Method, apparatus and electronic device for determining reference line in autonomous driving

Abstract

The present application provides a method, apparatus, and electronic device for determining a reference line in autonomous driving. The method includes: obtaining lane sequence information corresponding to the current position of the vehicle, that is, the position, orientation, and boundary information corresponding to multiple anchor points; inputting the lane sequence information into a preset reference line smoothing problem-solving model for solution to obtain a state sequence and a control output sequence, and further determining the reference line corresponding to the lane sequence information. Among them, the above model is an optimal control solution model under preset constraint conditions; the objective function is: the integral of the curvature and the derivative of the curvature of the entire curve segment corresponding to the reference line is minimized; the preset constraint conditions include: the starting position and the ending position of the entire curve segment are respectively consistent with the positions of the corresponding anchor points, the state relationship between spline curve segments satisfies the system state transition equation derived from the quintic polynomial state equation, and the anchor point boundary constraint conditions. The present application can improve the determination accuracy and efficiency of the reference line in autonomous driving at the same time.

Description

Technical Field

[0001] The present application relates to the technical field of autonomous driving, and in particular, to a method, an apparatus, and an electronic device for determining a reference line in autonomous driving. Background Art

[0002] Designing a reference line smoothing algorithm that takes into account both efficiency and accuracy is one of the difficulties in the field of autonomous driving. First, since many current autonomous driving motion planning algorithms are based on the frenet coordinate system, and the deviation between the frenet coordinate system and the Cartesian coordinate system depends on the smoothness of the reference line, the quality of the trajectory output by the planning module also strongly depends on the quality of the reference line. In addition, autonomous driving is a large system, but the computing power and resources of the in-vehicle computing platform are limited. The reference line module is required to minimize the occupation of computing resources and efficiently and stably generate high-quality reference lines.

[0003] Among the existing reference line smoothing algorithms, there are Bezier curve algorithms, piecewise spline curve algorithms, multi-stage helix algorithms, discrete point smoothing algorithms, etc. Among them, the multi-stage helix algorithm has many parameters and requires frequent use of the Newton shooting method, resulting in low algorithm efficiency; the Bezier curve algorithm and the discrete point smoothing algorithm are difficult to output a reference line with stable curvature and high quality, and the curvature fluctuates or mutates; in the piecewise spline curve algorithm, the objective function and the constraint matrix in the constructed quadratic programming problem are huge and sparse, resulting in high spatial complexity and low solution efficiency.

[0004] In summary, it is difficult for the existing reference line smoothing algorithms to take into account both efficiency and accuracy at the same time. Summary of the Invention

[0005] The purpose of the present application is to provide a method, an apparatus, and an electronic device for determining a reference line in autonomous driving, which can improve the accuracy and efficiency of determining the reference line in autonomous driving at the same time.

[0006] In a first aspect, an embodiment of the present application provides a method for determining a reference line in autonomous driving. The method includes: obtaining lane sequence information corresponding to the current position of the vehicle; the lane sequence information includes: the positions, orientations, and boundary information corresponding to multiple anchor points; inputting the lane sequence information into a preset reference line smoothing problem-solving model for solution to obtain a state sequence and a control output sequence; wherein, the reference line smoothing problem-solving model is an optimal control solution model under preset constraint conditions; the objective function corresponding to the model is: the integral of the curvature and the derivative of the curvature of the entire curve segment corresponding to the reference line is minimized; the entire curve segment is divided into multiple spline curve segments by multiple anchor points; the system state of each spline curve segment is characterized by a fifth-degree polynomial; the preset constraint conditions include: the starting position and the ending position of the entire curve segment are respectively consistent with the positions of the corresponding anchor points, the state relationship between the spline curve segments satisfies the system state transition equation derived from the fifth-degree polynomial state equation, and the anchor point boundary constraint conditions; determining the reference line corresponding to the lane sequence information based on the state sequence and the control output sequence.

[0007] Further, the step of obtaining the lane sequence information corresponding to the current position of the vehicle includes: determining the map information within a specified range according to the current position of the vehicle; performing equidistant sampling on the lane lines corresponding to the map information to obtain multiple anchor points; obtaining the position, orientation, and boundary information corresponding to each anchor point; the boundary information includes the front and rear boundaries and the left and right boundaries.

[0008] Further, in the case where the spline curve segments need to be spliced, the preset constraint conditions further include: the orientation of the ending position of the entire curve segment is consistent with the orientation of the corresponding anchor point; the orientation, curvature, and derivative of the curvature corresponding to the starting position of the entire curve segment are consistent with the initial state of the corresponding anchor point.

[0009] Further, the system state transition equation is as follows:

[0010]

[0011] Wherein,

[0012]

[0013]

[0014] x i (s) = a i0 + a i1 s + a i2 s 2 + a i3 s 3 + a i4 s 4 + a i5 s 5;

[0015] y i (s) = b i0 +b i1 s + b i2 s 2 +b i3 s 3 +b i4 s 4 +b i5 s 5 ;

[0016]

[0017]

[0018] wherein, and respectively represent the system states corresponding to the (i + 1)-th and i-th spline curve segments; respectively represent the abscissa system state and ordinate system state corresponding to the i-th spline curve segment; represents the control output corresponding to the i-th spline curve segment; respectively represent the abscissa control output and ordinate control output corresponding to the i-th spline curve segment; s is the arc length of each spline curve segment; x i (s), y i (s) respectively represent the abscissa fifth-order polynomial state equation and ordinate fifth-order polynomial state equation corresponding to the i-th spline curve segment; a i0 , a i1 , a i2 , a i3 , a i4 , a i5 , b i0 , b i1 , b i2 , b i3 , b i4 , b i5 are all coefficients of the spline curve segment.

