively to obtain the corresponding target contour point set X Object And the obstacle contour point set X Obstacles .

[0081] Step S40: constructing a current occupied grid map according to the lane discrete point set and the obstacle target point trace set, and generating a drivable area boundary according to the current occupied grid map.

[0082] It should be understood that in the initial grid map, the default parameter of each grid is 0. You can set different filling values ​​for different point sets when they occupy the grid. For example, you can preset the filling rule as:

[0083]

[0084] Based on the above filling rules, since various data point sets (X Lane , X Object , X RoadEdge , X Obstacles ) are coordinate points represented by the coordinate system based on the initial grid map. Therefore, a binary search is performed in the initial grid map to determine the corresponding grid of each coordinate point in the initial grid map and fill in the corresponding value.

[0085] It can be understood that after completing the grid filling of each coordinate point, the current occupied grid map can be obtained, and then the driving boundaries of the left and right sides of the current vehicle can be determined according to the current occupied grid map.

[0086] In addition, you can refer to Figure 3 The overall process of the method for generating the boundary of the vehicle drivable area in this application is described. Figure 3 This is an overall flow chart of the method for generating the boundary of the vehicle drivable area of ​​this application.

[0087] exist Figure 3 In the method, the map boundary size of the current vehicle can be set first to obtain the initial grid map; then the lane line data, curb data, obstacle data and traffic target data obtained by the fusion of the forward monocular vision and millimeter-wave radar are sampled and processed respectively, so as to convert the non-uniform data forms into point set information, and the initial grid map is updated in combination with the filling rule to construct the current occupied grid map, and finally the drivable area boundary is output based on the current occupied grid map.

[0088] This embodiment aims at the problem of abnormal steering wheel control and lane deviation caused by the perception algorithm in the existing vehicle ADAS for lane stains, rainy day water stains, and backlight scenes. The perception algorithm does not clearly recognize the ground lane markings, and the recognition error causes abnormal steering wheel control and lane deviation. The lane line, curb, obstacle, traffic target and other data that can be obtained by the existing visual perception are processed in a unified manner to construct the current occupancy grid map, and then generate the boundary of the drivable area of ​​the current vehicle, so as to make up for the abnormal vehicle driving trajectory control caused by over-reliance on lane line data in the above-mentioned difficult lane line recognition scenes. And specifically adopts the perception data after the fusion of forward-looking monocular vision and millimeter-wave radar as the data source, which can adapt to the ADAS perception solutions of existing mainstream manufacturers without increasing hardware costs.

[0089] Based on the first embodiment of the present application, in the second embodiment of the present application, the same or similar contents as those in the above-mentioned embodiment 1 can be referred to the above introduction, and will not be repeated in the following. Figure 4 , Figure 4 This is a flow chart of the second embodiment of the method for generating the boundary of the vehicle drivable area of ​​the present application.

[0090] In this embodiment, in order to specifically illustrate the sampling process of obstacle target data, step S30 includes: steps S301 to S303.

[0091] Step S301: determining the target position and target size of the obstacle target according to the obstacle target data.

[0092] It should be noted that the sampling and processing process of obstacle target data can still be specifically described here by taking traffic target data as an example. The target position is the midpoint position of the closest edge of the traffic target to the current vehicle (Pos x ,Pos y ), the target size is the length Length and width Width of the traffic target.

[0093] In addition to the above-mentioned target position and target size, the traffic target data may also include: the nearest identified edge SideNear of the traffic target in the identification frame. The nearest identified edge can be divided into four categories: front protection position, rear protection position, left side and right side.

[0094] It should also be noted that when the vehicle is driving, we can only focus on the traffic targets in front of the rear axle of the current vehicle. Therefore, the positional relationship between the traffic targets and the current vehicle can be divided into two categories: forward targets (driving in the opposite direction of the current vehicle) and lateral targets (driving in the same direction as the current vehicle).

