A reference line smoothing method, system, vehicle and storage medium

Abstract

The application discloses a reference line smoothing method, comprising the steps of: intercepting a first reference line segment on the original reference line according to the vehicle positioning, obtaining interpolation points by equidistant sampling, and constructing a first B-spline curve based on a cubic quasi-uniform B-spline; extending the reference line as the vehicle advances, intercepting a second reference line segment from the tail point of the first reference line segment, obtaining interpolation points by equidistant sampling, and constructing a second B-spline curve based on a cubic quasi-uniform B-spline; selecting at least part of the control vertices of the first B-spline curve and the second B-spline curve, and constructing a third B-spline curve; selecting a splicing point in the overlapping area of the first B-spline curve and the third B-spline curve, splicing the two curves, and finally obtaining the current smoothed reference line and outputting the same. The application also discloses a corresponding system, a storage medium and a vehicle. By implementing the application, a C2 continuous smoothed reference line can be obtained, and the application has the advantages of simple construction method, small calculation amount and stable structure.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving technology, and more specifically to a reference line smoothing method, system, vehicle, and storage medium. Background Art

[0002] The current mainstream reference line smoothing algorithms are the smoothing algorithm based on quintic polynomial spline curves, the smoothing algorithm based on double helix lines and the smoothing algorithm based on discrete points, but all of these existing reference line smoothing algorithms have shortcomings.

[0003] Among them, the smoothing algorithm based on the quintic polynomial spline curve is to fit the reference line through multiple segments of quintic polynomial spline curves. It considers factors such as the deviation between the points on the reference line and the spline curve, smoothness, etc., establishes the objective function and constraints in the form of quadratic programming, and completes the solution of the coefficients of multiple segments of quintic polynomials through the OSQP solver. Although the quintic polynomial spline curve can fit the reference line in a high-order manner and is continuous at the internal nodes C2 (that is, it is second-order continuous and differentiable at the splicing point), it only requires that the position error of the two reference lines at the splicing point is less than a limit value. When the splicing point C2 continuity is used as a constraint, the "Runge phenomenon" of the high-order polynomial causes the spline curve to have excessive deviations at the internal nodes, making such a spline curve unusable.

[0004] The double-helix-based reference line smoothing algorithm uses double-helix fitting as a reference. It not only considers factors such as node deviation and smoothness, but also adds curvature constraints. It can solve the problem of reference line smoothing for large angles. However, it is a nonlinear problem and converges slowly using the IPOPT solver, so it takes the longest time to run.

[0005] The reference line smoothing algorithm based on discrete points also uses the microscopic distribution of three adjacent discrete points to establish a quadratic programming model that considers factors such as node deviation, smoothness, and curvature. This model then directly obtains the coordinates of the discrete points through numerical solution. While this algorithm can account for curvature constraints and is computationally faster, the output discrete points do not meet strict second-order continuity requirements. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to provide a reference line smoothing method, system, vehicle and storage medium, which can obtain a C2 continuous smooth reference line and have the advantages of simple construction method, low calculation amount and stable structure.

[0007] To solve the above technical problems, as one aspect of the present invention, a reference line smoothing method is provided for autonomous driving, which comprises at least the following steps:

[0008] According to the vehicle positioning, a first segment of the reference line to be smoothed is intercepted on the original reference line, and interpolation points are obtained by equidistant sampling; a first B-spline curve is constructed based on a cubic quasi-uniform B-spline, the first B-spline curve passes through the interpolation points obtained by sampling the first segment of the reference line, and corresponding control vertices are obtained;

[0009] As the vehicle advances, the reference line is extended, a second segment of the reference line is intercepted starting from the tail point of the first segment of the reference line, and interpolation points are obtained by equidistant sampling; a second B-spline curve is constructed based on a cubic quasi-uniform B-spline, the second B-spline curve passes through the interpolation points obtained by sampling the second segment of the reference line, and corresponding control vertices are obtained;

[0010] At least part of the control vertices of the first B-spline curve and the second B-spline curve are selected, a third B-spline curve is constructed based on a cubic quasi-uniform B-spline according to the selected control vertices, and the first B-spline curve and the third B-spline curve have an overlapping area;

[0011] A splicing point is selected in the overlapping area, the first B-spline curve and the third B-spline curve are spliced based on the splicing point, and a latest smoothed reference line is formed;

[0012] A segment of a predetermined length is intercepted in the latest smoothed reference line as a current smoothed reference line based on the current positioning of the vehicle, and the current smoothed reference line is output.

[0013] The step of intercepting a first segment of the reference line to be smoothed on the original reference line according to the vehicle positioning and obtaining interpolation points by equidistant sampling further comprises:

[0014] A projection point of the vehicle positioning is obtained on the original reference line to be smoothed, and a first segment of the reference line is formed by cutting a predetermined length before and after the projection point;

[0015] A plurality of interpolation points are obtained on the first segment of the reference line by equidistant sampling.