[0019] Furthermore, the above objective function is as follows:

[0020]

[0021] wherein, respectively represent the second derivatives of the abscissa fifth-order polynomial state equation and ordinate fifth-order polynomial state equation corresponding to the i-th spline curve; respectively represent the third-order derivatives of the fifth-degree polynomial state equation of the abscissa and the fifth-degree polynomial state equation of the ordinate corresponding to the i-th spline curve; ω1 and ω2 are the weights corresponding to the second-order derivative term and the third-order derivative term respectively; i = 0, 1, ..., N, where N is a positive integer.

[0022] Further, the above anchor boundary constraint means that the position at the corresponding arc length is within a rectangular area consistent with the orientation of the anchor point; the rectangular area is determined by the boundary information of the anchor point; the formula corresponding to the anchor boundary constraint is as follows:

[0023]

[0024]

[0025] where, Δ l , Δ r , Δ b , Δ f respectively represent the left boundary, right boundary, front boundary and rear boundary corresponding to the anchor point; represents the position of the j-th anchor point, represents the orientation of the j-th anchor point, j = 1, 2, ..., M, where M is the total number of anchor points; x i (s j ), y i (s j ) respectively represent the abscissa and ordinate of the reference point corresponding to the j-th anchor point.

[0026] Further, the above steps of determining the reference line corresponding to the lane sequence information based on the state sequence and the control output sequence include: substituting the state sequence and the control output sequence into formulas (2)-(6) to obtain the spline coefficients corresponding to each spline curve segment; performing equal-arc-length interpolation according to the spline coefficients corresponding to each spline curve segment and the fifth-degree polynomial state equation, and solving for the position, orientation, curvature and derivative of the curvature on the reference line to determine the reference line corresponding to the lane sequence information.

[0027] Further, after the above steps of determining the reference line corresponding to the lane sequence information, the method further includes: performing post-processing operations on the result of the reference line; the post-processing operations include at least one of the following: sampling, duplicate removal, feasibility check, constructing reference line-related objects.

[0028] Second aspect, an embodiment of the present application further provides a device for determining a reference line in autonomous driving. The device includes: an information acquisition module, configured to acquire lane sequence information corresponding to the current position of the vehicle; the lane sequence information includes: positions, orientations, and boundary information respectively corresponding to a plurality of anchor points; a solution module, configured to input the lane sequence information into a preset reference line smoothing problem solution model for solution to obtain a state sequence and a control output sequence; wherein, the reference line smoothing problem solution model is an optimal control solution model under preset constraint conditions; the objective function corresponding to the model is: the integral of the curvature and the derivative of the curvature of the entire curve segment corresponding to the reference line is minimized; the entire curve segment is divided into a plurality of spline curve segments by a plurality of anchor points; the system state of each spline curve segment is characterized by a fifth-degree polynomial; the preset constraint conditions include: the starting position and the ending position of the entire curve segment are respectively consistent with the positions of the corresponding anchor points, the state relationship between the spline curve segments satisfies the system state transition equation derived from the fifth-degree polynomial state equation, and the anchor point boundary constraint conditions; a reference line determination module, configured to determine a reference line corresponding to the lane sequence information based on the state sequence and the control output sequence.

[0029] Third aspect, an embodiment of the present application further provides an electronic device, including a processor and a memory, where the memory stores computer-executable instructions that can be executed by the processor, and the processor executes the computer-executable instructions to implement the method described in the first aspect above.

[0030] In the method, device, and electronic device for determining a reference line in autonomous driving provided by the embodiments of the present application, first, lane sequence information corresponding to the current position of the vehicle is acquired; the lane sequence information includes: positions, orientations, and boundary information respectively corresponding to a plurality of anchor points; the lane sequence information is input into a preset reference line smoothing problem solution model for solution to obtain a state sequence and a control output sequence; wherein, the reference line smoothing problem solution model is an optimal control solution model under preset constraint conditions; the objective function corresponding to the model is: the integral of the curvature and the derivative of the curvature of the entire curve segment corresponding to the reference line is minimized; the entire curve segment is divided into a plurality of spline curve segments by a plurality of anchor points; the system state of each spline curve segment is characterized by a fifth-degree polynomial; the preset constraint conditions include: the starting position and the ending position of the entire curve segment are respectively consistent with the positions of the corresponding anchor points, the state relationship between the spline curve segments satisfies the system state transition equation derived from the fifth-degree polynomial state equation, and the anchor point boundary constraint conditions; a reference line corresponding to the lane sequence information is determined based on the state sequence and the control output sequence. In the embodiments of the present application, the reference line smoothing problem is described as an optimal control problem and solved using the mathematical form of model predictive control. It is a reference line smoothing algorithm with relatively low time complexity and space complexity, which can output a high-quality reference line with smooth curvature and low peaks, meet the requirements of the planning and control modules, reduce resource occupancy, and improve efficiency at the same time. Description of the Drawings

[0031] To more clearly illustrate the specific embodiments of the present application or the technical solutions in the prior art, the following will briefly introduce the drawings required for use in the description of the specific embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0032] Figure 1 It is a flowchart of a method for determining a reference line in an autonomous driving provided by an embodiment of the present application;

[0033] Figure 2 It is a block diagram for describing an optimal control problem provided by an embodiment of the present application;

[0034] Figure 3 It is a flowchart of another method for determining a reference line in an autonomous driving provided by an embodiment of the present application;

[0035] Figure 4 It is a block diagram of the structure of a device for determining a reference line in an autonomous driving provided by an embodiment of the present application;

[0036] Figure 5 It is a schematic diagram of the structure of an electronic device provided by an embodiment of the present application. Specific embodiments

[0037] The following will clearly and completely describe the technical solutions of the present application in conjunction with the embodiments. Obviously, the described embodiments are some, rather than all, of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present application.

