Vehicle lateral positioning method, device and system

Abstract

The present invention provides a vehicle lateral positioning method, device and system, wherein the method comprises: when a vehicle is traveling in a tunnel road, obtaining lane line information of the lane currently located by the vehicle, passable area position information and first environmental point cloud data of the tunnel road; calculating the centerline position information and lane width information of the lane according to the lane line information; determining the boundary information of the tunnel road according to the centerline position information, the passable area position information and the first environmental point cloud data; determining the lateral position information of the vehicle in the tunnel road according to the centerline position information, the width information and the boundary information. The method realizes the lateral positioning of the vehicle through the fusion of visual detection and laser radar, avoiding the production of high-precision maps of complex high-dimensional features and databases of environmental feature values. In addition, the algorithm adopted by the method is concise, efficient, and has strong real-time performance, and has stronger robustness than complex feature matching algorithms.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle positioning, and in particular to a vehicle lateral positioning method, device and system. Background Art

[0002] With the development of smart cars, technologies related to autonomous driving have received widespread attention in the industry. The software modules of autonomous driving mainly include perception module, navigation and positioning module, decision-making and planning module, motion control module, etc. Among them, the positioning module provides centimeter-level positioning for smart cars in real time to ensure that the vehicle can drive safely and automatically on the road.

[0003] When smart cars drive automatically in tunnels, accurate positioning is a difficult problem for autonomous driving technology. Usually, vehicles can drive normally in the longitudinal direction based on information such as lane lines and traffic areas, so the longitudinal positioning accuracy is usually not required to be high; however, lateral positioning requires accurate confirmation of the current lane and centimeter-level positioning of the lane to ensure safe driving. Therefore, solving the lateral positioning of smart cars in tunnels is the core of ensuring safe autonomous driving of vehicles in tunnels.

[0004] Currently, autonomous driving positioning mainly includes global satellite navigation system (GNSS) positioning, inertial unit (IMU) positioning and high-precision map positioning. GPS positioning, which is common in global satellite navigation system (GNSS) positioning, achieves the positioning of the target by calculating the distance between multiple satellites and the vehicle. The GPS ranging method is simple, but it will produce measurement errors due to factors such as the speed of light propagating in different media. In order to reduce the GPS positioning error, differential GPS technology is used to calibrate the error by receiving and comparing the GPS data of the base station to achieve centimeter-level positioning. This positioning method has a good positioning effect in an open environment, but has poor positioning accuracy or even fails in an environment with obstruction or electronic interference. Therefore, it is not suitable for positioning in a tunnel environment. The method of obtaining positioning using an inertial unit (IMU) will produce cumulative errors and needs to be calibrated by GPS in a timely manner. When driving in a long tunnel, the positioning accuracy of the IMU will drop rapidly until it fails, and it is impossible to achieve accurate lateral positioning of the vehicle in the tunnel for a long time. The high-precision map positioning method is to formulate a high-precision map, rely on laser radar or visual detection, and match the high-precision map features to achieve accurate positioning. This positioning method relies on the characteristics of the current environment. When a smart car is driving in a tunnel, the differences in environmental characteristics are often small, which poses certain challenges to feature matching and positioning. Therefore, it is also difficult to stably and accurately give the lateral positioning of the vehicle in the tunnel. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide a vehicle lateral positioning method, device and system to solve the problem of inaccurate lateral positioning of the vehicle when the vehicle is traveling in a tunnel in the prior art.

[0006] In order to solve the above technical problems, the present invention provides a vehicle lateral positioning method in a first aspect, comprising:

[0007] When the vehicle is traveling in a tunnel road, the lane line information of the lane currently located by the vehicle, the position information of the passable area, and the first environmental point cloud data of the tunnel road are obtained;

[0008] Calculating the centerline position information of the lane and the lane width information according to the lane line information;

[0009] Determining boundary information of the tunnel road according to the centerline position information, the traversable area position information, and the first environmental point cloud data;

[0010] The lateral position information of the vehicle in the tunnel road is determined according to the centerline position information, the lane width information and the boundary information.

[0011] In a specific implementation, the determining the boundary information of the tunnel road according to the centerline position information, the traversable area position information and the first environmental point cloud data specifically includes:

[0012] Determine the position information of the search starting point according to the passable area position information and the centerline position information, wherein the search starting point is an initial search position of the tunnel road boundary point;

[0013] Determine the position information of a first candidate point of the tunnel road boundary point according to the first environmental point cloud data;

[0014] Determine the tunnel road boundary point information according to the centerline position information, the search starting point position information and the first candidate point position information;

[0015] The road boundary curve information is determined according to the tunnel road boundary point information.

[0016] In a specific implementation, the determining the location information of the search starting point according to the passable area location information and the centerline location information specifically includes:

[0017] Projecting the centerline position information and the passable area position information into a two-dimensional grid map respectively, and correspondingly obtaining the centerline position information in the two-dimensional grid map and the passable area position point information in the two-dimensional grid map;

[0018] According to the relative position relationship between the passable area position points in the two-dimensional grid map and the center line in the two-dimensional grid map, the passable area position points in the two-dimensional grid map are divided into left passable area position points and right passable area position points;

[0019] Calculating a first distance value between the position point in the left passage area and the origin of the vehicle coordinate system, and calculating a second distance value between the position point in the right passage area and the origin of the vehicle coordinate system;

[0020] The passable area position point corresponding to the minimum value in the first distance values ​​is used as the search starting point for the left road boundary point, and the passable area position point corresponding to the minimum value in the second distance values ​​is used as the search starting point for the right road boundary point.

[0021] In a specific implementation, determining the position information of the first candidate point of the tunnel road boundary point according to the first environmental point cloud data specifically includes:

[0022] Convert the first environmental point cloud data into a coordinate system to obtain second environmental point cloud data in a vehicle coordinate system;

[0023] Projecting the second environment point cloud data into the two-dimensional grid map, obtaining the third environment point cloud data in the two-dimensional grid map, and recording the height value of the third environment point cloud;

[0024] Selecting an environmental point cloud to be analyzed from the third environmental point cloud, respectively calculating a first difference between a height value of the environmental point cloud to be analyzed and a height value of a third environmental point cloud of a set area, determining whether the first difference is greater than a first set threshold, and if so, counting once to obtain a cumulative count corresponding to the set area;

[0025] It is determined whether the accumulated count is greater than a count threshold; if so, the first environment point cloud corresponding to the environment point cloud to be analyzed is the first candidate point.