[0016] The step of constructing a first B-spline curve based on a cubic quasi-uniform B-spline further comprises:

[0017] Constructing nodes in the cubic quasi-uniform B-spline:

[0018] Solving control vertices of the cubic quasi-uniform B-spline curve, the control vertices include interpolation points obtained by sampling the first segment of the reference line;

[0019] A predetermined formula is used to solve the coordinates of the points on the spline curve, thereby constructing the first B-spline curve.

[0020] The step of extending the reference line as the vehicle moves forward, extending the reference line forward from the end point of the first reference line as the starting point to intercept the second reference line, and sampling at equal intervals to obtain interpolation points further includes:

[0021] As the vehicle moves forward, when the distance between the projection point corresponding to the vehicle positioning and the end point of the first reference line is less than a set value, a predetermined length is intercepted from the original reference line to be smoothed as a second reference line, and the starting point of the interpolation point on the second reference line coincides with the end point of the interpolation point on the first reference line;

[0022] A plurality of interpolation points are obtained on the second reference line based on an equidistant sampling method.

[0023] The step of selecting at least some of the control vertices of the first B-spline curve and the second B-spline curve, and constructing a third B-spline curve based on the cubic quasi-uniform B-spline according to the selected control vertices further includes:

[0024] Selecting a first number of control vertices located in a rear column of the first B-spline curve, and selecting a second number of control vertices located in a front column of the second B-spline curve;

[0025] According to the selected control vertices, a third B-spline curve is constructed based on the cubic quasi-uniform B-spline.

[0026] The step of selecting a splicing point in the overlapping area and splicing the first B-spline curve and the third B-spline curve based on the splicing point to form the latest smooth reference line further includes:

[0027] A point in a smooth area of ​​the overlapping area is selected as a splicing point, and the first B-spline curve before the splicing point and the third B-spline curve after the splicing point are spliced ​​to form a latest reference line.

[0028] Another aspect of the present invention further provides a reference line smoothing system for autonomous driving, which at least comprises:

[0029] A first B-spline curve acquisition unit is configured to intercept a first reference line segment from the original reference line to be smoothed based on the vehicle positioning, and to obtain interpolation points by sampling at equal intervals; construct a first B-spline curve based on a cubic quasi-uniform B-spline, pass the first B-spline curve through the interpolation points sampled from the first reference line segment, and obtain corresponding control vertices;

[0030] a second B-spline curve acquisition unit, configured to extend the reference line as the vehicle moves forward, extend the reference line forward from the end point of the first reference line as the starting point to intercept the second reference line, and obtain interpolation points by sampling at equal intervals; construct a second B-spline curve based on a cubic quasi-uniform B-spline, pass the second B-spline curve through the interpolation points sampled from the second reference line, and obtain corresponding control vertices;

[0031] a third B-spline curve acquiring unit, configured to select at least some of the control vertices of the first B-spline curve and the second B-spline curve, and construct a third B-spline curve based on the selected control vertices and a cubic quasi-uniform B-spline, so that the first B-spline curve and the third B-spline curve have an overlapping area;

[0032] a splicing processing unit, configured to select a splicing point in the overlapping area, and splice the first B-spline curve and the third B-spline curve based on the splicing point to form an updated smooth reference line;

[0033] The smooth reference line output unit is used to intercept a section of a predetermined length from the latest smooth reference line based on the current positioning of the vehicle and output the section as the current smooth reference line.

[0034] Wherein, the first B-spline curve acquisition unit further includes:

[0035] A first reference line acquisition unit is configured to obtain a projection point of the vehicle positioning on the original reference line to be smoothed; and cut a predetermined length before and after the projection point to form a first reference line;

[0036] A first interpolation processing unit is configured to obtain a plurality of interpolation points on the first reference line based on an equidistant sampling method;

[0037] The first curve construction unit is used to construct a first B-spline curve based on a cubic quasi-uniform B-spline, so that the first B-spline curve passes through interpolation points sampled from the first reference line segment and obtains corresponding control vertices.

[0038] Wherein, the second B-sample curve acquisition unit further includes:

[0039] A distance judgment unit is used to judge whether the distance between the projection point corresponding to the vehicle positioning and the end point of the first reference line is less than a set value;

[0040] A second reference line acquisition unit is configured to, when the distance determination unit determines that the distance is less than , intercept a predetermined length of the original reference line to be smoothed as a second reference line, wherein the starting point of the interpolation point on the second reference line coincides with the ending point of the interpolation point on the first reference line;

[0041] A second interpolation processing unit is configured to obtain a plurality of interpolation points on the second reference line based on an equidistant sampling method;

[0042] The second curve construction unit is used to construct a second B-spline curve based on the cubic quasi-uniform B-spline, so that the second B-spline curve passes through the interpolation points sampled from the second reference line segment and obtains corresponding control vertices.

