Vehicle automatic driving reference line smoothing method, device, system, equipment and medium

Abstract

The present application provides a method, apparatus, system, equipment and medium for smoothing a reference line for an automatic driving vehicle. The method comprises: determining a point to be smoothed in a target reference line, wherein the point to be smoothed comprises a starting point and an end point; determining a first position constraint, a first-order derivative constraint and a second-order derivative constraint of the starting point, and determining a second position constraint and a heading angle constraint of the end point; smoothing the starting point based on the first position constraint, the first-order derivative constraint and the second-order derivative constraint, and smoothing the end point based on the second position constraint and the heading angle constraint; obtaining a smoothed target reference line based on the smoothed starting point and the smoothed end point. By the above method, the curvature change of the point is smoothed in combination with the first-order derivative constraint and the second-order derivative constraint of the starting point, so that the curvature of the smoothed reference line will not change suddenly, and the smooth driving of the vehicle can be effectively controlled when the reference line is used for path planning.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of automobile driving technology, and in particular to a method, device, system, equipment and medium for smoothing a reference line for automatic vehicle driving. Background Art

[0002] The planning algorithm of autonomous driving is highly dependent on the smoothness of the road reference line. If the smoothness of the reference line cannot meet the requirements of subsequent trajectory planning.

[0003] Current reference line smoothing technology mainly targets point position constraints and heading angle constraints. Although it can ensure that the point positions and heading angles of the reference line remain continuous to a certain extent during the splicing process, it usually cannot ensure that the curvature of the reference line is also continuous. This will result in the reference line not achieving a good smoothing effect, and in turn, the vehicle control system will still produce large fluctuations when using the reference line for path planning. Summary of the Invention

[0004] Embodiments of the present application provide a method, apparatus, system, device, and medium for smoothing a vehicle autonomous driving reference line to solve at least one of the above-mentioned problems.

[0005] According to one aspect of an embodiment of the present application, a method for smoothing a reference line for autonomous driving of a vehicle is provided, comprising:

[0006] Determine the points to be smoothed in the target reference line, wherein the points to be smoothed include a starting point and an end point;

[0007] Determining a first position constraint, a first-order derivative constraint, and a second-order derivative constraint of the starting point, and determining a second position constraint and a heading angle constraint of the end point;

[0008] The starting point is smoothed based on the first position constraint, the first-order derivative constraint, and the second-order derivative constraint, and the end point is smoothed based on the second position constraint and the heading angle constraint; and a smoothed target reference line is obtained based on the smoothed starting point and the smoothed end point.

[0009] In the embodiment of the present application, the above method can be used to achieve smoothing of the position, heading angle and curvature of the reference line. The curvature of the smoothed reference line will not change suddenly, and the smooth driving of the vehicle can be effectively controlled when the reference line is used for path planning.

[0010] In one embodiment, determining the points to be smoothed in the reference line includes:

[0011] Obtaining a first measurement data frame at the current moment and a second measurement data frame at the previous moment;

[0012] The overlapping reference line between the first reference line corresponding to the first measurement data frame and the second reference line corresponding to the second measurement data frame is determined as a target reference line, and the starting point and the end point of the target reference line are determined as points to be smoothed on the target reference line.

[0013] In the embodiment of the present application, the above method can be used to quickly determine the points to be smoothed in the reference line, thereby improving the smoothing efficiency of the target reference line.

[0014] In one embodiment, the determining of the first position constraint, the first-order derivative constraint, and the second-order derivative constraint of the starting point includes:

[0015] determining a first transverse coordinate constraint of the starting point according to the transverse coordinate position information of the starting point, a plurality of first optimization coefficients related to the transverse coordinate of the starting point, and initial distance information;

[0016] determining a first longitudinal coordinate constraint of the starting point according to the longitudinal coordinate position information of the starting point, a plurality of second optimization coefficients related to the longitudinal coordinate of the starting point, and the initial distance information;

[0017] Determining a first position constraint of the starting point according to the first horizontal coordinate constraint and the first vertical coordinate constraint, and determining a first-order derivative constraint and a second-order derivative constraint of the starting point according to the first position constraint;

[0018] and / or,

[0019] The determining of the second position constraint and the heading angle constraint of the endpoint includes:

[0020] determining a second abscissa constraint of the endpoint according to the abscissa position information of the endpoint, a plurality of third optimization coefficients related to the abscissa of the endpoint, and a distance between the endpoint and initial distance information;

[0021] determining a second longitudinal coordinate constraint of the endpoint according to the longitudinal coordinate position information of the endpoint, a plurality of fourth optimization coefficients related to the longitudinal coordinate of the endpoint, and a distance between the endpoint and initial distance information;

[0022] The second position constraint of the end point is determined according to the second horizontal coordinate constraint and the second vertical coordinate constraint, and the second position constraint is differentiated, and the heading angle constraint of the end point is determined according to the derivative result of the second position constraint and the sine value corresponding to the end point.

[0023] In the embodiment of the present application, the above method can be used to efficiently determine position constraints, heading angle constraints and multi-order derivative constraints to further improve the smoothing efficiency of the reference line.

[0024] In one embodiment, determining the first horizontal coordinate constraint of the starting point includes: constructing a first Nth-order polynomial regarding the initial distance information based on the horizontal coordinate position information of the starting point, a plurality of first optimization coefficients regarding the horizontal coordinate of the starting point, and initial distance information, and determining the first horizontal coordinate constraint of the starting point based on the first Nth-order polynomial;

[0025] Determining the first ordinate constraint of the starting point includes: constructing a second Nth-order polynomial regarding the initial distance information based on the ordinate position information of the starting point, a plurality of second optimization coefficients regarding the ordinate of the starting point, and the initial distance information, and determining the first ordinate constraint of the starting point based on the second Nth-order polynomial;

[0026] and / or,

[0027] Determining the second horizontal coordinate constraint of the endpoint includes: constructing a third Nth-order polynomial regarding the distance based on the horizontal coordinate position information of the endpoint, a plurality of third optimization coefficients regarding the horizontal coordinate of the endpoint, and the distance between the endpoint and the initial distance information, and determining the second horizontal coordinate constraint of the endpoint based on the third Nth-order polynomial;

[0028] The determining of the second longitudinal coordinate constraint of the end point includes: constructing a fourth Nth-order polynomial about the distance based on the longitudinal coordinate position information of the end point, a number of fourth optimization coefficients about the longitudinal coordinate of the end point, and the distance between the end point and the initial distance information, and determining the second longitudinal coordinate constraint of the end point based on the fourth Nth-order polynomial.