[0026] In a specific implementation, determining the tunnel road boundary point information according to the location information of the search starting point and the location information of the first candidate point specifically includes:

[0027] Processing the first candidate point to obtain a second candidate point, and projecting the second candidate point into the two-dimensional grid map;

[0028] According to the relative position relationship between the second candidate point in the two-dimensional grid map and the center line in the two-dimensional grid map, the second candidate point is divided into a left road candidate boundary point and a right road candidate boundary point;

[0029] The left road boundary point information is obtained according to the left road candidate boundary point and the search starting point of the left road boundary point, and the right road boundary point information is obtained according to the right road candidate boundary point and the search starting point of the right road boundary point.

[0030] In a specific implementation, the processing of the first candidate point to obtain the second candidate point specifically includes:

[0031] Traversing the first candidate points, obtaining first candidate points marked by the same laser beam of the laser radar, and clustering the first candidate points marked by the same laser beam;

[0032] Calculate a third distance value between each first candidate point in any class and the horizontal direction of the laser radar;

[0033] Obtaining a minimum distance value and a maximum distance value among the third distance values ​​in the arbitrary class, and calculating a second difference between the minimum distance value and the maximum distance value;

[0034] It is determined whether the second difference is greater than a second set threshold value. If so, the first candidate point in the arbitrary class is the second candidate point.

[0035] In a specific implementation, the obtaining of the left road boundary point information according to the left road candidate boundary point and the search starting point of the left road boundary point specifically includes:

[0036] Taking the search starting point of the left road boundary point as the initial search center, searching the two-dimensional grid map line by line according to the set rules;

[0037] If the left road candidate boundary point is searched, the position coordinates of the searched left road candidate boundary point are obtained, a first slope value of the lane centerline at the same ordinate as the searched left road candidate boundary point is calculated, next search starting position information is determined according to the position coordinates of the searched left road candidate boundary point and the first slope value, and the next search starting position is used as the first current search center, and the search is performed according to the set rules until the search of all rows of the two-dimensional grid map is completed, wherein the searched left road candidate boundary point is the left road boundary point;

[0038] If the left road candidate boundary point is not found, a second slope value of the lane centerline at the same ordinate as the current search center is calculated, and next search start position information is determined according to the coordinate position information of the current search center and the second slope value, and the next search start position is used as the second current search center, and the two-dimensional grid map is searched row by row according to the set rules until all rows of the two-dimensional grid map are searched, wherein the searched left road candidate boundary point is the left road boundary point; and / or

[0039] The obtaining of the right road boundary point information according to the right road candidate boundary point and the search starting point of the right road boundary point specifically includes:

[0040] Taking the search starting point of the right road boundary point as the initial search center, searching the two-dimensional grid map line by line according to the set rules;

[0041] If the right road candidate boundary point is searched, the position coordinates of the searched right road candidate boundary point are obtained, the third slope value of the lane centerline at the same ordinate as the searched right road candidate boundary point is calculated, the next search starting position is determined according to the position coordinates of the searched right road candidate boundary point and the third slope value, and the next search starting position is used as the third current search center, and the search is performed according to the set rules until the search of all rows of the two-dimensional grid map is completed, wherein the searched right road candidate boundary point is the right road boundary point;

[0042] If the right road candidate boundary point is not searched, the fourth slope value of the lane centerline at the same vertical coordinate as the current search center is calculated, and the next search starting position information is determined according to the coordinate position information of the current search center and the fourth slope value. The next search starting position is used as the fourth current search center, and the two-dimensional grid map is searched row by row according to the set rules until the search of all rows of the two-dimensional grid map is completed, wherein the searched right road candidate boundary point is the right road boundary point.

[0043] In a specific implementation, determining the next search start position information according to the searched left road candidate boundary point position coordinates and the first slope value specifically includes:

[0044] Determine the ordinate value of the searched left road candidate boundary point as the ordinate value of the next search position;

[0045] A first product is obtained by multiplying the horizontal coordinate value of the searched left road candidate boundary point by the first slope value, and a first sum value of the first product and the horizontal coordinate of the searched left road candidate boundary point is calculated, wherein the first sum value is the horizontal coordinate value of the next search starting position.

[0046] In a specific implementation, the calculating of the first slope value of the lane centerline at the same ordinate as the searched left road candidate boundary point specifically includes:

[0047] Acquire a first lateral coordinate value of a lane centerline point at the same ordinate as the searched left road candidate boundary point;

[0048] Obtaining a second lateral coordinate value of a lane centerline point in a row of grids below the lane centerline point;

[0049] A third difference between the second horizontal coordinate value and the first horizontal coordinate value is calculated, and the third difference is the first slope value.

[0050] In a specific implementation, determining the road boundary curve information according to the tunnel road boundary point information specifically includes:

[0051] The left road boundary points are fitted to obtain left road boundary curve information, and the right road boundary points are fitted to obtain right road boundary curve information.

[0052] In a specific implementation, determining the lateral position information of the vehicle in the tunnel road according to the centerline position information, the width information and the boundary information specifically includes:

[0053] Calculating a fourth distance value from the vehicle coordinate origin to the left road boundary curve and a fifth distance value from the vehicle coordinate origin to the right road boundary curve;

[0054] Calculate a total width value of the tunnel road according to the fourth distance value and the fifth distance value, and calculate a total number of lanes of the tunnel road according to the total width value and the width information;

[0055] determining the lane in which the vehicle is traveling according to the fourth distance value and the width information or determining the lane in which the vehicle is traveling according to the fifth distance value and the width information;

[0056] A sixth distance value between the vehicle coordinate origin and the center line is calculated, and a lateral position coordinate of the vehicle in the lane in which the vehicle is traveling is determined according to the sixth distance value.