[0095] Furthermore, since the closest edge categories of different traffic targets to the current vehicle may be different, in order to facilitate the subsequent unified sampling process, the midpoint positions (Pos x ,Pos y ) for conversion, and there are also differences in the conversion methods for forward targets and backward targets.

[0096] Therefore, before step S302, the method further includes steps S001 to S003:

[0097] Step S001: Determine the nearest identifiable edge of each obstacle target according to the target position, and determine whether the nearest identifiable edge is the lateral nearest edge.

[0098] Step S002: If yes, a nearest edge conversion operation is performed on the nearest identified edge, and the target position is updated according to the conversion result, wherein the nearest edge conversion operation is used to convert the nearest identified edge into a rear-preserved nearest edge.

[0099] Step S003: If not, when the nearest identified edge is the front nearest edge and the obstacle target is a lateral target, a nearest edge conversion operation is performed on the nearest identified edge, and each target position is updated according to the conversion result.

[0100] You can refer to here Figure 5 The transformation rules of the nearest identified edge are explained. Figure 5 Logic diagram of transformation rules for the most recently identified edge.

[0101] After determining the category of the nearest recognized edge of each obstacle target (traffic target) according to the above target position, it can be first determined whether the nearest recognized edge SideNear is the lateral nearest edge (left or right).

[0102] If yes, directly perform the nearest edge conversion operation (based on the left and right sides) to convert the lateral nearest edge into the rear-preserving nearest edge and output it as the reference edge;

[0103] If not, then further determine whether SideNear is a front-preserving edge. If SideNear is not a front-preserving edge, that is, SideNear is a back-preserving edge, then the back-preserving edge can be directly output as the reference edge. If SideNear is a front-preserving edge, proceed to the next step of determination.

[0104] Determine whether the traffic target to which the front protection position edge belongs is the forward target of the current vehicle. If so, the front protection position edge can be directly output as the reference edge; if not, determine that the traffic target to which the front protection position edge belongs is a lateral target, and perform a nearest edge conversion operation (based on the front protection position) to convert the front protection position edge into the rear protection nearest edge and output it as the reference edge.

[0105] After obtaining each reference edge, the midpoint position coordinates of each reference edge, i.e., the updated target position, can be obtained. For the nearest identified edges of different categories, the conversion formula of the target position can be expressed as:

[0106] Front guard position edge:

[0107] Left position edge:

[0108] Right position edge:

[0109] Step S302: Based on a preset contour sampling rule, determine an obstacle target sampling interval according to the target position and the target size.

[0110] Step S303: sampling the obstacle target according to the obstacle target sampling interval to obtain an obstacle target point trace set.

[0111] You can refer to here Figure 6 The contour sampling process is explained. Figure 6 This is a schematic diagram of traffic target data contour sampling. Figure 6 Figure 6-a is a schematic diagram of transverse profile sampling. Figure 6 Figure 6-b is a schematic diagram of longitudinal profile sampling.

[0112] First, the obstacle target may be transversely sampled according to the transverse sampling interval to obtain a plurality of transverse contour points.

[0113] For example, reference Figure 6 6-a, yes Two points are sampled on each side of the reference, and the remaining five points are obtained by length translation, for a total of 10 points. The sampling interval of each horizontal point can be expressed as: Each transverse contour point can be expressed as:

[0114]

[0115] Then, the obstacle target is sampled longitudinally according to the longitudinal sampling interval to obtain a plurality of longitudinal contour points.

[0116] For example, reference Figure 6 In 6-b, after sampling the transverse profile, 6 points are sampled in the longitudinal profile, for a total of 12 points. The sampling interval of each longitudinal point can be expressed as: Each longitudinal contour point can be expressed as:

[0117]

[0118] Finally, determine the target contour point set X Object , the target contour point set X Object Including: each of the transverse contour points and each of the longitudinal contour points

[0119] In a specific implementation, the target contour point set X can be determined based on the above process. Object , and the obstacle data is sampled in the same way as the above traffic target data sampling process, and the obstacle contour point set X can be obtained Obstacles .

