CN111177934B - Method, apparatus and storage medium for reference path planning - Google Patents
Method, apparatus and storage medium for reference path planning Download PDFInfo
- Publication number
- CN111177934B CN111177934B CN201911425379.XA CN201911425379A CN111177934B CN 111177934 B CN111177934 B CN 111177934B CN 201911425379 A CN201911425379 A CN 201911425379A CN 111177934 B CN111177934 B CN 111177934B
- Authority
- CN
- China
- Prior art keywords
- path
- reference path
- curvature
- value
- course angle
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 72
- 238000009499 grossing Methods 0.000 claims abstract description 14
- 238000012545 processing Methods 0.000 claims abstract description 12
- 238000004590 computer program Methods 0.000 claims description 16
- 230000006870 function Effects 0.000 claims description 16
- 230000008859 change Effects 0.000 claims description 6
- 238000000354 decomposition reaction Methods 0.000 claims description 5
- 230000007613 environmental effect Effects 0.000 claims description 2
- 238000004364 calculation method Methods 0.000 abstract description 35
- 238000010586 diagram Methods 0.000 description 17
- 206010048669 Terminal state Diseases 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 6
- 230000008569 process Effects 0.000 description 6
- 239000011159 matrix material Substances 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000007423 decrease Effects 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- 241000282414 Homo sapiens Species 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012806 monitoring device Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000000630 rising effect Effects 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
Images
Classifications
-
- G06Q50/40—
Abstract
The invention discloses a method, equipment and a storage medium for planning reference paths, wherein the first reference paths are classified according to the path characteristics of the first reference paths to obtain the path types of the first reference paths, the endpoint state space is obtained according to the path characteristics of the first reference paths, and the first reference paths are smoothed through an interpolation model according to the endpoint state space to obtain second reference paths, wherein the interpolation model corresponds to the path types, so that the problems that a large amount of numerical calculation is needed for smoothing processing of the reference paths to obtain the optimal solution, the calculation resources are consumed more, the solution efficiency is low are solved, the consumption of the calculation resources is reduced, and the calculation efficiency is improved.
Description
Technical Field
The present application relates to the field of automated driving technologies, and in particular, to a method, an apparatus, and a storage medium for reference path planning.
Background
With the development of science and technology, autodrive automobiles have emerged. The automatic driving automobile depends on artificial intelligence technology, and the automatic driving software, the vision and radar sensor, the global positioning system, the communication network, the monitoring device and the like cooperate with each other to ensure that the vehicle provided with the automatic driving system can automatically and safely serve human beings without the operation of a driver. With the wide application of the automatic driving technology, in the related technology, the automatic driving automobile plans a reasonable reference path according to the map information. And performing online smoothing treatment on the reference path in a decision planning module, and planning a driving path according with vehicle dynamics by combining information such as local traffic environment, obstacles and the like. In the related art, a large amount of numerical calculation is required to obtain the optimal solution for the smoothing processing of the reference path, so that the calculation resources are consumed, and the solution efficiency is low.
Aiming at the problems that in the related art, a large amount of numerical optimization iteration is needed for smoothing processing of a reference path, vehicle-mounted computing resources are consumed, and the solving efficiency is low, an effective solution is not provided at present.
Disclosure of Invention
The invention provides a reference path planning method, equipment and a storage medium, aiming at solving the problems that in the related planning technology, a large amount of numerical calculation is needed for smoothing processing of a reference path to obtain an optimal solution, calculation resources are consumed, and the solution efficiency is low.
According to an aspect of the present invention, there is provided a method of reference path planning, the method comprising:
classifying the first reference path according to the path characteristics of the first reference path to obtain the path type of the first reference path;
and according to the endpoint state space, smoothing the first reference path through an interpolation model to obtain a second reference path, wherein the interpolation model corresponds to the path type.
In one embodiment, the smoothing the first reference path by an interpolation model according to the endpoint state space to obtain a second reference path includes one of:
under the condition that the first reference path is a straight line path, solving the first reference path through a straight line path interpolation model;
under the condition that the first reference path is a deflection path, solving the first reference path through a deflection path interpolation model;
and under the condition that the first reference path is a special path, solving the first reference path through a special path interpolation model.
In one embodiment, after the obtaining the type of the first reference path, the method further includes:
under the condition that the first reference path is a special path and the complexity of the special path is greater than a decomposition threshold value, decomposing the special path into a combination of a plurality of types of paths to obtain a plurality of endpoint state spaces;
and interpolating the special path according to the endpoint state space to obtain a second reference path.
In one embodiment, after the obtaining the second reference path, the method further includes:
according to the path characteristics of the second reference path, evaluating the second reference path through a target loss function to obtain a path error;
under the condition that the path error is not within a preset error range, selecting the parameter value of the path characteristic again within a preset deviation threshold value of the path characteristic to obtain an offset endpoint state space;
and determining a third reference path through the target loss function according to the offset endpoint state space.