[0057] A second aspect of the present invention provides a vehicle lateral positioning device, comprising:

[0058] An acquisition unit, used for acquiring lane line information of a lane currently located by the vehicle, position information of a passable area, and first environmental point cloud data of the tunnel road when the vehicle is traveling in the tunnel road;

[0059] A calculation unit, used for calculating the center line position information of the lane and the width information of the lane according to the lane line information;

[0060] A boundary information determination unit, configured to determine the boundary information of the tunnel road according to the lane centerline position information, the passable area position information and the first environmental point cloud data;

[0061] A lateral position determination unit is used to determine the lateral position information of the vehicle in the tunnel road according to the centerline position information, the width information and the boundary information.

[0062] In a specific implementation, the boundary information determination unit includes:

[0063] A search starting point determination unit, configured to determine the position information of the search starting point according to the passable area position information and the center line position information, wherein the search starting point is an initial search position of a tunnel road boundary point;

[0064] A first candidate point determination unit, configured to determine position information of a first candidate point of the tunnel road boundary point according to the first environmental point cloud data;

[0065] a road boundary point information determining unit, configured to determine the tunnel road boundary point information according to the position information of the search starting point and the position information of the first candidate point;

[0066] A road boundary determination unit is used to determine the road boundary information according to the tunnel road boundary point information.

[0067] In a specific implementation, the road boundary point information determination unit specifically includes:

[0068] A second candidate point determination unit, configured to process the first candidate point to obtain a second candidate point, and project the second candidate point into the two-dimensional grid map;

[0069] a candidate road boundary point determination unit, configured to divide the second candidate point in the two-dimensional grid map into a left road candidate boundary point and a right road candidate boundary point according to a relative position relationship between the second candidate point in the two-dimensional grid map and a center line in the two-dimensional grid map;

[0070] A left and right road boundary point determination unit, configured to obtain the left road boundary point information according to the left road candidate boundary point and the search starting point of the left road boundary point, and to obtain the right road boundary point information according to the right road candidate boundary point and the search starting point of the right road boundary point;

[0071] The fitting unit is used to fit the left road boundary point information and the right road boundary point information respectively, and correspondingly obtain the left road boundary curve information and the right road boundary curve information.

[0072] The third aspect of the present invention provides a vehicle lateral positioning system, comprising: a visual detection device, a laser radar and the aforementioned vehicle lateral positioning device, wherein:

[0073] The visual detection device is used to collect and process image information of the current driving direction of the vehicle when the vehicle is driving in the tunnel road, and output lane line information of the lane where the vehicle is currently located and passable area position information;

[0074] The laser radar device is used to collect environmental information of the tunnel road and output first environmental point cloud data of the tunnel road.

[0075] In a specific embodiment, the visual detection device is an EPM system, and the laser radar is a 64-line laser radar.

[0076] The beneficial effect of the embodiment of the present invention is that the method obtains the lane line information of the current lane of the vehicle in the tunnel road, the position information of the passable area and the environmental point cloud data of the tunnel road, and determines the boundary information of the tunnel road according to the lane line information, the position information of the passable area and the environmental point cloud data, and determines the lateral position information of the vehicle according to the lane line information and the boundary information. The method realizes the lateral positioning of the vehicle through the fusion of visual detection and laser radar, avoids the production of high-precision maps with complex high-dimensional features and databases of environmental feature values. In addition, the algorithm used in the method is simple, efficient, and has strong real-time performance, and has stronger robustness than complex feature matching algorithms. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] In order to more clearly illustrate the embodiments of the present invention 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, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0078] Figure 1 1 is a flow chart of a vehicle lateral positioning method according to a first embodiment of the present invention;

[0079] Figure 2 is a more specific flow chart of step S3 of a vehicle lateral positioning method according to the first embodiment of the present invention;

[0080] Figure 3 is a more specific flowchart of step S31 of a vehicle lateral positioning method according to the first embodiment of the present invention;

[0081] Figure 4 is a more specific flowchart of step S32 of a vehicle lateral positioning method according to the first embodiment of the present invention;

[0082] Figure 5 is a more specific flowchart of step S33 of a vehicle lateral positioning method according to the first embodiment of the present invention;

[0083] Figure 6 This is a schematic diagram of clustering first candidate points of a vehicle lateral positioning method according to the first embodiment of the present invention;

[0084] Figure 7 is a structural schematic diagram of a vehicle lateral positioning device according to a second embodiment of the present invention;

[0085] Figure 8 It is a structural schematic diagram of a vehicle lateral positioning system according to Embodiment 3 of the present invention. DETAILED DESCRIPTION

[0086] The following descriptions of the embodiments refer to the accompanying drawings to illustrate specific embodiments in which the present invention may be implemented.

[0087] The following reference Figure 1 As shown, the first embodiment of the present invention provides a vehicle lateral positioning method, comprising the following steps:

[0088] S1. When a vehicle is traveling in a tunnel road, lane line information of a lane currently located by the vehicle, location information of a passable area, and first environmental point cloud data of the tunnel road are obtained.

[0089] Specifically, the autonomous vehicle is equipped with Mobileye's EPM system. The EPM system obtains image information of the road where the vehicle is currently located through the vehicle-mounted camera. The EPM system processes the image information and outputs the lane line information and the passable area location information of the lane where the vehicle is currently traveling. The lane line information includes the polynomial coefficients of the lane line and the longitudinal distance of the lane line, and the passable area location information is discrete location point information. When there is a vehicle in front of the vehicle, the discrete location point information is composed of the discrete road point location information and the boundary point location information of the position of the vehicle in front.

[0090] Specifically, a vehicle-mounted laser radar is used to scan the environmental information of the tunnel road where the vehicle is located in real time to obtain a data packet in a pcap format, wherein the data packet is polar coordinate environmental point cloud data centered on the laser radar, namely, first environmental point cloud data.

[0091] S2. Calculate the centerline position information of the lane and the lane width information according to the lane line information.