[0038] In the existing reference line smoothing algorithms, in the multi-stage spiral line algorithm, there are many parameters and the Newton shooting method needs to be frequently used, resulting in low reference line smoothing efficiency; in the Bezier curve algorithm and the discrete point smoothing algorithm, the curvature fluctuates or mutates, and it is difficult to output a reference line with stable curvature and high quality; in the piecewise spline curve algorithm, the objective function and the constraint matrix in the constructed quadratic programming problem are huge and sparse, with high space complexity and low solution efficiency. It is difficult for the existing reference line smoothing algorithms to simultaneously take into account both efficiency and accuracy. Based on this, the embodiments of the present application provide a method, device, and electronic device for determining a reference line in an autonomous driving, which can improve the accuracy and efficiency of determining the reference line in an autonomous driving at the same time. For ease of understanding of this embodiment, first, a method for determining a reference line in an autonomous driving disclosed in the embodiments of the present application will be introduced in detail.

[0039] Figure 1A method for determining a reference line in autonomous driving provided by an embodiment of the present application, the method includes the following steps:

[0040] Step S102, obtaining lane sequence information corresponding to the current position of the vehicle; the lane sequence information includes: positions, orientations, and boundary information respectively corresponding to a plurality of anchor points.

[0041] In this step, the above lane sequence information is information obtained by equally spaced sampling on the lane line within a specified range with the current position of the vehicle as the reference. For example, lane line anchor point information is extracted from the map information within a range of 50 meters backward and 300 meters forward of the current position of the vehicle. The position corresponding to the anchor point can be represented by the abscissa and ordinate, the orientation corresponding to the anchor point can be obtained by position difference, and the boundary information corresponding to the anchor point can include the front boundary, rear boundary, left boundary, and right boundary.

[0042] Step S104, inputting the lane sequence information into a preset reference line smoothing problem solving model for solution to obtain a state sequence and a control output sequence; wherein, the reference line smoothing problem solving model is an optimal control solving model under preset constraint conditions; the objective function corresponding to the model is: the integral of the curvature and the derivative of the curvature of the entire curve segment corresponding to the reference line is minimized; the entire curve segment is divided into multiple spline curve segments by a plurality of anchor points; the system state of each spline curve segment is characterized by a fifth-degree polynomial; the preset constraint conditions include: the starting position and the ending position of the entire curve segment are respectively consistent with the positions of the corresponding anchor points, the state relationship between the spline curve segments satisfies the system state transition equation derived from the fifth-degree polynomial state equation, and the anchor point boundary constraint conditions.

[0043] In an embodiment of the present application, the reference line smoothing problem is described as an optimal control problem, that is, the above reference line smoothing problem solving model is an optimal control solving model under preset constraint conditions. Since the smaller the curvature and the derivative of the curvature of the reference line, the smoother the reference line, the objective function corresponding to the model is that the integral of the curvature and the derivative of the curvature of the entire curve segment corresponding to the reference line is minimized. In addition, in order to ensure the smoothness of the reference line, and the points on the reference line cannot deviate too far from the current lane, and to ensure that the algorithm quickly solves to obtain a smooth result, the preset constraint conditions in the embodiment of the present application include the following conditions:

[0044] The starting position and the ending position of the entire curve segment are consistent with the positions of the corresponding anchor points;

[0045] The state relationship between the spline curve segments satisfies the system state transition equation derived from the fifth-degree polynomial state equation;

[0046] The anchor point boundary constraint conditions, that is, the position at the corresponding arc length is within a rectangular area consistent with the orientation of the anchor point; the rectangular area is determined by the boundary information of the anchor point.

[0047] In this embodiment, a model predictive control mathematical form is used for solving, which is a reference line smoothing algorithm with relatively low time complexity and space complexity. It can output a high-quality reference line with smooth curvature and low peaks, meeting the requirements of the planning and control modules, reducing resource occupation while improving efficiency.

[0048] Step S106: Based on the state sequence and the control output sequence, determine the reference line corresponding to the lane sequence information.

[0049] Based on the above model, by performing backward inference and solution through the obtained state sequence and control output sequence, the reference line corresponding to the lane sequence information can be determined.

[0050] In the method for determining a reference line in an autonomous driving provided by an embodiment of the present application, first, obtain the lane sequence information corresponding to the current position of the vehicle, that is, the position, orientation, and boundary information corresponding to multiple anchor points; then input the lane sequence information into a preset reference line smoothing problem solving model for solution to obtain a state sequence and a control output sequence; among them, the reference line smoothing problem solving model is an optimal control solving model under preset constraint conditions; the objective function corresponding to the model is: the integral of the curvature and the derivative of the curvature of the entire curve segment corresponding to the reference line is minimized; the entire curve segment is divided into multiple spline curve segments by multiple anchor points; the system state of each spline curve segment is characterized by a fifth-degree polynomial; the preset constraint conditions include: the starting position and the ending position of the entire curve segment are respectively consistent with the positions of the corresponding anchor points, the state relationship between the spline curve segments satisfies the system state transition equation derived from the fifth-degree polynomial state equation, and the anchor point boundary constraint conditions. Finally, based on the state sequence and the control output sequence, determine the reference line corresponding to the lane sequence information. In the embodiment of the present application, the reference line smoothing problem is described as an optimal control problem and is solved using a model predictive control mathematical form, which is a reference line smoothing algorithm with relatively low time complexity and space complexity. It can output a high-quality reference line with smooth curvature and low peaks, meeting the requirements of the planning and control modules, reducing resource occupation while improving efficiency.

[0051] An embodiment of the present application also provides another method for determining a reference line in an autonomous driving. This method is implemented based on the previous embodiment. This embodiment focuses on describing the acquisition process of the lane sequence information, the model construction process, and the specific content.