[0043] Wherein, the third B-sample curve acquisition unit further includes:

[0044] a control vertex selection unit, configured to select a first number of control vertices located in a rear column of the first B-spline curve, and select a second number of control vertices located in a front column of the second B-spline curve;

[0045] The third curve construction unit is configured to construct a third B-spline curve based on the selected control vertices and the cubic quasi-uniform B-spline, so that the first B-spline curve and the third B-spline curve have an overlapping area.

[0046] The splicing processing unit is specifically configured to select a point in a smooth area in the overlapping area as a splicing point, and splice a first B-spline curve before the splicing point with a third B-spline curve after the splicing point to form a latest reference line.

[0047] In another aspect of the present invention, a computer-readable storage medium is provided, on which a computer program is stored, wherein the computer program implements the steps of the reference line smoothing method as described above when executed by a processor.

[0048] In another aspect of the present invention, a vehicle is provided, comprising a reference line generating device, wherein the reference line generating device is integrated with the aforementioned reference line smoothing system.

[0049] The implementation of the embodiments of the present invention has the following beneficial effects:

[0050] The present invention provides a reference line smoothing method, system, vehicle, and storage medium. By constructing three cubic quasi-uniform B-spline curves, the resulting reference line can be C2-continuous at both nodes and internal points. Furthermore, the smoothing of the reference line is second-order continuous at the nodes and infinite-order continuous at internal points. Compared with existing quintic polynomial spline curves, the method effectively avoids the "Runge phenomenon," and the result obtained is an analytical solution that precisely passes through the interpolation points, with the advantages of low computational complexity and stable structure.

[0051] In this embodiment, the latter reference line is an extension of the former reference line, and is conveniently constructed based on a cubic quasi-uniform B-spline curve. It is at least second-order continuous at the splicing point, and the construction method is very simple.

[0052] In this embodiment, a reference line smoothing method based on a cubic quasi-uniform B-spline curve is used to interpolate the reference line sampling points, outputting a C2-continuous curve. The B-spline curve also controls vertex overlap, ensuring that the spliced ​​curves overlap within the sampling area and achieving C2 continuity at the splicing points. This method is applicable not only to reference line smoothing but also to path and velocity curve smoothing. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, without paying any creative work, other drawings obtained based on these drawings still fall within the scope of the present invention.

[0054] Figure 1 A schematic diagram of the main process of an embodiment of a reference line smoothing method provided by the present invention;

[0055] Figure 2 This is a schematic diagram of the principle of intercepting the first section of the reference line to be smoothed according to the present invention;

[0056] Figure 3 This is a schematic diagram of the principle of intercepting the second section of the reference line to be smoothed according to the present invention;

[0057] Figure 4 A schematic diagram of the principle of constructing a first B-spline curve according to the present invention;

[0058] Figure 5 A schematic diagram of the principle of constructing a second B-spline curve involved in the present invention;

[0059] Figure 6 A schematic diagram of the principle of constructing the third B-spline curve involved in the present invention;

[0060] Figure 7 A schematic structural diagram of an embodiment of a reference line smoothing system provided by the present invention;

[0061] Figure 8 for Figure 7 A schematic structural diagram of the first B-spline curve acquisition unit;

[0062] Figure 9 for Figure 7 A schematic structural diagram of the second B-spline curve acquisition unit;

[0063] Figure 10 for Figure 7 Schematic diagram of the structure of the third B-spline curve acquisition unit. DETAILED DESCRIPTION

[0064] In order to make the objectives, technical solutions and advantages of the present invention more clear, the present invention will be described in further detail below with reference to the accompanying drawings.

[0065] like Figure 1FIG. 1 is a schematic diagram showing the main process of an embodiment of a reference line smoothing method provided by the present invention. Figures 2 to 6 As shown, in this embodiment, the reference line smoothing method is used for autonomous driving, which at least includes the following steps:

[0066] Step S10: based on the vehicle positioning, intercept a first reference line segment from the original reference line to be smoothed, and sample interpolation points at equal intervals; construct a first B-spline curve based on a cubic quasi-uniform B-spline, pass the first B-spline curve through the interpolation points sampled from the first reference line segment, and obtain corresponding control vertices;

[0067] Please refer to Figure 2 The step S10 further includes:

[0068] The projection point of the vehicle positioning is obtained on the original reference line to be smoothed (hereinafter referred to as the original reference line); a predetermined length is cut before and after the projection point to form the first reference line (i.e., reference line 1); specifically, at the initial moment, the vehicle positioning point L is (1) Project a point onto the original reference line, and intercept the first section of the reference line to be smoothed on the original reference line according to the forward cutoff length l1 and the backward cutoff length l2;

[0069] Based on the equidistant sampling method, multiple interpolation points are obtained on the first reference line, and the interpolation point is obtained: p0 (1) ,p1 (1) ,…,p m1 (1) ;

[0070] It is understood that, in the present invention, a B-spline curve is constructed based on a cubic quasi-uniform B-spline, wherein the mathematical expression of the B-spline is:

[0071]

[0072] Among them, d i (i=0,1,…,n) is the control vertex, N i,k (u)(i=0,1,…,n) is the class-specific B-spline basis function, which is composed of a non-decreasing knot vector U: u0≤u1≤…≤u n+k+1 The determined k-order piecewise polynomial satisfies:

[0073]

[0074] It can be deduced from formula 2:

[0075]

[0076] It shows that p(u) is in the interval [u i ,ui+1 ] only with k+1 control vertices d i-k ,d i-k+1 ,…,d i Etc. and k+1 basis functions N i-k,k (u), N i-k+1,k (u),…,N i,k (u) related to.