[0029] In this embodiment, the above method can be used to perform fitting using polynomials to determine the corresponding various constraints, so that the smoothing result is more stable, thereby further improving the smoothing effect of the reference line.

[0030] In one embodiment, the method further comprises:

[0031] Normalizing the initial distance information and the distance between the endpoint and the initial distance information according to a preset normalization coefficient, wherein the preset normalization coefficient is determined based on the initial distance information or the distance and the total length of the target reference line;

[0032] The first horizontal coordinate constraint and the first vertical coordinate constraint of the starting point are respectively determined according to the normalized initial distance information, and the second horizontal coordinate constraint and the second coordinate constraint of the end point are respectively determined according to the normalized distance.

[0033] In this embodiment, the above method can effectively reduce the amount of calculation to improve the smoothing efficiency of the reference line.

[0034] In one embodiment, the method further comprises:

[0035] A planned path is generated based on the smoothed target reference line, and the planned path is sent to a target control device so that the target control device controls the autonomous driving vehicle to travel based on the planned path.

[0036] In this embodiment, the above method can be used to effectively control the driving stability of the vehicle when the control device uses the reference line for path planning.

[0037] According to a second aspect of an embodiment of the present application, a vehicle automatic driving reference line smoothing device is provided, comprising:

[0038] A module for determining points to be smoothed, configured to determine points to be smoothed in a target reference line, wherein the points to be smoothed include a starting point and an end point;

[0039] a constraint determination module configured to determine a first position constraint, a first-order derivative constraint, and a second-order derivative constraint of the starting point, and to determine a second position constraint and a heading angle constraint of the end point;

[0040] A smoothing module is configured to smooth the starting point based on the first position constraint, the first-order derivative constraint, and the second-order derivative constraint, and to smooth the end point based on the second position constraint and the heading angle constraint; and obtain a smoothed target reference line based on the smoothed starting point and the smoothed end point.

[0041] In one embodiment, the vehicle automatic driving reference line smoothing device can be used to perform any possible implementation of the first aspect described above.

[0042] According to a third aspect of an embodiment of the present application, an autonomous driving system is provided, including a planning device and a control device:

[0043] The planning device is used to obtain a smoothed target reference line according to the vehicle automatic driving reference line smoothing method, and generate a planned path according to the smoothed target reference line;

[0044] The control device is used to receive the planned path and control the autonomous driving vehicle to travel based on the planned path.

[0045] In one embodiment, the planning device may be used to execute any possible implementation of the first aspect.

[0046] According to a fourth aspect of an embodiment of the present application, an electronic device is provided, comprising: a memory, a processor and a computer program, wherein the computer program is stored in the memory, and the processor runs the computer program to execute the vehicle automatic driving reference line smoothing method.

[0047] According to a fifth aspect of an embodiment of the present application, a computer-readable storage medium is provided, wherein the storage medium includes a computer program, and the computer program is used to implement the vehicle automatic driving reference line smoothing method.

[0048] According to a sixth aspect of an embodiment of the present application, a computer program product is provided, comprising a computer program, which, when executed by a processor, describes a method for smoothing a reference line for automatic driving of a vehicle.

[0049] According to the seventh aspect of an embodiment of the present application, a chip is provided, comprising a memory and a processor, wherein the memory is used to store a computer program, and the processor is used to call and run the computer program from the memory to execute the vehicle automatic driving reference line smoothing method.

[0050] The embodiments of the present application provide a method, apparatus, system, device, and medium for smoothing a reference line for an autonomous vehicle. The method, apparatus, system, and medium determine the points to be smoothed in a target reference line, including a starting point and an end point, and determine the first position constraint, first-order derivative constraint, and second-order derivative constraint of the starting point, and determine the second position constraint and heading angle constraint of the end point. The method then smoothes the starting point based on the first position constraint, first-order derivative constraint, and second-order derivative constraint, and smoothes the end point based on the second position constraint and heading angle constraint. Finally, the method obtains a smoothed target reference line based on the smoothed starting point and the smoothed end point. In this process, the position constraint, first-order derivative constraint, and second-order derivative constraint of the starting point, and the position constraint and heading angle constraint of the end point are selected to smooth the reference line. Compared with related technologies, in addition to smoothing the point position and heading angle, the curvature change of the point is also smoothed, so that the curvature of the smoothed reference line does not change suddenly. When the reference line is used for path planning, the vehicle can be effectively controlled to travel smoothly. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1a A possible framework diagram of a method for smoothing a reference line for autonomous driving of a vehicle provided in an embodiment of the present application;

[0052] Figure 1b Schematic diagram of curvature change of reference line after smoothing in related art;

[0053] Figure 2 A schematic diagram of a flow chart of a method for smoothing a reference line for autonomous driving of a vehicle provided in an embodiment of the present application;

[0054] Figure 3 A schematic flow chart of another method for smoothing a reference line for autonomous driving of a vehicle provided in an embodiment of the present application;

[0055] Figure 4 This is a schematic diagram of the curvature change of the main road curve after reference line smoothing in an embodiment of the present application;

[0056] Figure 5 A schematic diagram of the curvature change after reference line smoothing of a main road curve in the related art;

[0057] Figure 6 A schematic diagram of the structure of a reference line smoothing device for automatic driving of a vehicle provided in an embodiment of the present application;

[0058] Figure 7 A schematic diagram of the structure of an autonomous driving system provided in an embodiment of the present application;

[0059] Figure 8 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0060] In order to make the purpose, technical solutions and advantages of the present application clearer, the technical solutions in the embodiments of the present application will be described in more detail below in conjunction with the drawings in the embodiments of the present application. In the drawings, the same or similar reference numerals throughout represent the same or similar parts or parts with the same or similar functions. The described embodiments are part of the embodiments of the present application, not all of the embodiments. The embodiments described below with reference to the drawings are exemplary and are intended to be used to explain the present application, and should not be understood as limitations on the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0061] Figure 1a A possible framework diagram provided for an embodiment of the present application is as follows: Figure 1a As shown, the autonomous driving system includes modules for data perception, positioning, map processing, etc., which are used to provide the decision-making module with dynamic obstacle status, map and positioning information. The decision-making module is used to provide the planning module with decision-making behaviors such as following the vehicle and changing lanes. The planning module generates the planned trajectory, and the control module is used to control the vehicle to track the planned trajectory.