[0120] In this embodiment, it is considered that when sampling the contour of the obstacle target, the nearest edge, i.e. the nearest recognized edge, of each obstacle target may be different, so the nearest recognized edge is converted according to the specific category of the nearest recognized edge and the position relationship between the obstacle target and the current vehicle, so that the reference edge output after conversion is unified to update the position of each target, which is conducive to performing a unified horizontal and vertical contour sampling process for each obstacle target. And because the calculation method in this embodiment is simple, the computing power requirement is low, and it is more lightweight, thereby reducing the resource consumption required to generate the boundary of the vehicle drivable area.

[0121] Based on the first and second embodiments of the present application, in the third embodiment of the present application, the same or similar contents as those of the first and second embodiments can be referred to the above description, and will not be described in detail later. Figure 7 , Figure 7 This is a flow chart of the third embodiment of the method for generating the boundary of the vehicle drivable area of ​​the present application.

[0122] In this embodiment, in order to specifically explain how to determine the boundary of the drivable area at the moment according to the constructed current occupancy grid map, step S40 specifically includes: steps S401 to S403:

[0123] Step S401: numerically fill the initial grid map with the lane discrete point set and the obstacle target point trace set according to a preset filling rule to obtain a current occupied grid map, wherein the initial grid map is generated based on a pre-configured map size.

[0124] Step S402: extracting the center point position of each occupied grid in the current occupied grid map, and generating a set of boundary points of the driving area at the current moment according to the center point position.

[0125] It can be understood that the boundary trajectory tree grows from bottom to top, and the node position of the tree is the center point of the grid. The boundary coordinate sequence is saved to obtain the boundary measurement sequence extracted after the current occupancy grid map is established under the current measurement. That is, the set of boundary points of the driving area can be the boundary measurement sequence Z k , and Z k ={Z1Z2...Z n},n≤Key_X; Z1=(P 1,x ,P 1,y ).

[0126] Figure 8 This is an example of the current occupancy grid map. Based on the current grid map, the boundary measurement sequence Z is performed. k The extraction process can be taken as an example of the extraction process of the left boundary, and is expressed as the following steps:

[0127] First, it is known that the specifications of the current occupied grid map are: the number of vertical grids is Key_X, the number of horizontal grids is Key_Y, and the coordinate relationship can be established for each grid;

[0128] Step 1: Start from the X axis and search in the positive direction of the Y axis, and initialize the row number flag to j=0;

[0129] Step 2: Increment the row number flag (j++);

[0130] Step 3: Search the grid for the first non-zero position by direction to get the column number and record it as i;

[0131] Step 4: Calculate the center point of the grid coordinates according to row number j and column number i. Taking the first retrieved point as an example, the calculation formula is as follows:

[0132]

[0133] Z1=(P 1,x ,P 1,y )

[0134] Step 5: Fill the coordinate points into set Z in sequence k ;

[0135] Step 6: The row number flag is equal to Key_Y, and the obtained boundary sequence is output.

[0136] It can be understood that based on the above steps 1 to 6, the set of boundary points Z of the driving area at the current moment can be generated. k .

[0137] Step S403: updating the driving area boundary point set at the previous moment according to the driving area boundary point set at the current moment, and obtaining the drivable area boundary at the current moment according to the update result.

[0138] It should be noted that in order to obtain the latest drivable area boundary, the existing boundary data can be updated through the current measurement. Fig. 9 Explain the update process. Fig. 9 Update the flow diagram for the set of travel area boundary points.

[0139] It should also be noted that, if the current moment is represented as moment k, the previous moment can be represented as moment k-1.

[0140] First, coordinate transformation is performed on the set of boundary points of the driving area at time k-1, and the boundary data of the driving area at time k-1 is obtained by deleting the boundary points that have been driven. Then, the boundary data of the driving area at time k is obtained by using the set of boundary points Z of the driving area at time k. k Linear interpolation is performed on the boundary data at time k-1; finally, the above linear interpolation result is updated through Kalman filtering to obtain the boundary of the drivable area at the current moment.