[0077] In order to ensure that the first segment of the B-spline curve and the third segment of the B-spline curve coincide in the overlapping area, in addition to ensuring that the control vertices are the same, it is also necessary to ensure that the basis functions are equal in the overlapping area. Therefore, a quasi-uniform B-spline is used, and its node vector is expressed as:

[0078] u0 = u1 = ... = u k =0, u n+1 =u n+2 =…=u n+k+1 , u i =(ik) / (n-k+1)(when i=k+1,…,n+1) Equation 4

[0079] When the interpolation points p0, p1, ..., p m , construct a k-degree quasi-uniform B-spline curve passing through the interpolation point. The general steps are: first construct the node vector according to formula 4, where n = m + 2, and then inversely calculate the control vertex d i (i=0,1,…,n), and finally solve the coordinates of the point u∈[0,1] on the spline curve according to Equation 2 and Equation 1, as follows Figure 4 As shown. The difficulty lies in inversely finding the control vertex. Assuming that the interpolation point passes through the node, the equation group is established:

[0080]

[0081] There are n-k+2 equations and n+1 unknown vertices in Formula 5, so k-1 constraint equations are added, including tangent vector conditions, free endpoint conditions, virtual node conditions, parabola conditions and non-node conditions. The free endpoint conditions can be used in the present invention.

[0082] In a specific example of the present invention, the step of constructing the first B-spline curve based on the cubic quasi-uniform B-spline in step S10 further includes:

[0083] Step S100, constructing nodes in a cubic quasi-uniform B-spline: Specifically, since the first cubic quasi-uniform B-spline curve (i.e., the first B-spline curve) constructed in the present invention passes through the interpolation point p0 (1) ,p1 (1) ,…,p m1 (1) , then k=3, n1=m1+2, and the node is calculated according to the above formula 4:

[0084] u0=u1=u2=u3=0,u n1+1 =u n1+2 =u n1+3 =u n1+4 ,u i =(i-3) / (n1-2) (when i=4,…,n1+1).

[0085] Step S101, solving the control vertices of the cubic quasi-uniform B-spline curve, wherein the control vertices include interpolation points sampled from the first reference line segment;

[0086] Specifically, according to Equation 5, the control vertices of the cubic quasi-uniform B-spline curve are inversely solved. Since d0 (1) =p0 (1) ,d n1 (1) =p m1 (1) , the simplified equations are established as:

[0087]

[0088] Among them, the coefficient a i , b i , c i and e i It can be expressed as:

[0089]

[0090] Where Δu i =u i+1 -u i , i=2,…,n-2, where n=n1.

[0091] According to the free endpoint condition, it satisfies:

[0092]

[0093] The numerical solution of Equation 6 can be obtained by using the pursuit method, thereby obtaining the control vertices of the B-spline curve passing through the interpolation point.

[0094] Step S102: Using a predetermined formula to solve the coordinates of points on the spline curve, thereby constructing a first B-spline curve.

[0095] According to the above formula 1 and formula 2, we can get [u3,u n1-1 ]=[0,1], thereby obtaining the discrete points on the smooth reference line at the initial moment to construct the first B-spline curve. Specifically, the first B-spline curve can refer to Figure 4 shown.

[0096] Step S11: Extending the reference line as the vehicle moves forward, extending the reference line forward from the end point of the first reference line as the starting point to intercept a second reference line segment, and sampling interpolation points at equal intervals; constructing a second B-spline curve based on a cubic quasi-uniform B-spline, making the second B-spline curve pass through the interpolation points sampled from the second reference line segment, and obtaining corresponding control vertices;

[0097] As the vehicle moves forward, when the distance between the projection point corresponding to the vehicle positioning and the end point of the first reference line is less than the set value, the current reference line length is insufficient to meet the subsequent trajectory planning, so the reference line needs to be extended forward. Specifically, a predetermined length ( Figure 3 Δl in the figure is reserved as the second reference line (i.e., reference line 2), and the starting point of the interpolation point on the second reference line coincides with the end point of the interpolation point on the first reference line;

[0098] Based on the equidistant sampling method, multiple interpolation points are obtained on the second reference line, such as Figure 3 As shown, the interpolation point p0 is sampled from the equidistant (2) ,p1 (2) ,…,p m2 (2) Since the extended reference line is extended forward based on the end point of the reference line of the previous frame, the starting point of the interpolation point on the second reference line coincides with the end point of the interpolation point on the first reference line, that is, p0 (2) =p m (1) =d n1 (1) .