[0062] In the planning module, a reference line smoothing algorithm is needed to smooth the reference line issued by the decision module so that the control module can track the reference line more smoothly and achieve stable driving of the vehicle.

[0063] It can be understood that the main purpose of the smooth reference line is to make the position, heading angle, curvature and other attributes of each point on the reference line continuous, so as to control the vehicle to track the planned trajectory with a relatively gentle acceleration (lateral acceleration or longitudinal acceleration).

[0064] In related technologies, the reference line smoothing algorithm for the planning module usually uses a parameterized polynomial as the reference line fitting path. The description of the points on the reference line is as follows:

[0065] x=a0+a1s+a2s 2 +a3s 3 +a4s 4 +a5s 5 (1)

[0066] y=b0+b1s+b2s 2 +b3s 3 +b4s 4 +b5s 5 (2)

[0067] Where x and y represent the horizontal and vertical positions of a point, respectively. a0-a5 correspond to the optimization parameters of the horizontal and vertical coordinates, respectively. b0-b5 correspond to the optimization parameters of the vertical coordinates, respectively. s represents the distance between a point on the reference line and the starting point. During the reference line smoothing process, the smoothing result of the previous frame needs to be spliced ​​with the smoothing result of the current frame. This involves the smoothing problem of the splicing point.

[0068] In many examples, the splicing points use position constraints and heading angle constraints. In one implementation, if the position constraints and heading angle constraints of the start and end points are selected, the description is as follows:

[0069]

[0070]

[0071]

[0072]

[0073]

[0074]

[0075] For example, if the reference line has one segment, corresponding to the above formula, there are 12 optimization variables (i.e., a0-a5, b0-b5), and the equality constraints are at least 6 (i.e., formulas (3)-(8) above). Similarly, if the reference line has two segments, the optimization variables are 24.

[0076] Where x0 is the horizontal coordinate position constraint of the starting point, y0 is the vertical coordinate position constraint of the starting point, s0 is the initial distance information, and x t is the horizontal coordinate position constraint of the end point, y t is the ordinate position constraint of the end point, is the first-order derivative of the horizontal coordinate position constraint of the starting point, is the first-order derivative of the ordinate position constraint of the starting point, θ0 is the heading angle of the starting point, sinθ0 represents the sine value of the heading angle of the starting point, is the first-order derivative of the end point horizontal coordinate position constraint, is the first-order derivative of the ordinate position constraint of the end point, θ t is the heading angle of the end point, sinθ t is the sine of the end point heading angle.

[0077] The above smoothing algorithm can easily obtain the position continuity and heading angle continuity of the splicing points to a certain extent. However, since it does not involve the smoothing of the curvature in the reference line, it is usually difficult to ensure the curvature continuity between the splicing points. Specifically, the above smoothing algorithm smoothes the reference line, and its smoothing result is as follows: Figure 1b As shown, Figure 1b In the figure, the horizontal axis is the number of points, the difference between two points is 0.02s, and the vertical axis represents the curvature of the reference line at that point. As can be seen from the figure, the curvature basically has a sudden change at each splicing point.

[0078] The applicant studied the curvature change during the smoothing process of the reference line, combined with the theoretical calculation formula of curvature Where, are the first and second derivatives of the function y, respectively, and k represents the curvature. To prevent sudden changes in the curvature between the splicing points, the curvature between the splicing points must be constrained to be the same. According to the curvature change formula, the first and second derivatives between the splicing points must be the same.

[0079] In view of this, an embodiment of the present application provides a method for smoothing a reference line for an automatic driving vehicle, by first determining the points to be smoothed in the target reference line, the points to be smoothed including the starting point and the end point, and determining the first position constraint, first-order derivative constraint, and second-order derivative constraint of the starting point, as well as the second position constraint and heading angle constraint of the end point, then smoothing the starting point based on the first position constraint, the first-order derivative constraint, and the second-order derivative constraint, and smoothing the end point based on the second position constraint and the heading angle constraint, and finally obtaining a smoothed target reference line based on the smoothed starting point and the smoothed end point. In this process, the position constraint, first-order derivative constraint, and second-order derivative constraint of the starting point, and the position constraint and heading angle constraint of the end point are selected to smooth the reference line. Compared with the related art, in addition to smoothing the point position and heading angle, the curvature change of the point is also smoothed, so that the curvature of the smoothed reference line will not change suddenly. When using the reference line for path planning, the smooth driving of the vehicle can be effectively controlled.

[0080] The above briefly describes the scenario diagram of the present application. Below, taking the execution subject server as an example, the vehicle automatic driving reference line smoothing method provided by the embodiment of the present application is described in detail.

[0081] Please refer to Figure 2 , Figure 2 A flow chart of a method for smoothing a reference line for autonomous driving of a vehicle is provided for an embodiment of the present application, including steps S201-S204.

[0082] Step S201: Determine the points to be smoothed in the target reference line, where the points to be smoothed include a starting point and an end point.

[0083] It is understood that the target reference line in this embodiment is the road reference line that the vehicle can use for navigation. The reference line is the path that the autonomous vehicle should follow if there are no surrounding obstacles. If the reference line is directly derived from the original lane centerline of the HD map, the smoothness of the reference line will not meet the requirements of subsequent trajectory planning due to map production errors and uneven road connections. Smoothing the reference line can make vehicle control more stable.