[0052] The acquisition process of the lane sequence information is as follows:

[0053] (1) According to the current position of the vehicle, determine the map information within a specified range.

[0054] In specific implementation, the corresponding position on the map is found according to the current position of the vehicle. Since the planning and control module needs to limit the speed according to the lane curvature information, and the time domain of the planning module is usually about 8s, the map information 50 meters backward and 300 meters forward from the current position of the vehicle is read as the basic data.

[0055] (2) Perform equally spaced sampling on the lane lines corresponding to the map information to obtain multiple anchor points; obtain the position, orientation, and boundary information corresponding to each anchor point; the boundary information includes the front and rear boundaries and the left and right boundaries.

[0056] According to the above map information, lane sequence information is extracted to obtain a series of continuous lane sequences. For example, for the lane lines on the map information within 350 meters, coarse-grained sampling is performed at an interval of 25 meters, that is, sampling is performed every 25 meters to obtain an anchor point, and the position, orientation, and boundary information of the anchor point are obtained. Since the map is stored in the form of polyline segments, there is a certain error in this sampling. Among them, the position directly takes the center point of the lane boundary, and the orientation is obtained by position difference, that is, calculated according to the horizontal and vertical coordinates of adjacent two points; the boundary information is obtained by subtracting half of the vehicle width and the safety buffer from the lane boundary. For example, the left boundary is 20 cm, the right boundary is 20 cm, and the front and rear boundaries are both 2 meters. This boundary information can be set differently according to different vehicle type requirements, or can be scaled according to the vehicle type and road conditions.

[0057] Through the above sampling, for the 350-meter lane line, sampling at an interval of 25 meters can obtain 15 anchor points, and the reference line between adjacent two anchor points can be regarded as a spline curve segment.

[0058] The model construction process is as follows:

[0059] In the embodiment of the present application, the reference line smoothing problem is described as an optimal control problem. Given the initial state and the target state, through a series of control sequences u i , the control object reaches the target state under certain constraint conditions and the cost is minimized. As Figure 2 shown.

[0060] Considering the requirements of high-order smoothness and fitting ability of the reference line, a system state transition model is derived based on the fifth-degree polynomial. Let the state of the system be x i (s), y i (s). Taking x i (s) as an example, the change of the state is described by the fifth-degree polynomial, then there is the following formula (1), where s is the arc length of each spline curve segment, corresponding to the time domain length t of the optimal control problem. To avoid ambiguity, it is expressed as s here.

[0061] xi (s) = a i0 + a i1 s + a i2 s 2 + a i3 s 3 + a i4 s 4 + a i5 s 5 ;

[0062]

[0063]

[0064]

[0065]

[0066]

[0067] In the above formula (1), the starting coordinate corresponding to the i-th spline curve segment is x i , y i . Taking x i as an example, there is the following formula (2):

[0068] x i = x i (0) = a i0 ;

[0069]

[0070]

[0071]

[0072]

[0073]

[0074] Let Then the state transition equation of the system can be written as the following formula (3):

[0075] Similarly,

[0076] Since there is the following formula (4):

[0077]

[0078] According to formula (2), there is the following formula (5):

[0079]

[0080] Therefore, the system state transition equation can be expressed as Equation (6):

[0081]

[0082] where

[0083] Thus, by using the above system state transition equation, the generated reference line satisfies fifth-order continuity within a segment and second-order continuity between segments. In summary, the reference line smoothing problem can be written in the following mathematical form of an optimal control problem (7):

[0084]

[0085] s.t. x0(s0) = x init ; y0(s0) = y init ;

[0086] x N (s N ) = x goal ; y N (s N ) = y goal ;

[0087]

[0088]

[0089]

[0090]

[0091]

[0092]

[0093]

[0094] In the above mathematical form (7), the first expression is the objective function, which represents the respectively represent the second derivatives of the fifth-degree polynomial state equations of the abscissa and ordinate corresponding to the i-th spline curve; respectively represent the third-order derivatives of the fifth-degree polynomial state equations of the abscissa and ordinate corresponding to the $i$-th spline curve; $\omega_1$ and $\omega_2$ are the weights corresponding to the second-order derivative term and the third-order derivative term respectively; $i = 0, 1, \cdots, N$, where $N$ is a positive integer representing the number of spline curve segments. The objective function contains quadratic terms and cubic terms, representing the integrals of the kappa and dkappa of the curve, and it is expected that the curve is as smooth as possible. represents the system state corresponding to the $i$-th spline curve segment; represents the control output corresponding to the $i$-th spline curve segment; thus, the solution result is the state sequence and the control output sequence

[0095] s.t. are equality or inequality constraints, specifically including the following 9 constraint equations, namely Constraint One, Constraint Two... Constraint Nine. Constraint One and Constraint Two indicate that the starting point and the target point positions of the entire curve segment are consistent with the anchor point positions; $x_0(s_0), y_0(s_0)$ represent the initial positions of the entire curve segment corresponding to the reference line, $x$ init , $y$ init represent the anchor point positions corresponding to the initial positions of the entire curve segment. $x$ N $(s$ N ), $y$ N $(s$ N ) represent the end positions of the entire curve segment corresponding to the reference line, $x$ goal , $y$ goal represent the anchor point positions corresponding to the end positions of the entire curve segment.