[0099] At the same time, the interpolation point p0 of the cubic quasi-uniform B-spline curve is known (2) ,p1 (2) ,…,p m2 (2) , the second segment of the cubic quasi-uniform B-spline curve (i.e., the second B-spline curve) can be constructed using the same method as in step S10, and its node vector is:

[0100] u0=u1=u2=u3=0,u n2+1 =u n2+2 =u n2+3 =u n2+4 ,u i =(i-3) / (n2-2)(when i=4,…,n2+1)

[0101] Its control vertices are:

[0102] d0 (2) =p0 (2) =p m (1) =d n1(1) ,d1 (2) ,…,d n2 (2) =p m2 (2) (n2=m2-2).

[0103] Specifically, the second B-spline curve can refer to Figure 5 shown.

[0104] Step S12: selecting at least some of the control vertices of the first B-spline curve and the second B-spline curve, and constructing a third B-spline curve based on the selected control vertices and a cubic quasi-uniform B-spline, so that the first B-spline curve and the third B-spline curve have an overlapping area;

[0105] Specifically, a first number of control vertices in the rear column of the first B-spline curve is selected, and a second number of control vertices in the front column of the second B-spline curve is selected; and a third B-spline curve is constructed based on the cubic quasi-uniform B-spline according to the selected control vertices.

[0106] It is understandable that in this step, in order to ensure that the two reference lines are at least C2 continuous at the splicing point, the local performance of the B-spline curve is used: at the node [u i ,u i+1 ] points within the range are only related to the control point p i-3 ,p i-2 ,p i-1 and p i Based on the cubic quasi-uniform B-spline curve, a third bridging B-spline curve is constructed, whose control vertex is {p i (3)}(i=0,1,…,n2+7), where the first 7 control points are obtained from the first B-spline curve control points. n1-7 (1) ,…,d n1-1 (1) , the next n2+1 control points are obtained from the second B-spline curve d0 (2) =d n1 (1) ,d1 (2) ,…,d n2 (2) , there are n2+8 control vertices {p i (3)}(i=0,1,…,n3)where n3=n2+7, such as Figure 6 As shown. The constructed node vector is:

[0107] u0 (3) =u1 (3) =u2 (3) =u3 (3)=0,u i (3) =(i-3) / (n3-2)(i=4,…,n3+1),u n3+1 (3) =u n3+2 (3) =u n3+3 (3) =u n3+4 (3)

[0108] According to the properties of the cubic quasi-uniform B-spline curve, the third B-spline curve is in the interval [u6 (3) ,u7 (3) ] and d n1-4 (1) ,d n1-3 (1) ,d n1-2 (1) ,d n1-1 (1) The four control vertices are related, and the basis function N 3,3 (u), N 4,3 (u), N 5,3 (u), N 6,3 (u) in [u6 (3) ,u7 (3) The value on ] is related to the basis function N of the first B-spline curve n1-6,3 (u), N n1-5,3 (u), N n1-4,3 (u), N n1-3,3 (u) In the curve [u n1-3 (1) ,u n1-2 (1) ] are equal, then the third B-spline curve is in the interval [u6 (3) ,u7 (3) ] and the first B-spline curve on the curve [u n1-3 (1) ,u n1-2 (1) ] on the interval [u 13 (3) ,u n3+1 (3) ] and the second B-spline curve in [u6 (2) ,u n2+1 (2) ] coincide with each other.

[0109] Step S13, selecting a splicing point in the overlapping area, and splicing the first B-spline curve and the third B-spline curve based on the splicing point to form a latest smooth reference line;

[0110] A point in a smooth area of ​​the overlapping area is selected as a splicing point, and the first B-spline curve before the splicing point and the third B-spline curve after the splicing point are spliced ​​to form a latest reference line.

[0111] It is understandable that since the third B-spline curve is in the interval [u7 (3) ,u 10 (3) ] and the control vertex of the first B-spline curve in the interval [u n1-2 (1) ,u n1+1 (1) ] have the same control vertices but different basis functions, then the third B-spline curve does not pass through the interpolation point p of the first B-spline curve m1-1 (1) , p m1 (1) , similarly, the third B-spline curve does not pass through the interpolation point p0 of the second B-spline curve (2) and p1 (2) In order to avoid excessive deviation, the points in the smooth area can be selected as the splicing points of the first B-spline curve and the second B-spline curve.

[0112] Splice the first B-spline curve and the third B-spline curve together, because the third B-spline curve is in the interval [u6 (3) ,u7 (3) ] coincides with the first B-spline curve, and the interval [u6 (3) ,u7 (3) ] is used as the joining point of the two curves, and u7 is used here. (3) Take as an example to introduce the process of splicing two curves.