[0141] Further, in order to specifically illustrate the above updating process, step S403 includes: steps S4031 to S4033:

[0142] Step S4031: Acquire the set of boundary points of the driving area at the last moment, and update the set of boundary points of the driving area at the last moment according to the driving displacement of the current vehicle from the last moment to the current moment, so as to obtain a set of candidate boundary points of the drivable area.

[0143] At time k, the current vehicle position and lateral angle changes relative to the vehicle position at time k-1 are: Δx, Δy, Δθ, and the set of driving boundary points output at time k-1 is Γ k-1 , which can be expressed as: Γ k-1 ={Γ1Γ2...Γ n},n≤Key_X;Γ1=(P 1,x,Γ ,P 1,y,Γ ). Among them, Γ n Represents a coordinate point. Then perform the following steps AB:

[0144] Step A: For Γ k-1 After calculating the coordinate system rotation and translation for each item, we can get The calculation of a single item is as follows:

[0145]

[0146] Step B: Each P 1,x,Γ The value of is judged, if P 1,x,Γ If ≤0, the point will be removed from the set.

[0147] Through the above steps AB, the result after the coordinates of the set of boundary points of the driving area at time k-1 are transferred and the points that have been driven across the boundary are eliminated is recorded as the candidate set of boundary points of the driving area.

[0148] Step S4032: performing linear interpolation on the candidate drivable area boundary point set according to the drivable area boundary point set at the current moment to obtain a drivable area boundary point set.

[0149] It can be understood that the current driving area boundary point set Z k The longitudinal position of the candidate drivable area boundary point set Perform linear interpolation to obtain the set of boundary points of the drivable area From this Z k and The number of points in the same.

[0150] Specifically, using Z k The vertical coordinate in The horizontal trend gets a vertical coordinate with Z k Consistent sequence Its expansion can be expressed as:

[0151] You can refer to here Fig.10 Specify the set of boundary points of the drivable area The calculation process, Fig.10 Schematic diagram of the calculation method of the set of boundary points of the drivable area.

[0152] exist Fig.10 middle:

[0153]

[0154] By analogy, we can obtain the set of boundary points of the drivable area.

[0155] Step S4033: filtering the set of boundary points of the drivable area, and generating the boundary of the drivable area at the current moment according to the filtering process.

[0156] It should be noted that Kalman filtering can provide a more accurate and robust state estimation by comprehensively considering the predicted and observed values. Perform Kalman filtering to finally determine the boundary of the drivable area at the current moment Γ k .

[0157] Specifically, the above Kalman filtering process is as follows:

[0158] (1) Prediction: Based on the above set of boundary points of the drivable area Denote it as the one-step prediction of Kalman filtering, and the process noise as R k , the variance of the output at time k-1 is Φ k-1 ;and

[0159] Among them, R k and Φ k-1 Both are diagonal matrices of [Key_X×Key_Y].

[0160] (2) Update: Let the measurement noise be Q k , Q k It is also a diagonal matrix of [Key_X×Key_Y].

[0161] The Kalman gain is calculated by the following formula:

[0162]

[0163]

[0164] Since the longitudinal position P i,x,Γ Keep consistent with the measurement, do not update this item, only the lateral position P i,y,Γ Update and express it as: P of the i-th item at time k i,y,Γ The update process can be expressed as follows:

[0165]

[0166] Repeat the above prediction-update steps to obtain the current drivable area boundary Γ k .

[0167] In addition, you can also refer to Fig.11 The process of determining the boundary of the drivable area at the moment according to the constructed current occupancy grid map is described below. Fig.11 is the transformation graph of the drivable area boundary.