[0096] Constraint Three represents the system state transition equation, which describes the state relationship of each sub-spline curve segment. Since the system equation is derived from a fifth-degree polynomial curve, there is fifth-order continuity within each spline curve segment. Since the system is a third-order integral system, it implies that the second-order continuity is satisfied at the connection points of two spline curves, representing the curvature continuity at the connection points. In the system state transition equation,

[0097]

[0098]

[0099]

[0100] $x$ i $(s)=a$ i0 $+a$ i1 $s + a$ i2 $s$ 2 $+ a$ i3 $s$ 3 $+ a$ i4 $s$ 4 $+ a$ i5s 5 ;

[0101] y i (s) = b i0 +b i1 s + b i2 s 2 +b i3 s 3 +b i4 s 4 +b i5 s 5 ;

[0102]

[0103]

[0104] wherein, and respectively represent the system states corresponding to the (i + 1)-th and i-th spline curve segments; respectively represent the abscissa system state and ordinate system state corresponding to the i-th spline curve segment; represents the control output corresponding to the i-th spline curve segment; respectively represent the abscissa control output and ordinate control output corresponding to the i-th spline curve segment; s is the arc length of each spline curve segment; x i (s), y i (s) respectively represent the abscissa fifth-degree polynomial state equation and ordinate fifth-degree polynomial state equation corresponding to the i-th spline curve segment; a i0 , a i1 , a i2 , a i3 , a i4 , a i5 , b i0 , b i1 , b i2 , b i3 , b i4 , b i5 are all coefficients of the spline curve segment.

[0105] Constraints Four and Five represent the bounding box boundary constraints at the anchor points, that is, the position at the corresponding arc length should be within the rectangle consistent with the orientation of the anchor point. Δ l , Δ r , Δ b , Δ f respectively represent the left boundary, right boundary, front boundary and rear boundary corresponding to the anchor point; represents the position of the j-th anchor point, represents the orientation of the j-th anchor point, j = 1, 2,..., M, where M is the total number of anchor points; xi (s j ) and y i (s j ) represent the abscissa and ordinate of the reference point corresponding to the j-th anchor point respectively.

[0106] Constraint six indicates that the orientation of the end point of the entire curve segment is the same as that of the last anchor point. represents the first derivative of the anchor point position corresponding to the end point position of the entire curve segment, that is, the orientation; θ goal represents the orientation of the end point position of the entire curve segment.

[0107] Constraints seven, eight, and nine indicate that the first, second, and third derivatives at the starting point of the entire curve segment are the same as the initial state of the anchor point. Since the last four constraints are only used in the splicing case, the high-order derivative information of the initial state of the anchor point is taken from the reference line of the previous frame.

[0108] In the above solution model, the input data is that is, the position and orientation of the anchor point, and the output data is When running the reference line smoothing algorithm in autonomous driving, in order to solve for a shorter length, there is usually a splicing logic. Each time a newly added section is smoothed, it needs to be spliced with the reference line of the previous frame. At this time, the last four constraints in the above model are required.

[0109] After determining the above solution model, call the relevant solver for optimal control to solve and obtain the state sequence and control output sequence. Using the solution result, the parametric spline coefficients of each spline curve are inversely solved according to Equation (2). Then, equi-arc length interpolation is performed to solve for information such as position, orientation, curvature kappa, and dkappa on the reference line.

[0110] Finally, post-processing can also be performed on the solved reference line results, including sampling, duplicate removal (removing points that are too close), feasibility checking (checking whether it deviates too far from the original anchor point), and constructing reference line-related objects (constructing objects containing information such as reference line position, parameters, and length for downstream path planning and obstacle projection). Finally, the reference line is output.

[0111] Figure 3 The flowchart of another method for determining the reference line in autonomous driving is shown, specifically including the following steps:

[0112] 1. Read the corresponding map according to the positioning information;

[0113] 2. Calculate the anchor point information, including position, orientation, boundary, etc.;

[0114] 3. Construct the reference line smoothing problem dimensions, including state, time domain length, and system input;

[0115] 4. Construct equality constraints, including the system state transition equation, initial and target states;

[0116] 5. Construct inequality constraints, including the bounding box;

[0117] 6. Construct the objective function and related weights;

[0118] 7. Call the solver to solve;

[0119] 8. Use the solution result to calculate the spline curve parameters;

[0120] 9. Perform equal arc length interpolation to solve information such as the position, orientation, and curvature of the reference line;

[0121] 10. Post-process the reference line, remove duplicates, and construct reference line-related objects;

[0122] 11. Output the reference line.

[0123] The method for determining the reference line in autonomous driving provided by the embodiments of the present application has the following advantages:

[0124] 1. Describe the reference line smoothing problem as an optimal control problem, derive the system model based on the fifth-order polynomial, use the state transition equation to maintain the second-order smoothing of the system, and have the advantage of high-order smoothing within the reference line segment. The output reference line has a smooth curvature and a low peak value, meeting the requirements of the planning module and the control module.

[0125] 2. The mathematical form of the solution model is in the form of an optimal control problem, which is convenient to call an efficient mathematical solver for the optimal control problem to solve.

[0126] 3. The algorithm has a low time complexity and a low space complexity, reducing resource occupancy while improving efficiency. Here, the low space complexity means that the final established optimization problem form is the form of Equation (7), which is the MPC mathematical form. An optimal control solver can be called to solve it, instead of calling a quadratic programming solver. The existing algorithm for smoothing the reference line of the piecewise spline curve is in the form of quadratic programming, and the Hessian matrix in it is o(N^2), while in the present application it is o(N). Since the problem form of the present application can be solved by calling an optimal control solver, the number of solution iteration steps generally only needs to be more than a dozen times, which is much faster than the existing algorithms. The solution peak of the existing algorithm reaches more than 100 ms, while the present application only needs less than 1 ms.