[0113] Select the interval [u3 (1) =0,u n1-2 (1) ]=[0,1] as the front segment of the splicing curve; select the interval (u7 (3) ,u n3+1 (3) =1]=[0,1] as the back segment of the splicing curve, and then the two splicing curves are spliced ​​together to obtain a complete splicing smooth reference line.

[0114] Step S14 , based on the current positioning of the vehicle, a section of a predetermined length is intercepted from the latest smooth reference line as a current smooth reference line and outputted.

[0115] More specifically, for the reference line smoothing output at the initial moment, the first B-spline curve obtained above is used as the current smooth reference line; for the current smooth reference line output after extension, it is also necessary to obtain it by intercepting it from the latest smooth reference line based on the vehicle positioning and the lengths of the forward and backward reference lines.

[0116] like Figure 7 FIG. 1 is a schematic diagram showing a structure of an embodiment of a reference line smoothing system provided by the present invention. Figures 8 to 10 As shown, in this embodiment, the system 1 is used for autonomous driving, which at least includes:

[0117] A first B-spline curve acquisition unit 10 is configured to intercept a first reference line segment from the original reference line to be smoothed based on the vehicle positioning, and to obtain interpolation points by sampling at equal intervals; construct a first B-spline curve based on a cubic quasi-uniform B-spline, pass the first B-spline curve through the interpolation points sampled from the first reference line segment, and obtain corresponding control vertices;

[0118] A second B-spline curve acquisition unit 11 is configured to extend the reference line as the vehicle moves forward, extend the reference line forward from the end point of the first reference line as the starting point, and obtain interpolation points by sampling at equal intervals; construct a second B-spline curve based on a cubic quasi-uniform B-spline, pass the interpolation points sampled from the second reference line, and obtain corresponding control vertices;

[0119] a third B-spline curve acquiring unit 12, configured to select at least some of the control vertices of the first B-spline curve and the second B-spline curve, and construct a third B-spline curve based on the selected control vertices and a cubic quasi-uniform B-spline, so that the first B-spline curve and the third B-spline curve have an overlapping area;

[0120] A splicing processing unit 13 is configured to select a splicing point in the overlapping area, and splice the first B-spline curve and the third B-spline curve based on the splicing point to form a latest smooth reference line;

[0121] The smooth reference line output unit 14 is configured to intercept a section of a predetermined length from the latest smooth reference line based on the current position of the vehicle and output the section as a current smooth reference line.

[0122] like Figure 8 As shown, in one example, the first B-spline curve acquisition unit 10 further includes:

[0123] The first reference line acquisition unit 100 is used to obtain the projection point of the vehicle positioning on the original reference line to be smoothed; cut off a predetermined length before and after the projection point to form the first reference line;

[0124] A first interpolation processing unit 101 is configured to obtain a plurality of interpolation points on the first reference line based on an equidistant sampling method;

[0125] The first curve construction unit 102 is configured to construct a first B-spline curve based on a cubic quasi-uniform B-spline, so that the first B-spline curve passes through interpolation points sampled from the first reference line segment and obtains corresponding control vertices.

[0126] like Figure 9 As shown, in one example, the second B-sample curve acquisition unit 11 further includes:

[0127] The distance determination unit 110 is used to determine whether the distance between the projection point corresponding to the vehicle positioning and the end point of the first reference line is less than a set value;

[0128] A second reference line acquisition unit 111 is configured to, when the distance determination unit determines that the distance is less than , intercept a predetermined length of the original reference line to be smoothed as a second reference line, wherein the starting point of the interpolation point on the second reference line coincides with the ending point of the interpolation point on the first reference line;

[0129] A second interpolation processing unit 112 is configured to obtain a plurality of interpolation points on the second reference line based on an equidistant sampling method;

[0130] The second curve construction unit 113 is configured to construct a second B-spline curve based on the cubic quasi-uniform B-spline, so that the second B-spline curve passes through the interpolation points sampled from the second reference line segment and obtains corresponding control vertices.

[0131] like Figure 10 As shown, in one example, the third B-sample curve acquisition unit 12 further includes:

[0132] A control vertex selection unit 120 is configured to select a first number of control vertices located in a rear column of the first B-spline curve and a second number of control vertices located in a front column of the second B-spline curve;

[0133] The third curve construction unit 121 is configured to construct a third B-spline curve based on the selected control vertices and the cubic quasi-uniform B-spline, so that the first B-spline curve and the third B-spline curve have an overlapping area.

[0134] The splicing processing unit 13 is specifically configured to select a point in a smooth area in the overlapping area as a splicing point, and splice the first B-spline curve before the splicing point with the third B-spline curve after the splicing point to form a latest reference line.

[0135] In another aspect of the present invention, a computer-readable storage medium is provided, on which a computer program is stored, wherein the computer program is executed by a processor to implement the above-mentioned Figures 1 to 6 Describe the steps of the reference line smoothing method.