[0084] In this embodiment, by first determining the points to be smoothed on the reference line and then smoothing the points to be smoothed, compared to smoothing all discrete points on the reference line, targeted smoothing of the points to be smoothed can be achieved, thereby improving the smoothing efficiency of the reference line to a certain extent.

[0085] In one implementation, the reference line undergoes a complete smoothing process in each frame of measurement data (e.g., the measurement data frame of the road centerline collected by the perception module). The smoothing result for each frame is the reference line from the vehicle's position in that frame to a certain distance ahead, such as 400 meters (this parameter can be set). Therefore, in two consecutive frames of measurement data, there will be overlapping reference lines. This overlapping portion requires ensuring that the calculation results of the two frames are consistent. Therefore, in this embodiment, the junction point between the reference lines corresponding to the two frames of measurement data is determined as the point to be smoothed. Specifically, determining the point to be smoothed in the reference line in step S201 of this embodiment may include the following steps:

[0086] Obtaining a first measurement data frame at the current moment and a second measurement data frame at the previous moment;

[0087] The overlapping reference line between the first reference line corresponding to the first measurement data frame and the second reference line corresponding to the second measurement data frame is determined as a target reference line, and the starting point and the end point of the target reference line are determined as points to be smoothed on the target reference line.

[0088] During the vehicle's autonomous driving process, sensors continuously collect measurement data, namely road centerline data, and send it to downstream decision-making modules and planning modules to facilitate path planning. For two consecutive frames of measurement data, there will be overlapping parts between the reference lines. For this overlapping part, it is necessary to ensure that the calculation results of the two frames are consistent, that is, the reference line part needs to be smoothed.

[0089] In this embodiment, the starting point and end point of the target reference line are the splicing points of the reference lines corresponding to two consecutive frames of measurement data. The starting point and end point of the target reference line are determined as the points to be smoothed, which can improve the smoothing efficiency. Specifically, the target reference line is the overlapping part of the reference line of the previous frame (i.e., the second reference line) and the reference line of the current frame (i.e., the first reference line), wherein the starting point is the splicing end point of the reference line of the previous frame, the splicing starting point of the reference line of the current frame, and the end point is the splicing end point of the reference line of the current frame. In this embodiment, the starting point and the end point are smoothed in each frame, that is, the position, heading angle and curvature of the splicing starting point of the current reference line are ensured to be the same as the splicing end point (i.e., the starting point) of the reference line of the previous frame. In addition, the position and heading angle of the end point are ensured to be the same as the position and heading angle of the original reference line (the reference line of the current frame) at that point.

[0090] By using the above method of determining the target reference line, the points to be smoothed in the reference line can be quickly determined, and then the points to be smoothed can be smoothed, thereby effectively improving the smoothing efficiency of the target reference line.

[0091] Step S202: Determine a first position constraint, a first-order derivative constraint, and a second-order derivative constraint of the starting point, and determine a second position constraint and a heading angle constraint of the end point.

[0092] In this embodiment, in addition to position constraints and heading angle constraints on the splicing points, first-order derivative constraints and second-order derivative constraints are applied to the starting point to achieve smooth curvature of the target reference line and maintain curvature continuity between reference lines.

[0093] In one embodiment, the UTM coordinates and the distance information corresponding to the splicing point are combined to further determine the splicing point position constraint, the first-order / second-order derivative constraint, and the heading angle constraint. Specifically, step S202 determines the first position constraint, the first-order derivative constraint, and the second-order derivative constraint of the starting point, which may include the following steps:

[0094] determining a first transverse coordinate constraint of the starting point according to the transverse coordinate position information of the starting point, a plurality of first optimization coefficients related to the transverse coordinate of the starting point, and initial distance information;

[0095] determining a first longitudinal coordinate constraint of the starting point according to the longitudinal coordinate position information of the starting point, a plurality of second optimization coefficients related to the longitudinal coordinate of the starting point, and the initial distance information;

[0096] A first position constraint of the starting point is determined according to the first horizontal coordinate constraint and the first vertical coordinate constraint, and a first-order derivative constraint and a second-order derivative constraint of the starting point are determined according to the first position constraint.

[0097] It is understood that the distance corresponding to the starting point is the initial distance information, which can be obtained based on the measurement data frame. The first optimization coefficients may be one or more optimization coefficients. If there are multiple optimization coefficients, each optimization coefficient may have different values. The values ​​of the first optimization coefficients may be determined based on the optimization objective of reference line smoothing. For example, in straight roads and curves, since the optimization objectives of the two are generally different, the values ​​of the first optimization coefficients are generally different.

[0098] It should be noted that those skilled in the art can adapt the first optimization coefficient to actual applications, and the same applies to the second, third, and fourth optimization coefficients described below. Furthermore, the first, second, third, and fourth optimization coefficients in this embodiment are merely used to distinguish similar objects and have no other meaning. They can represent the same or different contents and are not specifically limited here.

[0099] To further improve the smoothing effect, in this embodiment, determining the first horizontal coordinate constraint of the starting point in the above step may include the following steps: constructing a first Nth-order polynomial regarding the initial distance information based on the horizontal coordinate position information of the starting point, a plurality of first optimization coefficients regarding the horizontal coordinate of the starting point, and the initial distance information, and determining the first horizontal coordinate constraint of the starting point based on the first Nth-order polynomial;

[0100] Determining the first vertical coordinate constraint of the starting point in the above steps may include the following steps: constructing a second Nth-order polynomial about the initial distance information based on the vertical coordinate position information of the starting point, a number of second optimization coefficients about the vertical coordinate of the starting point and the initial distance information, and determining the first vertical coordinate constraint of the starting point based on the second Nth-order polynomial.

[0101] Specifically, the first Nth-order polynomial and the second Nth-order polynomial are each a quintic polynomial. Using a quintic polynomial curve fitting can make the longitudinal acceleration control change smoothly. The first horizontal coordinate constraint and the first vertical coordinate constraint respectively satisfy the following formulas. In this embodiment, the following formulas correspond to the position constraint of the starting point:

[0102]

[0103]

[0104] In the formula, x0 represents the horizontal coordinate position information of the starting point, a0-a5 represent several first optimization coefficients respectively, s0 represents the initial distance information, y0 represents the vertical coordinate information of the starting point, and b0-b5 represent several second optimization coefficients respectively.