[0092] Specifically, it is assumed that the left and right lane lines of the lane use the cubic curve equation x=a3y 3 +a2y 2 +a1y+a0 represents, y represents the longitudinal distance from the lane line to the origin of the vehicle coordinate system, and x represents the lateral position of the lane line, where a positive value of x indicates that the lane line is on the right side of the vehicle, and a negative value of x indicates that the lane line is on the left side of the vehicle, and a3, a2, a1, and a0 are corresponding coefficients. According to the longitudinal distance range of the left and right lane lines output by the EPM system, the lateral position of the lane line corresponding to the longitudinal range is calculated by the above equation. At the same longitudinal distance, the lateral positions of the left and right lane lines are xL and xR respectively, and the center line of the lane is expressed as (xL+xR) / 2. The lane line width is expressed as |xL-xR|. By traversing the longitudinal distance range of the left and right lane lines, the center line position of the lane within the range is obtained. At the same time, the lane line width is averaged to obtain the lane line width.

[0093] S3. Determine the boundary information of the tunnel road according to the centerline position information, the traversable area position information and the first environmental point cloud data.

[0094] like Figure 2 As shown, in a specific implementation, the step S3 specifically includes:

[0095] S31. Determine the search starting point position information according to the passable area position information and the center line position information, wherein the search starting point is an initial search position of a tunnel road boundary point.

[0096] S32. Determine the position information of a first candidate point of the tunnel road boundary point according to the first environmental point cloud data.

[0097] S33. Determine the tunnel road boundary point information according to the centerline position information, the search start point position information and the position information of the first candidate point.

[0098] S34. Determine the road boundary curve information according to the tunnel road boundary point information.

[0099] like Figure 3 As shown, in a specific implementation, the step S31 specifically includes:

[0100] S311, projecting the center line position information and the passable area position information into a two-dimensional grid map respectively, and correspondingly obtaining the center line position information in the two-dimensional grid map and the passable area position point information in the two-dimensional grid map.

[0101] Take the center line of the lane as an example. Assume that the position coordinates of the center line of the lane are (x, y), where x is the horizontal coordinate of the center line of the lane, and y is the vertical coordinate of the center line of the lane. Its corresponding coordinates in the two-dimensional grid map are (u, v), where u is the horizontal coordinate of the center line of the lane in the two-dimensional grid map, and v is the vertical coordinate of the center line of the lane in the two-dimensional grid map. The calculation formula is:

[0102] u=x / R x +U

[0103] v=y / R y +V

[0104] Among them, R x is the lateral resolution of the two-dimensional grid, R y is the longitudinal resolution of the two-dimensional grid, U is the horizontal coordinate of the origin of the vehicle coordinate system in the two-dimensional grid map, and V is the vertical coordinate of the origin of the vehicle coordinate system in the two-dimensional grid map. In a specific implementation, the size of the two-dimensional grid map is 400×400, and the horizontal and vertical resolutions of the two-dimensional grid map are both 20 cm, then R x =R y =20cm; the vertical range of the two-dimensional grid map is 80 meters, and the horizontal range is 40 meters on both sides, so U=200, V=400.

[0105] According to the above calculation formula, the position information of the passable area can be projected into the two-dimensional grid map to obtain the position point information of the passable area in the two-dimensional grid map.

[0106] S312: According to the relative position relationship between the passable area position points in the two-dimensional grid map and the center line in the two-dimensional grid map, the passable area position points in the two-dimensional grid map are divided into left passable area position points and right passable area position points.

[0107] Specifically, according to whether the passable area position point in the two-dimensional grid map is on the left or right side of the lane centerline position in the two-dimensional grid map, the passable area position point in the two-dimensional grid map is correspondingly divided into a left passable area position point and a right passable area position point. When the passable area position point in the two-dimensional grid map is at the coordinate of the lane centerline in the two-dimensional grid map, it is a left passable area position point, and if it is on the right side, it is a right passable area position point.

[0108] S313: Calculate a first distance value between the position point in the left traffic area and the origin of the vehicle coordinate system, and calculate a second distance value between the position point in the right traffic area and the origin of the vehicle coordinate system.

[0109] In a specific embodiment, the origin of the vehicle coordinate system is the center point of the vehicle front, and the first distance value between each left traffic area position point and the center point of the vehicle front is calculated, and the second distance value between each right traffic area position point and the center point of the vehicle front is calculated.

[0110] S314: Use the passable area location point corresponding to the minimum value in the first distance values ​​as the search starting point for the left road boundary point, and use the passable area location point corresponding to the minimum value in the second distance values ​​as the search starting point for the right road boundary point.

[0111] like Figure 4 As shown, in a specific implementation, the step S32 specifically includes:

[0112] S321. Perform coordinate system conversion on the first environment point cloud data to obtain second environment point cloud data in a vehicle coordinate system.

[0113] Specifically, the first environmental point cloud data obtained by the laser radar is corrected and coordinate converted to obtain the second environmental point cloud data in the vehicle coordinate system.

[0114] S322: Project the second environment point cloud data into the two-dimensional grid map, obtain the third environment point cloud data in the two-dimensional grid map, and record the height value of the third environment point cloud.

[0115] The second environment point cloud data in the vehicle coordinate system is projected into the two-dimensional grid map to obtain the third environment point cloud data in the two-dimensional grid map, and the height value of the third environment point cloud is recorded.

[0116] S323. Select an environmental point cloud to be analyzed from the third environmental point cloud, calculate the first difference between the height value of the environmental point cloud to be analyzed and the height value of the third environmental point cloud of the set area, and determine whether the first difference is greater than a first set threshold. If so, count once to obtain the cumulative count corresponding to the set area.

[0117] S324: Determine whether the accumulated count is greater than a count threshold; if so, the first environment point cloud corresponding to the environment point cloud to be analyzed is the first candidate point.

[0118] Specifically, any point cloud is selected from the third environmental point cloud in the two-dimensional grid map as the point cloud to be analyzed, and the difference between the height value of the point cloud to be analyzed and the height value of each third environmental point cloud in the set area is calculated to determine whether the difference is greater than the first threshold value. If so, count once, obtain the cumulative count corresponding to the set area, and determine whether the cumulative count is greater than the counting threshold value. If greater than, the first environmental point cloud corresponding to the point cloud to be analyzed is the first candidate point. The set area is centered on the position where the point cloud data to be analyzed is projected into the two-dimensional grid map, and its size is an m×n area. Preferably, the value of m is between 10 and 20, and the value of n is between 10 and 20. Preferably, the value range of the first threshold is 10cm-20cm, and the counting threshold N can be N=0.25*m*n.