[0136] In another aspect of the present invention, a vehicle is provided, which includes a reference line generating device, wherein the reference line generating device is integrated with the aforementioned Figures 7 to 10 Describes the reference line smoothing system.

[0137] The implementation of the embodiments of the present invention has the following beneficial effects:

[0138] The present invention provides a reference line smoothing method, system, vehicle, and storage medium. By constructing three cubic quasi-uniform B-spline curves, the resulting reference line can be C2-continuous at both nodes and internal points. Furthermore, the smoothing of the reference line is second-order continuous at the nodes and infinite-order continuous at internal points. Compared with existing quintic polynomial spline curves, the method effectively avoids the "Runge phenomenon," and the result obtained is an analytical solution that precisely passes through the interpolation points, with the advantages of low computational complexity and stable structure.

[0139] In this embodiment, the latter reference line is an extension of the former reference line, and is conveniently constructed based on a cubic quasi-uniform B-spline curve. It is at least second-order continuous at the splicing point, and the construction method is very simple.

[0140] In this embodiment, a reference line smoothing method based on a cubic quasi-uniform B-spline curve is used to interpolate the reference line sampling points, outputting a C2-continuous curve. The B-spline curve also controls vertex overlap, ensuring that the spliced ​​curves overlap within the sampling area and achieving C2 continuity at the splicing points. This method is applicable not only to reference line smoothing but also to path and velocity curve smoothing.

[0141] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, apparatus, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0142] The computer program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other processing devices to cause a series of operational steps to be performed on the computer, other programmable apparatus or other processing devices to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide processes for implementing the functions specified in the flowchart block or blocks. Figure 1 The flowchart and / or block diagram in the variations disclosed herein illustrate the architecture, functionality, and operation of possible implementations of apparatuses and computer program products according to various embodiments. In this regard, each flowchart block and / or block in the flowcharts and / or block diagrams can represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical functions. It should also be noted that, in some alternative implementations, the flowchart blocks and / or blocks in the flowcharts and / or block diagrams can represent a Figure 1 The computer program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other processing devices to cause a series of operational steps to be performed on the computer, other programmable apparatus or other processing devices to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide processes for implementing the functions specified in the flowchart block or blocks.

[0143] The above-described embodiments are merely given as non-limiting examples. It will be understood by those skilled in the art that various modifications, changes, additions and omissions can be made without departing from the spirit and essential characteristics of the application. Therefore, the scope of the present application should be determined not by the foregoing description but by the appended claims.

Claims

1. A reference line smoothing method for autonomous driving, characterized in that: At least the following steps are included: Based on the vehicle positioning, a first reference line segment is intercepted on the original reference line to be smoothed, and interpolation points are obtained by sampling at equal intervals; a first B-spline curve is constructed based on a cubic quasi-uniform B-spline, and the first B-spline curve passes through the interpolation points sampled from the first reference line segment, and corresponding control vertices are obtained; As the vehicle moves forward, the reference line is extended, and a second reference line segment is extended forward from the end point of the first reference line segment as the starting point, and interpolation points are obtained by sampling at equal intervals; a second B-spline curve is constructed based on a cubic quasi-uniform B-spline, and the second B-spline curve is made to pass through the interpolation points sampled from the second reference line segment, and the corresponding control vertices are obtained; selecting at least some of the control vertices of the first B-spline curve and the second B-spline curve, and constructing a third B-spline curve based on the selected control vertices and a cubic quasi-uniform B-spline, so that the first B-spline curve and the third B-spline curve have an overlapping area; Selecting a splicing point in the overlapping area, and splicing the first B-spline curve and the third B-spline curve based on the splicing point to form a latest smooth reference line; Based on the current positioning of the vehicle, a section of a predetermined length is intercepted from the latest smooth reference line and output as a current smooth reference line.

2. The method according to claim 1, wherein The step of intercepting a first section of the reference line on the original reference line to be smoothed according to the vehicle positioning and obtaining interpolation points by equidistant sampling further includes: Obtaining a projection point of the vehicle positioning on the original reference line to be smoothed; cutting a predetermined length before and after the projection point to form a first reference line segment; A plurality of interpolation points are obtained on the first reference line based on an equidistant sampling method.

3. The method according to claim 2, wherein The step of constructing the first B-spline curve based on the cubic quasi-uniform B-spline further includes: Construct the nodes in a cubic quasi-uniform B-spline: Solving for control vertices of a cubic quasi-uniform B-spline curve, wherein the control vertices include interpolation points sampled from the first reference line segment; A predetermined formula is used to solve the coordinates of points on the spline curve, thereby constructing a first B-spline curve.