[0127] Based on the above method embodiments, the embodiments of the present application further provide a device for determining the reference line in autonomous driving. See Figure 4As shown, the device includes: an information acquisition module 42 for acquiring lane sequence information corresponding to the current position of the vehicle; the lane sequence information includes: the positions, orientations, and boundary information respectively corresponding to multiple anchor points; a solution module 44 for inputting the lane sequence information into a preset reference line smoothing problem solution model for solution to obtain a state sequence and a control output sequence; wherein, the reference line smoothing problem solution model is an optimal control solution model under preset constraint conditions; the objective function corresponding to the model is: the integral of the curvature and the derivative of the curvature of the entire curve segment corresponding to the reference line is minimized; the entire curve segment is divided into multiple spline curve segments by multiple anchor points; the system state of each spline curve segment is characterized by a fifth-degree polynomial; the preset constraint conditions include: the starting position and the ending position of the entire curve segment are respectively consistent with the positions of the corresponding anchor points, the state relationship between the spline curve segments satisfies the system state transition equation derived from the fifth-degree polynomial state equation, and the anchor point boundary constraint conditions; a reference line determination module 46 for determining the reference line corresponding to the lane sequence information based on the state sequence and the control output sequence.

[0128] The above information acquisition module is further configured to: determine map information within a specified range according to the current position of the vehicle; perform equidistant sampling on the lane lines corresponding to the map information to obtain multiple anchor points; acquire the positions, orientations, and boundary information respectively corresponding to each anchor point; the boundary information includes front and rear boundaries and left and right boundaries.

[0129] In the case where the above spline curve segments need to be spliced, the preset constraint conditions further include: the orientation of the ending position of the entire curve segment is consistent with the orientation of the corresponding anchor point; the orientation, curvature, and derivative of the curvature corresponding to the starting position of the entire curve segment are consistent with the initial state of the corresponding anchor point.

[0130] The above system state transition equation is as follows:

[0131]

[0132] Wherein,

[0133]

[0134]

[0135] x i (s) = a i0 + a i1 s + a i2 s 2 + a i3 s 3 + a i4 s 4 + a i5 s 5 ;

[0136] y i (s) = b i0 +b i1 s + b i2 s 2 +b i3 s 3 +b i4 s 4 +b i5 s 5 ;

[0137]

[0138]

[0139] Wherein, and respectively represent the system states corresponding to the (i + 1)-th and i-th spline curve segments; respectively represent the abscissa system state and ordinate system state corresponding to the i-th spline curve segment; represents the control output corresponding to the i-th spline curve segment; respectively represent the abscissa control output and ordinate control output corresponding to the i-th spline curve segment; s is the arc length of each spline curve segment; x i (s), y i (s) respectively represent the abscissa fifth-order polynomial state equation and ordinate fifth-order polynomial state equation corresponding to the i-th spline curve segment; a i0 , a i1 , a i2 , a i3 , a i4 , a i5 , b i0 , b i1 , b i2 , b i3 , b i4 , b i5 are all coefficients of the spline curve segment.

[0140] Furthermore, the above objective function is as follows:

[0141]

[0142] Wherein, respectively represent the second derivatives of the abscissa fifth-order polynomial state equation and ordinate fifth-order polynomial state equation corresponding to the i-th spline curve; respectively represent the third derivatives of the abscissa fifth-order polynomial state equation and ordinate fifth-order polynomial state equation corresponding to the i-th spline curve; ω1 and ω2 are the weights corresponding to the second derivative term and third derivative term respectively; i = 0, 1,..., N, and N is a positive integer.

[0143] Further, the above-mentioned anchor boundary constraint indicates that the position at the corresponding arc length is within a rectangular area consistent with the orientation of the anchor point; the rectangular area is determined by the boundary information of the anchor point; the formula corresponding to the anchor boundary constraint is as follows:

[0144]

[0145]

[0146] Among them, Δ l , Δ r , Δ b , Δ f respectively represent the left boundary, right boundary, front boundary and rear boundary corresponding to the anchor point; represents the position of the j-th anchor point, represents the orientation of the j-th anchor point, j = 1, 2,..., M, where M is the total number of anchor points; x i (s j ), y i (s j ) respectively represent the abscissa and ordinate of the reference point corresponding to the j-th anchor point.

[0147] The above-mentioned reference line determination module is further configured to: substitute the state sequence and the control output sequence into the relevant formula of the system state transition equation to obtain the spline coefficients corresponding to each spline curve segment; perform equal-arc length interpolation according to the spline coefficients corresponding to each spline curve segment and the quintic polynomial state equation to solve for the position, orientation, curvature and derivative of the curvature on the reference line, so as to determine the reference line corresponding to the lane sequence information.

[0148] The above-mentioned device further includes a post-processing module for: performing post-processing operations on the result of the reference line; the post-processing operations at least include one of the following: sampling, duplicate removal, feasibility check, constructing reference line-related objects.

[0149] The device provided by the embodiments of the present application has the same implementation principle and the same technical effects as those of the foregoing method embodiments. For the sake of brief description, for the parts not mentioned in the device embodiments, reference may be made to the corresponding contents in the foregoing method embodiments.

[0150] The embodiments of the present application also provide an electronic device, as Figure 5 shown, which is a schematic structural diagram of the electronic device. Among them, the electronic device includes a processor 51 and a memory 50. The memory 50 stores computer-executable instructions that can be executed by the processor 51, and the processor 51 executes the computer-executable instructions to implement the above method.

[0151] In Figure 5In the illustrated embodiment, the electronic device further includes a bus 52 and a communication interface 53, wherein the processor 51, the communication interface 53, and the memory 50 are connected through the bus 52.