[0119] like Figure 5 As shown, in a specific implementation, the step S33 specifically includes:

[0120] S331. Process the first candidate point to obtain a second candidate point, and project the second candidate point into the two-dimensional grid map.

[0121] When the vehicle vibrates and pitches, the critical laser point cloud height will change suddenly, resulting in a significant height difference, which will cause some continuous incorrectly marked road boundary candidate points. Therefore, it is necessary to process the first candidate points to obtain the final road boundary candidate points, i.e., the second candidate points.

[0122] Traverse the first candidate points, obtain the first candidate points marked by the same laser beam of the laser radar, and cluster the first candidate points marked by the same laser beam. Specifically, Figure 6As shown, two first candidate points at adjacent positions of the same laser beam mark are selected, assuming that they are A and B. The two first candidate points form an arc, and a 360-degree clockwise and counterclockwise search is performed along the arc. Assuming that the first candidate points C and D are in the search direction of the arc, C and D can be searched. According to the above method, the search of all first candidate points in the direction of the arc is completed, and all the searched first candidate points belong to the same class as the first candidate points A and B. Since the first candidate point E is not on the arc formed by the first candidate points A and B, the first candidate point E does not belong to the same class as the candidate point A. When the class search is completed, two first candidate points at adjacent positions are reselected and searched according to the above method until the clustering of all first candidate points is completed. For the first candidate point in each class, calculate the third distance value between the first candidate point in the class and the laser radar in the horizontal direction, obtain the minimum distance value and the maximum distance value of the third distance value in the class, calculate the second difference between the maximum distance value and the minimum distance value, and determine whether the second difference is greater than the second set threshold. If so, the first candidate point in the class is the second candidate point. Project the second candidate point data into the two-dimensional grid map to obtain the second candidate point data in the two-dimensional grid map. The second set threshold is preferably 20cm-40cm.

[0123] S332: According to the relative position relationship between the second candidate point in the two-dimensional grid map and the center line in the two-dimensional grid map, divide the second candidate point in the two-dimensional grid map into a left road candidate boundary point and a right road candidate boundary point.

[0124] Specifically, taking the vehicle forward direction as the positive direction, the second candidate point in the two-dimensional grid map is divided into a left road candidate boundary point and a right road candidate boundary point according to whether the second candidate point in the two-dimensional grid map is on the left side or the right side of the lane centerline. If the second candidate point in the two-dimensional grid map is on the left side of the lane centerline, it is a left road candidate boundary point, and if the second candidate point in the two-dimensional grid map is on the right side of the lane centerline, it is a right road candidate boundary point.

[0125] S333: Obtain the left road boundary point information according to the left road candidate boundary point and the search starting point of the left road boundary point, and obtain the right road boundary point information according to the right road candidate boundary point and the search starting point of the right road boundary point.

[0126] Specifically, when searching for a left road boundary point, the search starting point of the left road boundary point is used as the initial search center, and the two-dimensional grid map is searched row by row according to the set rules. If the left road candidate boundary point is found, the position coordinates of the searched left road candidate boundary point are obtained, and the first slope value of the lane centerline at the same vertical coordinate as the searched left road candidate boundary point is calculated. The next search starting position information is determined according to the position coordinates of the searched left road candidate boundary point and the first slope value, and the next search starting position is used as the first current search center, and the search is performed according to the set rules until the completion of the search. A search of all rows of a two-dimensional grid map, wherein the searched left road candidate boundary point is the left road boundary point, and if the left road candidate boundary point is not searched, a second slope value of a lane centerline at the same ordinate as the current search center is calculated, and next search starting position information is determined according to the coordinate position information of the current search center and the second slope value, and the next search starting position is used as the second current search center, and the two-dimensional grid map is searched row by row according to set rules until the search of all rows of the two-dimensional grid map is completed, wherein the searched left road candidate boundary point is the left road boundary point.

[0127] Specifically, when searching for a right road boundary point, the search starting point of the right road boundary point is used as the initial search center, and the two-dimensional grid map is searched row by row according to the set rules. If the right road candidate boundary point is searched, the position coordinates of the searched right road candidate boundary point are obtained, and the third slope value of the lane centerline at the same vertical coordinate as the searched right road candidate boundary point is calculated. The next search starting position is determined according to the position coordinates of the searched right road candidate boundary point and the third slope value, and the next search starting position is used as the third current search center, and the search is performed according to the set rules until the completion of the search. A search of all rows of a two-dimensional grid map, wherein the searched right road candidate boundary point is the right road boundary point, and if the right road candidate boundary point is not searched, a fourth slope value of a lane centerline at the same ordinate as a current search center is calculated, and next search start position information is determined according to the coordinate position information of the current search center and the fourth slope value, and the next search start position is used as the fourth current search center, and the two-dimensional grid map is searched row by row according to set rules until the search of all rows of the two-dimensional grid map is completed, wherein the searched right road candidate boundary point is the right road boundary point.

[0128] Specifically, after the candidate boundary points are divided into the left road candidate boundary points and the right road candidate boundary points by using the lane center line, the search starting point of the left road boundary points is near the left road boundary points, and the search starting point of the right road boundary points is near the right road boundary points. Therefore, the search starting point of the left road boundary points is used as the initial search starting point of the left road boundary points, and the search starting point of the right road boundary points is used as the initial search starting point of the right road boundary points. The two-dimensional grid map is searched row by row according to the set rules. The set rule refers to forming a line segment with a length of w with the current search center as the midpoint, and the straight line where the line segment is located is parallel to the straight line where the row of the two-dimensional grid map is located, and then searching row by row with the width of w. Preferably, w is 2 meters.