[0105] Furthermore, the position constraints of the first Nth order polynomial and the second Nth order polynomial are combined to determine the corresponding first-order derivative constraints and second-order derivative constraints. The first-order derivative constraints and the second-order derivative constraints satisfy the following formulas respectively:

[0106]

[0107]

[0108]

[0109]

[0110] Understandably, The corresponding formula represents the first-order derivative of the starting point (including the first-order derivative constraint of the horizontal coordinate and the first-order derivative constraint of the vertical coordinate), The corresponding formula represents the second-order derivative of the starting point (including the second-order derivative constraint of the horizontal coordinate and the second-order derivative constraint of the vertical coordinate).

[0111] Furthermore, determining the second position constraint and heading angle constraint of the endpoint in step S202 may include the following steps:

[0112] determining a second abscissa constraint of the endpoint according to the abscissa position information of the endpoint, a plurality of third optimization coefficients related to the abscissa of the endpoint, and a distance between the endpoint and initial distance information;

[0113] determining a second longitudinal coordinate constraint of the endpoint according to the longitudinal coordinate position information of the endpoint, a plurality of fourth optimization coefficients related to the longitudinal coordinate of the endpoint, and a distance between the endpoint and initial distance information;

[0114] The second position constraint of the end point is determined according to the second horizontal coordinate constraint and the second vertical coordinate constraint, and the second position constraint is differentiated, and the heading angle constraint of the end point is determined according to the derivative result of the second position constraint and the sine value corresponding to the end point.

[0115] It can be understood that the distance between the end point and the initial distance information, that is, the distance between the end point and the distance information corresponding to the starting point, is the cumulative distance traveled by the vehicle from the initial point of the reference line to that point.

[0116] Determining the second horizontal coordinate constraint of the endpoint in the above step may include the following steps: constructing a third Nth-order polynomial regarding the distance based on the horizontal coordinate position information of the endpoint, a plurality of third optimization coefficients regarding the horizontal coordinate of the endpoint, and the distance between the endpoint and the initial distance information, and determining the second horizontal coordinate constraint of the endpoint based on the third Nth-order polynomial;

[0117] Determining the second longitudinal coordinate constraint of the end point in the above steps may include the following steps: constructing a fourth Nth-order polynomial about the distance based on the longitudinal coordinate position information of the end point, several fourth optimization coefficients about the longitudinal coordinate of the end point, and the distance between the end point and the initial distance information, and determining the second longitudinal coordinate constraint of the end point based on the fourth Nth-order polynomial.

[0118] Specifically, the third N-order polynomial and the fourth N-order polynomial are quintic polynomials that satisfy the following formula:

[0119]

[0120]

[0121] Furthermore, the heading angle constraint of the end point satisfies the following formula:

[0122]

[0123] Where x t Indicates the horizontal coordinate position information of the end point, y t Indicates the vertical coordinate position information of the end point, s t represents the distance between the end point and the initial distance information, a0-a5 represent a number of third optimization coefficients, b0-b5 represent a number of fourth optimization coefficients, θ t Indicates the heading angle of the end point, sinθ t It should be noted that, in this embodiment, the target reference line is taken as an example of a segment, the third optimization coefficient is the same as the first optimization coefficient, and the second optimization coefficient is the same as the fourth optimization coefficient. In some embodiments, if the target reference line has multiple segments, the optimization coefficients of different segments may also be different.

[0124] In the above process of this embodiment, polynomial fitting is performed to determine the corresponding various constraints, so that the smoothing result is more stable, thereby further improving the smoothing effect of the reference line.

[0125] Step S203: smoothing the starting point based on the first position constraint, the first-order derivative constraint, and the second-order derivative constraint, and smoothing the end point based on the second position constraint and the heading angle constraint.

[0126] In this embodiment, for the smoothing of the starting point and the end point, taking the above formula as an example, for the position constraint of the starting point, that is, when the initial distance s=s0 of the starting point, x=x0, y=y0, the first-order derivative constraint and the second-order derivative constraint of the starting point, that is, when the initial distance s=s0 of the starting point, The position constraint of the end point is that when the distance between the end point and the starting point is s=st, x=xt, y=yt; the heading angle constraint of the end point is that when the end point x=xt, y=yt, θ=θ t .

[0127] Step S204: Acquire a smoothed target reference line based on the smoothed starting point and the smoothed end point.

[0128] In the embodiment of the present application, the above method can be used to achieve smoothing of the position, heading angle and curvature of the reference line. The curvature of the smoothed reference line will not change suddenly, and the smooth driving of the vehicle can be effectively controlled when the reference line is used for path planning.

[0129] In one embodiment, considering that the larger the value of the distance information is, the larger the error in the reference line smoothing algorithm will be. In particular, when a multi-order polynomial is used for calculation, not only is the computational complexity large, but also large errors are easily generated, this embodiment first normalizes the corresponding distance information. Specifically, the method further includes:

[0130] Normalizing the initial distance information and the distance between the endpoint and the initial distance information according to a preset normalization coefficient, wherein the preset normalization coefficient is determined based on the initial distance information or the distance and the total length of the target reference line;

[0131] The first horizontal coordinate constraint and the first vertical coordinate constraint of the starting point are respectively determined according to the normalized initial distance information, and the second horizontal coordinate constraint and the second coordinate constraint of the end point are respectively determined according to the normalized distance.

[0132] In this embodiment, the reference line is smoothed using a piecewise quintic polynomial. After normalization, when performing function programming, the distance information of each segment of the polynomial can be treated as 1, which greatly reduces the difficulty of function programming. Otherwise, it is necessary to add distance information as a function parameter, and if the distance information is too large, the calculation error will also become larger.