[0168] Fig.11 In the figure, 11-a is the current occupied grid map obtained after numerical filling (taking one side boundary as an example), 11-b is the center point position extracted from each occupied grid, 11-c is the driving area boundary point set at the current moment, and 11-d is the driving area boundary at the current moment obtained by updating the driving area boundary point set at the previous moment with the driving area boundary point set at the current moment.

[0169] In this embodiment, the current occupied grid map is obtained by filling the four types of data point sets into the initial grid map, thereby extracting the center point position of each occupied upper grid to obtain the current driving area boundary point set, which not only takes into account the current vehicle's motion state, but also takes into account lane lines, curbs, obstacles, traffic targets and other information to flexibly generate driving boundaries, further improving the vehicle's ability to cope with different scenarios. The driving area boundary point set at the previous moment is linearly interpolated and Kalman filtering is used to finally generate the current drivable area boundary, which can comprehensively consider the uncertainty of the predicted value and the observed value, so as to better adapt to environmental changes and improve the stability of the generated drivable area boundary.

[0170] Further, you can also refer to Fig.12 The driving boundary generated by the method of the present application is compared with the driving boundary generated by the traditional method. Fig.12 This is a schematic diagram comparing the boundary generation results of the vehicle drivable area boundary generation method and the traditional method.

[0171] exist Fig.12 In the example, the scene shown in 12-A is tested; the sampling results of the scene data including lane data and obstacle target data are shown in 12-B.

[0172] Traditional methods completely rely on lane line data for boundary extraction. The driving boundary extraction results are as follows: Fig.12 -C; the driving boundary extraction result of the present application method is shown as Fig.12 -D as shown;

[0173] The boundary extraction results of the traditional method and the method of this application are compared, such as Fig.12 -E as shown.

[0174] Since the method of the present application takes into account all currently available perception data in a unified manner, compared with traditional methods, it can further improve the flexibility of safety boundaries, thereby compensating for the current problem of abnormal vehicle driving trajectory control caused by over-reliance on lane line data in different scenarios, especially in scenarios where lane line recognition is difficult.

[0175] It should be noted that the above examples are only used to understand the present application and do not constitute a limitation on the method for generating the boundary of the vehicle drivable area of ​​the present application. More simple transformations based on this technical concept are all within the scope of protection of the present application.

[0176] The present application also provides a device for generating a boundary of a vehicle drivable area. Fig.13 , Fig.13 This is a schematic diagram of the module structure of the first embodiment of the vehicle drivable area boundary generation device of the present application, the device comprises:

[0177] The data acquisition module 1301 is used to acquire the lane data and obstacle target data of the current vehicle in real time;

[0178] The data processing module 1302 is used to sample the lane data to obtain a lane discrete point set; sample the obstacle target data to obtain an obstacle target point trace set;

[0179] The boundary generation module 1303 is used to construct a current occupied grid map according to the lane discrete point set and the obstacle target point trace set, and generate a drivable area boundary according to the current occupied grid map.

[0180] The lane data and the obstacle target data are collected by the forward-looking sensor and millimeter-wave radar installed on the current vehicle. The lane data includes: lane line data and curb data, and the obstacle target data includes obstacle data and traffic target data; the lane discrete point set includes: lane line discrete point set and curb point trace set; the obstacle target point trace set includes: obstacle contour point trace set and target contour point trace set.

[0181] Furthermore, the data processing module 1302 is also used to determine the lane sampling step and the number of lane sampling points based on a preset discrete sampling rule; sample the lane data according to the lane sampling step and the number of lane sampling points to obtain a lane discrete point set.

[0182] Furthermore, the data processing module 1302 is also used to determine the target position and target size of the obstacle target based on the obstacle target data; determine the obstacle target sampling interval based on the target position and the target size based on a preset contour sampling rule; sample the obstacle target according to the obstacle target sampling interval to obtain an obstacle target point trace set.