[0129] The following example illustrates that if the position coordinates of the searched road boundary candidate point are I[i0,j0], i0 is the horizontal coordinate, representing the column of the grid map, and j0 is the vertical coordinate, representing the row of the grid map, then the coordinates of the next search starting position I[i1,j1] are:

[0130] i1=i0

[0131] j1=k*j0+j0

[0132] k=u1-u0

[0133] Wherein, k is the slope, u0 is the horizontal coordinate of the lane centerline point at the next longitudinal distance from the searched road boundary candidate point, and u1 is the horizontal coordinate of the centerline point in the next row of the lane centerline point.

[0134] The first slope, the second slope, the third slope and the fourth slope are all calculated using the above k calculation method.

[0135] In a specific implementation, the step S34 is specifically: fitting the left road boundary points to obtain the left road boundary curve, and fitting the right road boundary points to obtain the right road boundary curve. The fitting can use a quadratic parabola model, a B-spline curve, a Bezier curve, etc.

[0136] S4. Determine the lateral position information of the vehicle on the road according to the centerline position information, the lane width information and the road boundary position information.

[0137] The total number of lanes of the tunnel road, the lane where the vehicle is currently located, and the lateral position of the vehicle in the lane where the vehicle is currently located are determined according to the centerline position information, the lane width information, the left road boundary curve, and the right road boundary curve.

[0138] Specifically, since the position of the vehicle coordinate origin in the two-dimensional grid map is fixed, after obtaining the left road boundary curve and the right road boundary curve of the tunnel road, the fourth distance L1 from the vehicle coordinate origin to the left road boundary curve and the fifth distance Lr from the vehicle coordinate origin to the right road boundary curve can be calculated, and the road width Tw of the tunnel road can be calculated according to the fourth distance L1 from the vehicle coordinate origin to the left road boundary curve and the fifth distance Lr from the vehicle coordinate origin to the right road boundary curve. The total number of lanes of the tunnel road can be obtained by dividing the tunnel width Tw by the current road width Lw. The specific lane in which the vehicle is located in the tunnel road can be determined according to the distance L1 from the vehicle coordinate origin to the left road boundary curve or the distance Lr from the vehicle coordinate origin to the right road boundary curve and the lane width. For example, the vehicle is traveling in the left lane 2. The lateral position of the vehicle in the lane in which it is traveling is determined according to the sixth distance from the vehicle coordinate origin to the center line of the lane. For example, the vehicle is 0.2 meters to the left of the center line of the left lane 2.

[0139] The vehicle lateral positioning method of the embodiment of the present invention obtains the lane line information of the current lane of the vehicle in the tunnel road, the position information of the passable area and the environmental point cloud data of the tunnel road, and determines the boundary information of the tunnel road based on the lane line information, the position information of the passable area and the environmental point cloud data, and determines the lateral position information of the vehicle based on the lane line information and the boundary information. The method realizes the lateral positioning of the vehicle through the fusion of visual detection and laser radar, avoids the production of high-precision maps with complex high-dimensional features and a database of environmental feature values. In addition, the algorithm used in the method is simple, efficient, and has strong real-time performance, and has stronger robustness than complex feature matching algorithms.

[0140] Embodiment 2 of the present invention further provides a vehicle lateral positioning device, such as Figure 7 As shown, the lateral positioning device 10 includes an acquisition unit 1, which is used to acquire the lane line information, the passable area position information and the first environmental point cloud data of the tunnel road of the lane where the vehicle is currently located when the vehicle is traveling in the tunnel road; a calculation unit 2, which is used to calculate the center line position information of the lane and the width information of the lane according to the lane line information; a boundary information determination unit 3, which is used to determine the boundary information of the tunnel road according to the center line position information of the lane, the passable area position information and the first environmental point cloud data; a lateral position determination unit 4, which is used to determine the lateral position information of the vehicle in the tunnel road according to the center line position information, the width information and the boundary information.

[0141] Among them, the boundary information determination unit 3 includes: a search starting point determination unit, used to determine the position information of the search starting point according to the passable area position information and the center line position information, and the search starting point is the initial search position of the tunnel road boundary point; a first candidate point determination unit, used to determine the position information of the first candidate point of the tunnel road boundary point according to the first environmental point cloud data; a road boundary point information determination unit, used to determine the tunnel road boundary point information according to the position information of the search starting point and the position information of the first candidate point; a road boundary determination unit, used to determine the road boundary information according to the tunnel road boundary point information.

[0142] The road boundary point information determination unit specifically includes: a second candidate point determination unit, used to process the first candidate point to obtain a second candidate point, and project the second candidate point into the two-dimensional grid map; a candidate road boundary point determination unit, used to divide the second candidate point into a left road candidate boundary point and a right road candidate boundary point according to the relative position relationship between the second candidate point in the two-dimensional grid map and the center line in the two-dimensional grid map; a left and right road boundary point determination unit, used to obtain the left road boundary point information according to the left road candidate boundary point and the search starting point of the left road boundary point, and to obtain the right road boundary point information according to the right road candidate boundary point and the search starting point of the right road boundary point; a fitting unit, used to fit the left road boundary point information and the right road boundary point information respectively, and correspondingly obtain the left road boundary curve information and the right road boundary curve information.

[0143] Embodiment 3 of the present invention further provides a vehicle lateral positioning system, such as Figure 8 As shown, the positioning system 100 includes: a visual detection device 20, a laser radar 30 and the aforementioned vehicle lateral positioning device, wherein the visual detection device 20 is used to collect and process image information of the current driving direction of the vehicle when the vehicle is traveling in a tunnel road, and output lane line information and passable area position information of the lane where the vehicle is currently located; the laser radar 30 is used to collect environmental information of the tunnel road and output first environmental point cloud data of the tunnel road.

[0144] The visual detection device 20 is a visual detection device that can stably output lane line and passable area information in a tunnel environment, and is preferably an EPM system.

[0145] The laser radar 30 is a multi-beam laser radar, such as a 4-beam laser radar, a 32-beam laser radar, a 64-beam laser radar, etc.

[0146] For the working principle of this embodiment and the beneficial effects it brings, please refer to the description of the first embodiment of the present invention, which will not be repeated here.