[0152] Among them, the memory 50 may include a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk memory. The communication connection between the system network element and at least one other network element is realized through at least one communication interface 53 (which can be wired or wireless), and the Internet, wide area network, local area network, metropolitan area network, etc. can be used. The bus 52 may be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus, etc. The bus 52 can be divided into an address bus, a data bus, a control bus, etc. For the sake of simplicity of representation, Figure 5 only a single bidirectional arrow is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0153] The processor 51 may be an integrated circuit chip with the ability to process signals. In the implementation process, each step of the above method can be completed by the integrated logic circuit of the hardware in the processor 51 or the instructions in the form of software. The above-mentioned processor 51 may be a general-purpose processor, including a central processing unit (CPU for short), a network processor (NP for short), etc.; it may also be a digital signal processor (DSP for short), an application specific integrated circuit (ASIC for short), a field-programmable gate array (FPGA for short), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the method disclosed in the embodiments of the present application can be directly embodied as being executed and completed by the hardware decoding processor, or executed and completed by a combination of the hardware and software modules in the decoding processor. The software module may be located in a mature storage medium in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. This storage medium is located in the memory, and the processor 51 reads the information in the memory and combines its hardware to complete the steps of the method in the foregoing embodiments.

[0154] The embodiments of the present application also provide a computer-readable storage medium, which stores computer-executable instructions. When the computer-executable instructions are called and executed by the processor, the computer-executable instructions cause the processor to implement the above method. For the specific implementation, reference can be made to the foregoing method embodiments, and details will not be repeated here.

[0155] The computer program product of the method, device, and electronic device provided by the embodiments of the present application includes a computer-readable storage medium storing program codes. The instructions included in the program codes can be used to execute the method described in the foregoing method embodiments. For the specific implementation, reference can be made to the method embodiments, and details will not be repeated here.

[0156] Unless otherwise specifically stated, the relative steps, numerical expressions, and values of the components and steps set forth in these embodiments do not limit the scope of the present application.

[0157] When the above-mentioned functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium executable by a processor. Based on such understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present application. The foregoing storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs.

[0158] In the description of the present application, it should be noted that the orientation or positional relationship indicated by the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the present application and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation to the present application. In addition, the terms "first", "second", "third" are only used for descriptive purposes and cannot be construed as indicating or implying relative importance.

[0159] Finally, it should be noted that the above-mentioned embodiments are only specific embodiments of the present application, used to illustrate the technical solutions of the present application, rather than limiting them. The protection scope of the present application is not limited thereto. Although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: any person skilled in the technical field of the present application can still modify the technical solutions described in the foregoing embodiments, or can easily think of changes, or perform equivalent replacements for some of the technical features; and these modifications, changes, or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be covered by the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for determining a reference line in autonomous driving, characterized in that, The method includes: Obtaining lane sequence information corresponding to the current position of the vehicle; the lane sequence information includes: the positions, orientations, and boundary information corresponding to multiple anchor points; Inputting the lane sequence information into a preset reference line smoothing problem-solving model for solution to obtain a state sequence and a control output sequence; Among them, the reference line smoothing problem-solving model is an optimal control solution model under preset constraint conditions; the objective function corresponding to the model is: the integral of the curvature and the derivative of the curvature of the entire curve segment corresponding to the reference line is minimized; the entire curve segment is divided into multiple spline curve segments by the multiple anchor points; the system state of each spline curve segment is characterized by a fifth-degree polynomial; the preset constraint conditions include: the starting position and the ending position of the entire curve segment are respectively consistent with the positions of the corresponding anchor points, the state relationship between the spline curve segments satisfies the system state transition equation derived from the fifth-degree polynomial state equation, and the anchor point boundary constraint conditions; the objective function is as follows: Among them, respectively represent the second derivatives of the fifth-degree polynomial state equation of the abscissa and the fifth-degree polynomial state equation of the ordinate corresponding to the i-th spline curve; respectively represent the third derivatives of the fifth-degree polynomial state equation of the abscissa and the fifth-degree polynomial state equation of the ordinate corresponding to the i-th spline curve; ω1 and ω2 are the weights corresponding to the second derivative term and the third derivative term respectively; i = 0, 1,..., N, and N is a positive integer; The anchor point boundary constraint condition indicates that the position at the corresponding arc length is within a rectangular area consistent with the orientation of the anchor point; the rectangular area is determined by the boundary information of the anchor point; the formula corresponding to the anchor point boundary constraint condition is as follows: Among them, Δ l , Δ r , Δ b , Δ f respectively represent the left boundary, right boundary, front boundary and rear boundary corresponding to the anchor point; represents the position of the j-th anchor point, represents the orientation of the j-th anchor point, j = 1, 2, ..., M, where M is the total number of anchor points; x i (x j ), y i (x j ) respectively represent the abscissa and ordinate of the reference point corresponding to the j-th anchor point; The system state transition equation is as follows: Among them, x i (s) = a i0 + a i1 s + a i2 s 2 + a i3 s 3 + a i4 s 4 + a i5 s 5 ; y i (s) = b i0 + b i1 s + b i2 s 2 + b i3 s 3 + b i4 s 4 + b i5 s 5 ; Wherein, and respectively represent the system state sequences corresponding to the (i + 1)-th and the i-th spline curve segments; respectively represent the abscissa system state sequence and the ordinate system state sequence corresponding to the i-th spline curve segment; represents the control output sequence corresponding to the i-th spline curve segment; respectively represent the abscissa control output sequence and the ordinate control output sequence corresponding to the i-th spline curve segment; s is the arc length of each spline curve segment; x i (s), y i (s) respectively represent the abscissa quintic polynomial state equation and the ordinate quintic polynomial state equation corresponding to the i-th spline curve segment; a i0 , a i1 , a i2 , a i3 , a i4 , a i5 , b i0 , b i1 , b i2 , b i3 , b i4 , b i5 are all coefficients of the spline curve segment; x i ,, y i respectively represent the abscissa and the ordinate of the starting point corresponding to the i-th spline curve segment; respectively represent the first derivative, the second derivative, the third derivative, the fourth derivative and the fifth derivative of x i ; respectively represent the first derivative, the second derivative, the third derivative, the fourth derivative and the fifth derivative of y i ; Substituting the state sequence and the control output sequence into the formulas involved in the system state transition equation to obtain the spline coefficients corresponding to each spline curve segment; performing equal-arc-length interpolation according to the spline coefficients corresponding to each spline curve segment and the fifth-degree polynomial state equation to solve for the position, orientation, curvature, and derivative of the curvature on the reference line to determine the reference line corresponding to the lane sequence information.