[0133] In this embodiment, s in two consecutive frames are used. t Take the normalization process as an example. For the first and second order derivatives, since the distance s field of the points on the reference line path is normalized during the smoothing process, it is also necessary to analyze the calculation method of the first and second order derivatives of the previous and next frames during the splicing process to ensure that they are the same in the UTM coordinate system and before normalization.

[0134] In the previous frame, the calculated point x t and y t The s at s t , after normalization, it is s1, and the coordinates are as follows:

[0135]

[0136]

[0137] In the next frame, the calculated point x t and y t The s at s t , after normalization, it is s2, and the coordinates are as follows:

[0138]

[0139]

[0140] make Among them, scale represents the normalization coefficient, length represents the total length of the target reference line, and num_spline represents s t The corresponding distances are:

[0141]

[0142]

[0143] thereby:

[0144]

[0145]

[0146] Where, Represents x t to s t The first derivative of Represents x t to s t The second derivative of represents y t to s t The first derivative of represents y t to s t The second derivative of .

[0147] In this embodiment, when performing first-order and second-order derivative constraints, it should be written as:

[0148]

[0149]

[0150]

[0151]

[0152] Where scale0 is the normalization coefficient of the starting point, scale1 and scale2 are s t The normalization coefficient of the corresponding distance.

[0153] In this embodiment, the above-mentioned normalization method can effectively reduce the amount of calculation to improve the smoothing efficiency of the reference line.

[0154] Please refer to Figure 3 , Figure 3 A flow chart of another method for smoothing a reference line for automatic driving of a vehicle provided in an embodiment of the present application. Based on the above embodiment, after generating a smoothed target reference line, this embodiment sends the smoothed target reference line to a downstream control device to achieve smooth driving of the vehicle. Specifically, in addition to the above steps S201-203, after step S204, the method also includes step S301.

[0155] Step S301: Generate a planned path based on the smoothed target reference line, and send the planned path to a target control device so that the target control device controls the autonomous driving vehicle to travel based on the planned path.

[0156] In this embodiment, the above method can be used to effectively control the driving stability of the vehicle when the control device uses the reference line for path planning.

[0157] In order to verify the advantages of this embodiment compared with the smoothing method in the related art, the curvature curve of the planned trajectory on the highway is compared through real vehicle debugging. Figure 4 and Figure 5 As shown, Figure 4 In order to use the curvature curve of the constraint 1st and 2nd derivatives of this embodiment on the main road curve, Figure 5 In the related art, only the curvature curve of the constrained heading angle is used on the main road curve. Figure 4 The curvature in is more stable, with almost no sudden changes in curvature, while in Figure 5 There are more mutation points in the .

[0158] Please refer to Figure 6 , Figure 6 A vehicle automatic driving reference line smoothing device provided in an embodiment of the present application is as follows: Figure 6 As shown, it includes a to-be-smoothed point determination module 61, a constraint determination module 62 and a smoothing module 63, wherein:

[0159] A to-be-smoothed point determination module 61 is configured to determine the to-be-smoothed points in the target reference line, wherein the to-be-smoothed points include a starting point and an end point;

[0160] a constraint determination module 62 configured to determine a first position constraint, a first-order derivative constraint, and a second-order derivative constraint of the starting point, and to determine a second position constraint and a heading angle constraint of the end point;

[0161] a smoothing module 63 configured to smooth the starting point based on the first position constraint, the first-order derivative constraint, and the second-order derivative constraint, and to smooth the end point based on the second position constraint and the heading angle constraint; and to obtain a smoothed target reference line based on the smoothed starting point and the smoothed end point.

[0162] In one embodiment, the vehicle automatic driving reference line smoothing device can be used to perform any possible implementation of the first aspect described above.

[0163] In one embodiment, the to-be-smoothed point determination module 61 includes:

[0164] an acquiring unit, configured to acquire a first measurement data frame at a current moment and a second measurement data frame at a previous moment;

[0165] The determining unit is configured to determine the overlapping reference line between the first reference line corresponding to the first measurement data frame and the second reference line corresponding to the second measurement data frame as a target reference line, and determine the starting point and end point of the target reference line as the points to be smoothed of the target reference line.

[0166] In one embodiment, the constraint determination module 62 includes:

[0167] a first horizontal coordinate constraint unit configured to determine a first horizontal coordinate constraint of the starting point according to the horizontal coordinate position information of the starting point, a plurality of first optimization coefficients regarding the horizontal coordinate of the starting point, and initial distance information;

[0168] a first longitudinal coordinate constraint unit configured to determine a first longitudinal coordinate constraint of the starting point according to the longitudinal coordinate position information of the starting point, a plurality of second optimization coefficients related to the longitudinal coordinate of the starting point, and the initial distance information;

[0169] a first constraint determining unit configured to determine a first position constraint of the starting point according to the first abscissa constraint and the first ordinate constraint, and to determine a first-order derivative constraint and a second-order derivative constraint of the starting point according to the first position constraint;

[0170] and / or,

[0171] The constraint determination module 62 includes:

[0172] a second horizontal coordinate constraint unit configured to determine a second horizontal coordinate constraint of the endpoint according to the horizontal coordinate position information of the endpoint, a plurality of third optimization coefficients regarding the horizontal coordinate of the endpoint, and a distance between the endpoint and the initial distance information;

[0173] a second longitudinal coordinate constraint unit configured to determine a second longitudinal coordinate constraint of the end point according to the longitudinal coordinate position information of the end point, a plurality of fourth optimization coefficients regarding the longitudinal coordinate of the end point, and a distance between the end point and the initial distance information;

[0174] a second constraint determination unit, configured to determine a second position constraint of the end point based on the second horizontal coordinate constraint and the second vertical coordinate constraint, to derive the second position constraint, and to determine the heading angle constraint of the end point based on the derivative result of the second position constraint and the sine value corresponding to the end point.