[0183] Furthermore, the data processing module 1302 is also used to determine the nearest recognizable edge of each obstacle target according to the target position, and determine whether the nearest recognizable edge is a lateral nearest edge; if so, perform a nearest edge conversion operation on the nearest recognizable edge, and update the target position according to the conversion result, the nearest edge conversion operation being used to convert the nearest recognizable edge into a rear-preserving nearest edge; if not, when the nearest recognizable edge is a front-preserving nearest edge and the obstacle target is a lateral target, perform a nearest edge conversion operation on the nearest recognizable edge, and update each target position according to the conversion result; based on the preset contour sampling rule, determine the obstacle target sampling interval according to the updated target position and the target size.

[0184] Furthermore, the boundary generation module 1303 is also used to numerically fill the initial grid map with the lane discrete point set and the obstacle target point trace set according to a preset filling rule to obtain a current occupied grid map, wherein the initial grid map is generated based on a pre-configured map size; extract the center point position of each occupied grid in the current occupied grid map, and generate a driving area boundary point set at the current moment based on each center point position; update the driving area boundary point set at the previous moment based on the driving area boundary point set at the current moment, and obtain the drivable area boundary at the current moment based on the update result.

[0185] Furthermore, the boundary generation module 1303 is also used to obtain a set of boundary points of the driving area at the previous moment, and update the set of boundary points of the driving area at the previous moment according to the driving displacement of the current vehicle from the previous moment to the current moment to obtain a set of candidate boundary points of the drivable area; perform linear interpolation on the candidate set of boundary points of the drivable area according to the set of boundary points of the driving area at the current moment to obtain a set of boundary points of the drivable area; perform filtering processing on the set of boundary points of the drivable area, and generate the boundary of the drivable area at the current moment according to the filtering processing.

[0186] This embodiment uniformly processes the data of lane lines, curbs, obstacles, traffic targets, etc. that can be obtained by existing visual perception, and constructs an occupancy grid map to confirm the boundaries of the current vehicle's drivable area. This allows the vehicle to choose a reasonable control trajectory when using ADAS for assisted driving to cope with scenarios such as lane line loss, vehicle blocking lane lines, road cones, etc., thereby further optimizing and improving user experience and vehicle safety.

[0187] The present application also provides a vehicle, on which an advanced driver assistance system is installed, the advanced driver assistance system comprising: at least one processor; and 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 generating the boundary of a vehicle drivable area in the above-mentioned embodiment one.

[0188] The present application also provides a computer-readable storage medium having computer-readable program instructions (ie, computer programs) stored thereon, the computer-readable program instructions being used to execute the method for generating a boundary of a vehicle drivable area in the above-mentioned embodiment.

[0189] The computer-readable storage medium provided in the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, systems or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, 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 above. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination of the above.

[0190] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the presence of other elements in the process, method, article or system including the element.

[0191] The above are only preferred embodiments of the present application, and the scope of the patent of the present application is not limited thereto. Any equivalent structure or equivalent process transformation made by using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the scope of patent protection of the present application. The above are only preferred embodiments of the present application, and the scope of the patent of the present application is not limited thereto. Any equivalent structure or equivalent process transformation made by using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the scope of patent protection of the present application.

Claims

1. A method for generating a boundary of a vehicle drivable area, characterized in that: The method comprises: Obtain the lane data and obstacle target data of the current vehicle in real time; Sampling the lane data to obtain a lane discrete point set; Sampling the obstacle target data to obtain an obstacle target point trace set; A current occupied grid map is constructed according to the lane discrete point set and the obstacle target point trace set, and a drivable area boundary is generated according to the current occupied grid map.

2. The method according to claim 1, characterized in that The step of sampling the lane data to obtain a lane discrete point set includes: Based on the preset discrete sampling rules, determine the lane sampling step length and the number of lane sampling points; The lane data is sampled according to the lane sampling step length and the number of lane sampling points to obtain a lane discrete point set.

3. The method according to claim 1, characterized in that The step of sampling the obstacle target data to obtain an obstacle target point trace set includes: Determine the target position and target size of the obstacle target according to the obstacle target data; Based on a preset contour sampling rule, determining an obstacle target sampling interval according to the target position and the target size; The obstacle target is sampled according to the obstacle target sampling interval to obtain an obstacle target point trace set.