2. The method according to claim 1, characterized in that, The step of obtaining lane sequence information corresponding to the current position of the vehicle includes: Determining map information within a specified range according to the current position of the vehicle; Performing equidistant sampling on the lane lines corresponding to the map information to obtain multiple anchor points; Obtaining the position, orientation, and boundary information corresponding to each anchor point; the boundary information includes front and rear boundaries and left and right boundaries.

3. The method according to claim 1, characterized in that, In the case where the reference lines need to be spliced, the preset constraint conditions further include: The orientation of the ending position of the entire curve segment is consistent with the orientation of the corresponding anchor point; The orientation, curvature, and derivative of the curvature corresponding to the starting position of the entire curve segment are consistent with the initial state of the corresponding anchor point.

4. The method according to claim 1, characterized in that, After the step of determining the reference line corresponding to the lane sequence information, the method further includes: Performing post-processing operations on the result of the reference line; the post-processing operations at least include one of the following: sampling, duplicate removal, feasibility check, constructing reference line-related objects.

5. A device for determining a reference line in autonomous driving, characterized in that, The device includes: An information acquisition module for acquiring lane sequence information corresponding to the current position of the vehicle; the lane sequence information includes: the positions, orientations, and boundary information corresponding to multiple anchor points; A solution module, configured to input the lane sequence information into a preset reference line smoothing problem solution model for solution, to obtain a state sequence and a control output sequence; wherein, the reference line smoothing problem solution model is an optimal control solution model under preset constraint conditions; the objective function corresponding to the model is: the integral of the curvature and the derivative of the curvature of the entire curve segment corresponding to the reference line is minimized; the entire curve segment is divided into a plurality of spline curve segments by the plurality of anchor points; the system state of each spline curve segment is characterized by a fifth-degree polynomial; the preset constraint conditions include: the starting position and the ending position of the entire curve segment are respectively consistent with the positions of the corresponding anchor points, the state relationship between the spline curve segments satisfies the system state transition equation derived from the fifth-degree polynomial state equation, and the anchor point boundary constraint conditions; the objective function is as follows: Among them, respectively represent the second derivatives of the fifth-order polynomial state equation of the abscissa and the fifth-order polynomial state equation of the ordinate corresponding to the i-th spline curve; respectively represent the third derivatives of the fifth-order polynomial state equation of the abscissa and the fifth-order polynomial state equation of the ordinate corresponding to the i-th spline curve; ω1 and ω2 are the weights corresponding to the second derivative term and the third derivative term respectively; i = 0, 1,..., N, and N is a positive integer; The anchor point boundary constraint condition indicates that the position at the corresponding arc length is within a rectangular area consistent with the orientation of the anchor point; the rectangular area is determined by the boundary information of the anchor point; the formula corresponding to the anchor point boundary constraint condition is as follows: Among them, Δ l , Δ r , Δ b , Δ f respectively represent the left boundary, right boundary, front boundary, and rear boundary corresponding to the anchor point; represents the position of the j-th anchor point, represents the orientation of the j-th anchor point, j = 1, 2,..., M, where M is the total number of anchor points; x i (s j ) and y i (s j ) respectively represent the abscissa and ordinate of the reference point corresponding to the j-th anchor point; The system state transition equation is as follows: Among them, x i (s) = a i0 +a i1 s + a i2 s 2 +a i3 s 3 +a i4 s 4 +a i5 s 5 ; y i (s) = b i0 +b i1 s + b i2 s 2 +b i3 s 3 +b i4 s 4 +b i5 s 5 ; wherein, and respectively represent the system state sequences corresponding to the (i + 1)-th and i-th spline curve segments; respectively represent the abscissa system state sequence and ordinate system state sequence corresponding to the i-th spline curve segment; represents the control output sequence corresponding to the i-th spline curve segment; respectively represent the abscissa control output sequence and ordinate control output sequence corresponding to the i-th spline curve segment; s is the arc length of each spline curve segment; x i (s), y i (s) respectively represent the abscissa fifth-order polynomial state equation and ordinate fifth-order polynomial state equation corresponding to the i-th spline curve segment; a i0 , a i1 , a i2 , a i3 , a i4 , a i5 , b i0 , b i1 , b i2 , b i3 , b i4 , b i5 are all coefficients of the spline curve segment; x i ,, y i , respectively represent the abscissa and ordinate of the starting point corresponding to the i-th spline curve segment; respectively represent the first derivative, second derivative, third derivative, fourth derivative, and fifth derivative of x i ,; respectively represent the first derivative, second derivative, third derivative, fourth derivative, and fifth derivative of y i ,. A reference line determination module, configured to substitute the state sequence and the control output sequence into the formulas involved in the system state transition equation, to obtain the spline coefficients corresponding to each spline curve segment; perform equal-arc-length interpolation according to the spline coefficients corresponding to each spline curve segment and the fifth-degree polynomial state equation, and solve for the position, orientation, curvature, and derivative of the curvature on the reference line, so as to determine the reference line corresponding to the lane sequence information.

6. An electronic device, characterized in that, It includes a processor and a memory, the memory stores computer-executable instructions that can be executed by the processor, and the processor executes the computer-executable instructions to implement the method according to any one of claims 1 to 4.

Images