[0175] In one embodiment, the first transverse coordinate constraint unit is specifically configured to construct a first Nth-order polynomial regarding the initial distance information based on the transverse coordinate position information of the starting point, a plurality of first optimization coefficients regarding the transverse coordinate of the starting point, and the initial distance information, and determine the first transverse coordinate constraint of the starting point based on the first Nth-order polynomial;

[0176] The first ordinate constraint unit is specifically configured to construct a second Nth-order polynomial regarding the initial distance information based on the ordinate position information of the starting point, a plurality of second optimization coefficients regarding the ordinate of the starting point, and the initial distance information, and determine the first ordinate constraint of the starting point based on the second Nth-order polynomial;

[0177] and / or,

[0178] The second horizontal coordinate constraint unit is specifically configured to construct a third Nth-order polynomial about the distance based on the horizontal coordinate position information of the end point, a plurality of third optimization coefficients about the horizontal coordinate of the end point, and the distance between the end point and the initial distance information, and determine the second horizontal coordinate constraint of the end point based on the third Nth-order polynomial;

[0179] The second vertical coordinate constraint unit is specifically configured to construct a fourth Nth-order polynomial about the distance based on the vertical coordinate position information of the end point, several fourth optimization coefficients about the vertical coordinate of the end point, and the distance between the end point and the initial distance information, and determine the second vertical coordinate constraint of the end point based on the fourth Nth-order polynomial.

[0180] In one embodiment, the apparatus further comprises:

[0181] A normalization module is configured to normalize the initial distance information and the distance between the end point and the initial distance information according to a preset normalization coefficient, wherein the preset normalization coefficient is determined based on the initial distance information or the distance and the total length of the target reference line; so as to determine the first horizontal coordinate constraint and the first vertical coordinate constraint of the starting point according to the normalized initial distance information, and determine the second horizontal coordinate constraint and the second coordinate constraint of the end point according to the normalized distance.

[0182] In one embodiment, the apparatus further comprises:

[0183] A generation module is configured to generate a planned path based on the smoothed target reference line, and send the planned path to a target control device so that the target control device controls the autonomous driving vehicle to travel based on the planned path.

[0184] Please refer to Figure 7 , Figure 7 An automatic driving system provided in an embodiment of the present application includes a planning device 71 and a control device 72:

[0185] The planning device 71 is used to obtain a smoothed target reference line according to the vehicle automatic driving reference line smoothing method, and generate a planned path according to the smoothed target reference line;

[0186] The control device 72 is used to receive the planned path and control the autonomous driving vehicle to travel based on the planned path.

[0187] In one embodiment, the above-mentioned planning device can be used to execute any possible implementation method in the above-mentioned method embodiment. The relevant description can be understood by the relevant description and effects corresponding to the steps in the corresponding method embodiment, and no further details are given here.

[0188] The present application also provides an electronic device. Figure 8 As shown, it includes: a memory 81, a processor 82 and a computer program, the computer program is stored in the memory 81, and the processor 82 runs the computer program to execute the vehicle automatic driving reference line smoothing method.

[0189] The relevant explanations can be understood by referring to the relevant descriptions and effects corresponding to the steps in the method embodiments, and will not be elaborated here.

[0190] An embodiment of the present application also provides a computer-readable storage medium, which includes a computer program, and the computer program is used to implement the vehicle automatic driving reference line smoothing method.

[0191] The relevant explanations can be understood by referring to the relevant descriptions and effects corresponding to the steps in the method embodiments, and will not be elaborated here.

[0192] An embodiment of the present application also provides a computer program product, including a computer program, which, when executed by a processor, provides a method for smoothing a vehicle automatic driving reference line.

[0193] The relevant explanations can be understood by referring to the relevant descriptions and effects corresponding to the steps in the method embodiments, and will not be elaborated here.

[0194] An embodiment of the present application also provides a chip, including a memory and a processor, wherein the memory is used to store a computer program, and the processor is used to call and run the computer program from the memory to execute the vehicle automatic driving reference line smoothing method.

[0195] The relevant explanations can be understood by referring to the relevant descriptions and effects corresponding to the steps in the method embodiments, and will not be elaborated here.

[0196] It will be understood by those skilled in the art that all or some of the steps, systems, and functional modules / units in the methods disclosed above may be implemented as software, firmware, hardware, and appropriate combinations thereof. In a hardware implementation, the division between the functional modules / units mentioned in the above description does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed by several physical components in cooperation. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or implemented as hardware, or implemented as an integrated circuit, such as an application-specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or non-transitory medium) and a communication medium (or temporary medium).

[0197] As is well known to those skilled in the art, the term computer storage media includes volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information (such as computer-readable instructions, data structures, program modules or other data). Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by a computer.

[0198] Furthermore, as is well known to those skilled in the art, communication media typically embodies computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0199] In the description of the embodiments of the present application, the term "and / or" merely represents a type of association relationship describing associated objects, indicating that three relationships may exist. For example, A and / or B may represent three situations: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the term "at least one" represents any combination of at least two of any one or more of a plurality of items. For example, at least one of A, B, and C may represent any one or more elements selected from a set including A, B, and C.

[0200] In the description of the embodiments of the present application, the terms "first", "second", "third", "fourth", etc. (if any) are used to distinguish similar objects and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0201] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some or all of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for smoothing a reference line for an automatic driving vehicle, characterized in that: include: Determine the points to be smoothed in the target reference line, wherein the points to be smoothed include a starting point and an end point; Determining a first position constraint, a first-order derivative constraint, and a second-order derivative constraint of the starting point, and determining a second position constraint and a heading angle constraint of the end point; Smoothing the starting point based on the first position constraint, the first-order derivative constraint, and the second-order derivative constraint, and smoothing the end point based on the second position constraint and the heading angle constraint; Obtain a smoothed target reference line based on the smoothed start point and the smoothed end point; The determining of the first position constraint, the first-order derivative constraint, and the second-order derivative constraint of the starting point includes: determining a first transverse coordinate constraint of the starting point according to the transverse coordinate position information of the starting point, a plurality of first optimization coefficients related to the transverse coordinate of the starting point, and initial distance information; determining a first longitudinal coordinate constraint of the starting point according to the longitudinal coordinate position information of the starting point, a plurality of second optimization coefficients related to the longitudinal coordinate of the starting point, and the initial distance information; Determining a first position constraint of the starting point according to the first transverse coordinate constraint and the first longitudinal coordinate constraint, and determining a first-order derivative constraint and a second-order derivative constraint of the starting point according to the first position constraint; The determining of the second position constraint and the heading angle constraint of the endpoint includes: determining a second abscissa constraint of the endpoint according to the abscissa position information of the endpoint, a plurality of third optimization coefficients related to the abscissa of the endpoint, and a distance between the endpoint and initial distance information; determining a second longitudinal coordinate constraint of the endpoint according to the longitudinal coordinate position information of the endpoint, a plurality of fourth optimization coefficients related to the longitudinal coordinate of the endpoint, and a distance between the endpoint and initial distance information; The second position constraint of the end point is determined according to the second horizontal coordinate constraint and the second vertical coordinate constraint, and the second position constraint is differentiated, and the heading angle constraint of the end point is determined according to the derivative result of the second position constraint and the sine value corresponding to the end point.