[0147] The above disclosure is only the preferred embodiment of the present invention, which certainly cannot be used to limit the scope of the present invention. Therefore, equivalent changes made according to the claims of the present invention are still within the scope of the present invention.

Claims

1. A vehicle lateral positioning method, characterized in that: include: When the vehicle is traveling in a tunnel road, the lane line information of the lane currently located by the vehicle, the position information of the passable area, and the first environmental point cloud data of the tunnel road are obtained; Calculating the centerline position information of the lane and the lane width information according to the lane line information; Determine the position information of the search starting point according to the centerline position information and the passable area position information, wherein the search starting point is an initial search position of the tunnel road boundary point; Determine the position information of a first candidate point of the tunnel road boundary point according to the first environmental point cloud data; Determine the tunnel road boundary point information according to the centerline position information, the search starting point position information and the first candidate point position information; determine the tunnel road boundary information according to the tunnel road boundary point information; Determine the lateral position information of the vehicle in the tunnel road according to the centerline position information, the lane width information and the boundary information; The determining of the location information of the search starting point according to the passable area location information and the centerline location information specifically includes: Projecting the centerline position information and the passable area position information into a two-dimensional grid map respectively, and correspondingly obtaining the centerline position information in the two-dimensional grid map and the passable area position point information in the two-dimensional grid map; According to the relative position relationship between the passable area position points in the two-dimensional grid map and the center line in the two-dimensional grid map, the passable area position points in the two-dimensional grid map are divided into left passable area position points and right passable area position points; Calculating a first distance value between the position point in the left passage area and the origin of the vehicle coordinate system, and calculating a second distance value between the position point in the right passage area and the origin of the vehicle coordinate system; The passable area position point corresponding to the minimum value in the first distance values ​​is used as the search starting point for the left road boundary point, and the passable area position point corresponding to the minimum value in the second distance values ​​is used as the search starting point for the right road boundary point.

2. The method according to claim 1, characterized in that The determining of the position information of the first candidate point of the tunnel road boundary point according to the first environmental point cloud data specifically includes: Convert the first environmental point cloud data into a coordinate system to obtain second environmental point cloud data in a vehicle coordinate system; Projecting the second environment point cloud data into a two-dimensional grid map, obtaining third environment point cloud data in the two-dimensional grid map, and recording a height value of the third environment point cloud; Selecting an environmental point cloud to be analyzed from the third environmental point cloud, respectively calculating a first difference between a height value of the environmental point cloud to be analyzed and a height value of a third environmental point cloud of a set area, determining whether the first difference is greater than a first set threshold, and if so, counting once to obtain a cumulative count corresponding to the set area; It is determined whether the accumulated count is greater than a count threshold; if so, the first environment point cloud corresponding to the environment point cloud to be analyzed is the first candidate point.

3. The method according to claim 1, characterized in that The determining of the tunnel road boundary point information according to the centerline position information, the search starting point position information and the first candidate point position information specifically includes: Processing the first candidate point to obtain a second candidate point, and projecting the second candidate point into a two-dimensional grid map; According to the relative position relationship between the second candidate point in the two-dimensional grid map and the center line in the two-dimensional grid map, the second candidate point is divided into a left road candidate boundary point and a right road candidate boundary point; The left road boundary point information is obtained according to the left road candidate boundary point and the search starting point of the left road boundary point, and the right road boundary point information is obtained according to the right road candidate boundary point and the search starting point of the right road boundary point.

4. The method according to claim 3, characterized in that The processing of the first candidate point to obtain the second candidate point specifically includes: Traversing the first candidate points, obtaining first candidate points marked by the same laser beam of the laser radar, and clustering the first candidate points marked by the same laser beam; Calculate a third distance value between each first candidate point in any class and the horizontal direction of the laser radar; Obtaining a minimum distance value and a maximum distance value among the third distance values ​​in the arbitrary class, and calculating a second difference between the minimum distance value and the maximum distance value; It is determined whether the second difference is greater than a second set threshold value. If so, the first candidate point in the arbitrary class is the second candidate point.

5. The method according to claim 3, characterized in that: The obtaining of the left road boundary point information according to the left road candidate boundary point and the search starting point of the left road boundary point specifically includes: Taking the search starting point of the left road boundary point as the initial search center, searching the two-dimensional grid map line by line according to the set rules; If the left road candidate boundary point is searched, the position coordinates of the searched left road candidate boundary point are obtained, a first slope value of the lane centerline at the same ordinate as the searched left road candidate boundary point is calculated, next search starting position information is determined according to the position coordinates of the searched left road candidate boundary point and the first slope value, and the next search starting position is used as the first current search center, and the search is performed according to the set rule until the search of all rows of the two-dimensional grid map is completed, wherein the searched left road candidate boundary point is the left road boundary point; If the left road candidate boundary point is not found, a second slope value of the lane centerline at the same ordinate as the current search center is calculated, and next search start position information is determined according to the coordinate position information of the current search center and the second slope value, and the next search start position is used as the second current search center, and the two-dimensional grid map is searched row by row according to the set rules until all rows of the two-dimensional grid map are searched, wherein the searched left road candidate boundary point is the left road boundary point; and / or The obtaining of the right road boundary point information according to the right road candidate boundary point and the search starting point of the right road boundary point specifically includes: Taking the search starting point of the right road boundary point as the initial search center, searching the two-dimensional grid map line by line according to the set rules; If the right road candidate boundary point is searched, the position coordinates of the searched right road candidate boundary point are obtained, and the third slope value of the lane centerline at the same ordinate as the searched right road candidate boundary point is calculated, and the next search starting position is determined according to the position coordinates of the searched right road candidate boundary point and the third slope value, and the next search starting position is used as the third current search center, and the search is performed according to the set rules until the search of all rows of the two-dimensional grid map is completed, wherein the searched right road candidate boundary point is the right road boundary point; If the right road candidate boundary point is not searched, the fourth slope value of the lane centerline at the same vertical coordinate as the current search center is calculated, and the next search starting position information is determined according to the coordinate position information of the current search center and the fourth slope value. The next search starting position is used as the fourth current search center, and the two-dimensional grid map is searched row by row according to the set rules until the search of all rows of the two-dimensional grid map is completed, wherein the searched right road candidate boundary point is the right road boundary point.