4. The method according to claim 3, characterized in that After the step of determining the obstacle target sampling interval according to the target position and the target size based on the preset contour sampling rule, the method further includes: Determine the nearest identifiable edge of each obstacle target according to the target position, and judge whether the nearest identifiable edge is the lateral nearest edge; If yes, a nearest edge conversion operation is performed on the nearest identified edge, and the target position is updated according to the conversion result, wherein the nearest edge conversion operation is used to convert the nearest identified edge into a rear-preserved nearest edge; If not, when the nearest identified edge is the nearest edge of the front guard and the obstacle target is a lateral target, a nearest edge conversion operation is performed on the nearest identified edge, and each target position is updated according to the conversion result; Accordingly, the step of determining the obstacle target sampling interval according to the target position and the target size based on the preset contour sampling rule includes: Based on the preset contour sampling rule, the obstacle target sampling interval is determined according to the updated target position and the target size.

5. The method according to claim 1, characterized in that The step of constructing a current occupied grid map according to the lane discrete point set and the obstacle target point trace set, and generating a drivable area boundary according to the current occupied grid map, further includes: Numerically filling an initial grid map with the lane discrete point set and the obstacle target point trace set according to a preset filling rule to obtain a current occupied grid map, wherein the initial grid map is generated based on a pre-configured map size; Extracting the center point position of each occupied grid in the current occupied grid map, and generating a set of boundary points of the driving area at the current moment according to the center point position; The driving area boundary point set at the previous moment is updated according to the driving area boundary point set at the current moment, and the drivable area boundary at the current moment is obtained according to the update result.

6. The method according to claim 5, characterized in that The step of updating the driving area boundary point set at the previous moment according to the driving area boundary point set at the current moment, and obtaining the drivable area boundary at the current moment according to the update result, comprises: Acquire a set of boundary points of the driving area at a previous moment, and update the set of boundary points of the driving area at a previous moment according to the driving displacement of the current vehicle from the previous moment to the current moment, so as to obtain a set of boundary points of a candidate drivable area; Performing linear interpolation on the candidate drivable area boundary point set according to the drivable area boundary point set at the current moment to obtain a drivable area boundary point set; The set of boundary points of the drivable area is filtered, and the boundary of the drivable area at the current moment is generated according to the filtering process.

7. The method according to any one of claims 1 to 6, characterized in that: The lane data and the obstacle target data are collected by a forward-looking sensor and a millimeter-wave radar installed on the current vehicle, the lane data includes: lane line data and curb data, and the obstacle target data includes obstacle data and traffic target data; Correspondingly, the lane discrete point set includes: a lane line discrete point set and a roadside point trace set; the obstacle target point trace set includes: an obstacle contour point trace set and a target contour point trace set.

8. A device for generating a boundary of a vehicle drivable area, characterized in that: The device comprises: Data acquisition module, used to obtain the lane data of the current vehicle and obstacle target data in real time; A data processing module, used to sample the lane data to obtain a lane discrete point set; and to sample the obstacle target data to obtain an obstacle target point trace set; The boundary generation module is used to construct a current occupied grid map according to the lane discrete point set and the obstacle target point trace set, and generate a drivable area boundary according to the current occupied grid map.

9. A storage medium, characterized in that: The storage medium stores a vehicle drivable area boundary generation program, and when the vehicle drivable area boundary generation program is executed by the processor, the steps of the vehicle drivable area boundary generation method according to any one of claims 1 to 7 are implemented.

10. A vehicle, characterized in that: The vehicle is equipped with an advanced driving assistance system, which includes: a memory, a processor, and a vehicle drivable area boundary generation program stored in the memory and executable on the processor, wherein the vehicle drivable area boundary generation program is configured to implement the steps of the vehicle drivable area boundary generation method as described in any one of claims 1 to 7.

Images