2. The method according to claim 1, characterized in that The step of determining the points to be smoothed in the reference line includes: Obtaining a first measurement data frame at the current moment and a second measurement data frame at the previous moment; The overlapping reference line between the first reference line corresponding to the first measurement data frame and the second reference line corresponding to the second measurement data frame is determined as a target reference line, and the starting point and the end point of the target reference line are determined as points to be smoothed on the target reference line.

3. The method according to claim 2, characterized in that Determining the first horizontal coordinate constraint of the starting point includes: constructing a first Nth-order polynomial about the initial distance information based on the horizontal coordinate position information of the starting point, a plurality of first optimization coefficients about the horizontal coordinate of the starting point, and initial distance information, and determining the first horizontal coordinate constraint of the starting point based on the first Nth-order polynomial; Determining the first ordinate constraint of the starting point includes: constructing a second Nth-order polynomial regarding the initial distance information based on the ordinate position information of the starting point, a plurality of second optimization coefficients regarding the ordinate of the starting point, and the initial distance information, and determining the first ordinate constraint of the starting point based on the second Nth-order polynomial; and / or, Determining the second horizontal coordinate constraint of the endpoint includes: constructing a third Nth-order polynomial regarding the distance based on the horizontal coordinate position information of the endpoint, a plurality of third optimization coefficients regarding the horizontal coordinate of the endpoint, and the distance between the endpoint and the initial distance information, and determining the second horizontal coordinate constraint of the endpoint based on the third Nth-order polynomial; The determining of the second longitudinal coordinate constraint of the end point includes: constructing a fourth Nth-order polynomial about the distance based on the longitudinal coordinate position information of the end point, a number of fourth optimization coefficients about the longitudinal coordinate of the end point, and the distance between the end point and the initial distance information, and determining the second longitudinal coordinate constraint of the end point based on the fourth Nth-order polynomial.

4. The method according to claim 2, characterized in that Also includes: Normalizing the initial distance information and the distance between the endpoint and the initial distance information according to a preset normalization coefficient, wherein the preset normalization coefficient is determined based on the initial distance information or the distance and the total length of the target reference line; The first horizontal coordinate constraint and the first vertical coordinate constraint of the starting point are respectively determined according to the normalized initial distance information, and the second horizontal coordinate constraint and the second coordinate constraint of the end point are respectively determined according to the normalized distance.

5. The method according to any one of claims 1 to 4, characterized in that Also includes: A planned path is generated based on the smoothed target reference line, and the planned path is sent to a target control device so that the target control device controls the autonomous driving vehicle to travel based on the planned path.

6. A vehicle automatic driving reference line smoothing device, characterized in that: include: A module for determining points to be smoothed, configured to determine points to be smoothed in a target reference line, wherein the points to be smoothed include a starting point and an end point; a constraint determination module configured to determine a first position constraint, a first-order derivative constraint, and a second-order derivative constraint of the starting point, and to determine a second position constraint and a heading angle constraint of the end point; a smoothing module configured to smooth the starting point based on the first position constraint, the first-order derivative constraint, and the second-order derivative constraint, and to smooth the end point based on the second position constraint and the heading angle constraint; and to obtain a smoothed target reference line based on the smoothed starting point and the smoothed end point; wherein, The smoothing module is specifically configured to determine a first transverse coordinate constraint of the starting point according to the transverse coordinate position information of the starting point, a plurality of first optimization coefficients related to the transverse coordinate of the starting point, and initial distance information; determining a first longitudinal coordinate constraint of the starting point according to the longitudinal coordinate position information of the starting point, a plurality of second optimization coefficients related to the longitudinal coordinate of the starting point, and the initial distance information; Determining a first position constraint of the starting point according to the first transverse coordinate constraint and the first longitudinal coordinate constraint, and determining a first-order derivative constraint and a second-order derivative constraint of the starting point according to the first position constraint; determining a second abscissa constraint of the endpoint according to the abscissa position information of the endpoint, a plurality of third optimization coefficients related to the abscissa of the endpoint, and a distance between the endpoint and initial distance information; determining a second longitudinal coordinate constraint of the endpoint according to the longitudinal coordinate position information of the endpoint, a plurality of fourth optimization coefficients related to the longitudinal coordinate of the endpoint, and a distance between the endpoint and initial distance information; The second position constraint of the end point is determined according to the second horizontal coordinate constraint and the second vertical coordinate constraint, and the second position constraint is differentiated, and the heading angle constraint of the end point is determined according to the derivative result of the second position constraint and the sine value corresponding to the end point.

7. An automatic driving system, characterized in that: Including planning device and control device: The planning device is used to obtain a smoothed target reference line according to the vehicle automatic driving reference line smoothing method according to any one of claims 1 to 5, and generate a planned path according to the smoothed target reference line; The control device is used to receive the planned path and control the autonomous driving vehicle to travel based on the planned path.

8. An electronic device, characterized in that: include: A memory, a processor and a computer program, wherein the computer program is stored in the memory, and the processor runs the computer program to execute the vehicle automatic driving reference line smoothing method as described in any one of claims 1 to 5.

9. A computer-readable storage medium, characterized in that The storage medium includes a computer program, which is used to implement the vehicle automatic driving reference line smoothing method as described in any one of claims 1 to 5.

Images