6. The method according to claim 5, characterized in that The determining of the next search start position information according to the searched position coordinates of the left road candidate boundary point and the first slope value specifically includes: Determine the ordinate value of the searched left road candidate boundary point as the ordinate value of the next search starting position; A first product is obtained by multiplying the horizontal coordinate value of the searched left road candidate boundary point by the first slope value, and a first sum value of the first product and the horizontal coordinate of the searched left road candidate boundary point is calculated, wherein the first sum value is the horizontal coordinate value of the next search starting position.

7. The method according to claim 6, characterized in that The calculating of the first slope value of the lane centerline at the same ordinate as the searched left road candidate boundary point specifically includes: Acquire a first lateral coordinate value of a lane centerline point at the same ordinate as the searched left road candidate boundary point; Obtaining a second lateral coordinate value of a lane centerline point in a row of grids below the lane centerline point; A third difference between the second horizontal coordinate value and the first horizontal coordinate value is calculated, and the third difference is the first slope value.

8. The method according to claim 3, characterized in that Also includes: Determining the tunnel road boundary curve information according to the tunnel road boundary point information specifically includes: The left road boundary points are fitted to obtain left road boundary curve information, and the right road boundary points are fitted to obtain right road boundary curve information.

9. The method according to claim 8, characterized in that Determining the lateral position information of the vehicle in the tunnel road according to the centerline position information, the lane width information and the boundary information specifically includes: The total number of lanes of the tunnel road, the lane where the vehicle is currently located, and the lateral position of the vehicle in the lane where the vehicle is currently located are determined according to the centerline position information, the lane width information, the left road boundary curve information, and the right road boundary curve information.

10. The method according to claim 9, characterized in that The determining the total number of lanes of the tunnel road, the lane where the vehicle is currently located, and the lateral position of the vehicle in the lane where the vehicle is currently located according to the centerline position information, the lane width information, the left road boundary curve information, and the right road boundary curve information specifically includes: Calculating a fourth distance value from the vehicle coordinate origin to the left road boundary curve and a fifth distance value from the vehicle coordinate origin to the right road boundary curve; Calculate a total width value of the tunnel road according to the fourth distance value and the fifth distance value, and calculate a total number of lanes of the tunnel road according to the total width value and the width information; Determine the lane the vehicle is currently in according to the fourth distance value and the width information or determine the lane the vehicle is currently in according to the fifth distance value and the width information; A sixth distance value between the vehicle coordinate origin and the center line is calculated, and a lateral position of the vehicle in a current lane is determined according to the sixth distance value.

11. A vehicle lateral positioning device, characterized in that: include: An acquisition unit, used for acquiring lane line information of a lane currently located by the vehicle, position information of a passable area, and first environmental point cloud data of the tunnel road when the vehicle is traveling in the tunnel road; A calculation unit, used for calculating the center line position information of the lane and the width information of the lane according to the lane line information; A boundary information determination unit, configured to determine the boundary information of the tunnel road according to the lane centerline position information, the passable area position information and the first environmental point cloud data; a lateral position determination unit, configured to determine lateral position information of the vehicle in the tunnel road according to the centerline position information, the width information and the boundary information; The boundary information determining unit comprises: A search starting point determination unit, configured to determine the position information of the search starting point according to the passable area position information and the center line position information, wherein the search starting point is an initial search position of a tunnel road boundary point; A first candidate point determination unit, configured to determine position information of a first candidate point of the tunnel road boundary point according to the first environmental point cloud data; a road boundary point information determining unit, configured to determine the tunnel road boundary point information according to the centerline position information, the position information of the search starting point and the position information of the first candidate point; A road boundary determination unit, configured to determine the tunnel road boundary information according to the tunnel road boundary point information; The search starting point determination unit is specifically used for: Projecting the centerline position information and the passable area position information into a two-dimensional grid map respectively, and correspondingly obtaining the centerline position information in the two-dimensional grid map and the passable area position point information in the two-dimensional grid map; According to the relative position relationship between the passable area position points in the two-dimensional grid map and the center line in the two-dimensional grid map, the passable area position points in the two-dimensional grid map are divided into left passable area position points and right passable area position points; Calculating a first distance value between the position point in the left passage area and the origin of the vehicle coordinate system, and calculating a second distance value between the position point in the right passage area and the origin of the vehicle coordinate system; The passable area position point corresponding to the minimum value in the first distance values ​​is used as the search starting point for the left road boundary point, and the passable area position point corresponding to the minimum value in the second distance values ​​is used as the search starting point for the right road boundary point.

12. The device according to claim 11, characterized in that The road boundary point information determination unit specifically includes: A second candidate point determination unit, configured to process the first candidate point to obtain a second candidate point, and project the second candidate point into a two-dimensional grid map; a candidate road boundary point determination unit, configured to divide the second candidate point in the two-dimensional grid map into a left road candidate boundary point and a right road candidate boundary point according to a relative position relationship between the second candidate point in the two-dimensional grid map and a center line in the two-dimensional grid map; A left and right road boundary point determination unit, configured to obtain left road boundary point information based on the left road candidate boundary point and the search starting point of the left road boundary point, and to obtain right road boundary point information based on the right road candidate boundary point and the search starting point of the right road boundary point; The fitting unit is used to fit the left road boundary point information and the right road boundary point information respectively, and correspondingly obtain the left road boundary curve information and the right road boundary curve information.

13. A vehicle lateral positioning system, characterized in that: include: A visual detection device, a laser radar, and a vehicle lateral positioning device as claimed in any one of claims 11 to 12, wherein: The visual detection device is used to collect and process image information of the vehicle's current driving direction when the vehicle is driving in the tunnel road, and output lane line information of the lane where the vehicle is currently located and passable area position information; The laser radar device is used to collect environmental information of the tunnel road and output first environmental point cloud data of the tunnel road.

14. The system according to claim 13, characterized in that: The visual detection device is an EPM system, and the laser radar is a multi-beam laser radar.

Images