4. The method according to claim 1, wherein The step of extending the reference line as the vehicle moves forward, extending the reference line forward from the end point of the first reference line as the starting point to intercept the second reference line, and sampling at equal intervals to obtain interpolation points further includes: As the vehicle moves forward, when the distance between the projection point corresponding to the vehicle positioning and the end point of the first reference line is less than a set value, a predetermined length is intercepted from the original reference line to be smoothed as a second reference line, and the starting point of the interpolation point on the second reference line coincides with the end point of the interpolation point on the first reference line; A plurality of interpolation points are obtained on the second reference line based on an equidistant sampling method.

5. The method according to claim 4, wherein The step of selecting at least some of the control vertices of the first B-spline curve and the second B-spline curve, and constructing a third B-spline curve based on the cubic quasi-uniform B-spline according to the selected control vertices further includes: Selecting a first number of control vertices located in a rear column of the first B-spline curve, and selecting a second number of control vertices located in a front column of the second B-spline curve; According to the selected control vertices, a third B-spline curve is constructed based on the cubic quasi-uniform B-spline.

6. The method according to claim 5, wherein The step of selecting a splicing point in the overlapping area and splicing the first B-spline curve and the third B-spline curve based on the splicing point to form a latest smooth reference line further includes: A point in a smooth area of ​​the overlapping area is selected as a splicing point, and the first B-spline curve before the splicing point and the third B-spline curve after the splicing point are spliced ​​to form a latest reference line.

7. A reference line smoothing system for autonomous driving, characterized in that: At least: A first B-spline curve acquisition unit is configured to intercept a first reference line segment from the original reference line to be smoothed based on the vehicle positioning, and to obtain interpolation points by sampling at equal intervals; construct a first B-spline curve based on a cubic quasi-uniform B-spline, pass the first B-spline curve through the interpolation points sampled from the first reference line segment, and obtain corresponding control vertices; a second B-spline curve acquisition unit, configured to extend the reference line as the vehicle moves forward, extend the reference line forward from the end point of the first reference line as the starting point to intercept the second reference line, and obtain interpolation points by sampling at equal intervals; construct a second B-spline curve based on a cubic quasi-uniform B-spline, pass the second B-spline curve through the interpolation points sampled from the second reference line, and obtain corresponding control vertices; a third B-spline curve acquiring unit, configured to select at least some of the control vertices of the first B-spline curve and the second B-spline curve, and construct a third B-spline curve based on the selected control vertices and a cubic quasi-uniform B-spline, so that the first B-spline curve and the third B-spline curve have an overlapping area; a splicing processing unit, configured to select a splicing point in the overlapping area, and splice the first B-spline curve and the third B-spline curve based on the splicing point to form an updated smooth reference line; The smooth reference line output unit is used to intercept a section of a predetermined length from the latest smooth reference line based on the current positioning of the vehicle and output the section as the current smooth reference line.

8. The system according to claim 7, wherein: The first B-spline curve acquiring unit further comprises: A first reference line acquisition unit is configured to obtain a projection point of the vehicle positioning on the original reference line to be smoothed; and cut a predetermined length before and after the projection point to form a first reference line; A first interpolation processing unit is configured to obtain a plurality of interpolation points on the first reference line based on an equidistant sampling method; The first curve construction unit is used to construct a first B-spline curve based on a cubic quasi-uniform B-spline, so that the first B-spline curve passes through interpolation points sampled from the first reference line segment and obtains corresponding control vertices.

9. The system according to claim 8, wherein The second B-sample curve acquisition unit further includes: A distance judgment unit is used to judge whether the distance between the projection point corresponding to the vehicle positioning and the end point of the first reference line is less than a set value; A second reference line acquisition unit is configured to, when the distance determination unit determines that the distance is less than , intercept a predetermined length of the original reference line to be smoothed as a second reference line, wherein the starting point of the interpolation point on the second reference line coincides with the ending point of the interpolation point on the first reference line; A second interpolation processing unit is configured to obtain a plurality of interpolation points on the second reference line based on an equidistant sampling method; The second curve construction unit is used to construct a second B-spline curve based on the cubic quasi-uniform B-spline, so that the second B-spline curve passes through the interpolation points sampled from the second reference line segment and obtains corresponding control vertices.

10. The system according to claim 9, wherein: The third B-sample curve acquisition unit further includes: a control vertex selection unit, configured to select a first number of control vertices located in a rear column of the first B-spline curve, and select a second number of control vertices located in a front column of the second B-spline curve; The third curve construction unit is configured to construct a third B-spline curve based on the selected control vertices and the cubic quasi-uniform B-spline, so that the first B-spline curve and the third B-spline curve have an overlapping area.

11. The system according to claim 10, wherein: The splicing processing unit is specifically configured to select a point in a smooth area in the overlapping area as a splicing point, and splice a first B-spline curve before the splicing point with a third B-spline curve after the splicing point to form a latest reference line.

12. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the reference line smoothing method according to any one of claims 1 to 6 are implemented.

13. A vehicle, characterized in that: It comprises a reference line generating device, in which the reference line smoothing system as described in any one of claims 7 to 11 is integrated.

Images