In one embodiment, the path features include a path arc length, a heading angle, a curvature value, and a curvature peak-valley, and the classifying the first reference path according to the path features of the first reference path to obtain the path type of the first reference path includes one of:
under the condition that the absolute value of the course angle increment is smaller than a first preset course angle and the absolute value of the curvature value is smaller than a preset curvature value, judging that the first reference path is a straight path;
under the condition that the absolute value of the course angle increment is greater than or equal to the first preset course angle or the absolute value of the curvature value is greater than or equal to the preset curvature value, and under the condition that the number of the curvature peaks and valleys is 1 or 2, the first reference path is judged to be a deflection path;
and under the condition that the absolute value of the course angle increment is greater than or equal to the first preset course angle or the absolute value of the curvature value is greater than or equal to the preset curvature value, and under the condition that the number of the curvature peaks and valleys is greater than 2, judging that the first reference path is a special path.
In one embodiment, after said determining that the first reference path is a deflected path, the method further comprises one of:
under the condition that the numerical value of the curvature peak valley is 1 and the absolute value of the course angle increment is greater than a second preset course angle, judging that the first reference path is a U-turn path;
under the condition that the numerical value of the curvature peak valley is 1, the absolute value of the course angle increment is smaller than or equal to the second preset course angle, and the curvature value is a positive value, the first reference path is judged to be a left-turning path;
when the value of the curvature peak valley is 1, the absolute value of the course angle increment is smaller than or equal to the second preset course angle, and the curvature value is a negative value, the first reference path is judged to be a right turning path;
under the condition that the numerical value of the curvature peak valley is 2, only a peak or only a valley exists in the curvature peak valley, and the absolute value of the course angle increment is greater than a third preset course angle, judging that the first reference path is a U-turn path;
under the condition that the numerical value of the curvature peak valley is 2, only a peak or only a valley exists in the curvature peak valley, and the absolute value of the course angle increment is smaller than or equal to the third preset course angle, judging that the first reference path is a left-turning path or a right-turning path according to the positive and negative of the curvature value;
when the curvature peak valley has a numerical value of 2, a peak and a valley exist in the curvature peak valley at the same time, and under the condition that the path arc length is greater than the preset arc length, the first reference path is judged to be an S-shaped path;
and when the curvature peak valley has a numerical value of 2, a peak and a valley exist in the curvature peak valley, and the first reference path is judged to be the lane change path under the condition that the path arc length is less than or equal to the preset arc length.
In one embodiment, before the path feature according to the first reference path, the method includes:
and acquiring the first reference path, and downloading the first reference path and then performing offline processing.
According to another aspect of the present invention, there is provided an apparatus for reference path planning, the apparatus comprising a classification module and an interpolation module:
the classification module is used for classifying the first reference path according to the path characteristics of the first reference path to obtain the path type of the first reference path;
the interpolation module is configured to obtain an endpoint state space according to the path feature of the first reference path, and smooth the first reference path through an interpolation model according to the endpoint state space to obtain a second reference path, where the interpolation model corresponds to the path type.
According to another aspect of the invention, there is provided a computer apparatus comprising acquisition means, a memory and a processor:
the memory stores a computer program;
the acquisition device is used for acquiring road information and environmental information;
the processor, when executing the computer program, implements any of the methods described above.
According to another aspect of the invention, there is provided a computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements any of the methods described above.
According to the method and the device, the first reference paths are classified according to the path characteristics of the first reference paths to obtain the path types of the first reference paths, the endpoint state space is obtained according to the path characteristics of the first reference paths, the first reference paths are smoothed through the interpolation model according to the endpoint state space to obtain the second reference paths, wherein the interpolation model corresponds to the path types, the problems that a large amount of numerical calculation needs to be carried out on smoothing processing of the reference paths to obtain the optimal solution, calculation resources are consumed relatively, the solving efficiency is low are solved, the consumption of the calculation resources is reduced, and the calculation efficiency is improved.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the invention and together with the description serve to explain the invention without limiting the invention to a proper form.
In the drawings:
fig. 1 is a schematic application environment diagram of a method for reference path planning according to an embodiment of the present invention;
FIG. 2 is a first flowchart of a method of reference path planning according to an embodiment of the invention;
FIG. 3 is a schematic diagram of a first reference path according to an embodiment of the present invention;
FIG. 4 is a schematic illustration of a turning path curvature according to an embodiment of the invention;
FIG. 5 is a second flowchart of a method of reference path planning according to an embodiment of the invention;
FIG. 6 is a flow chart diagram III of a method of reference path planning in accordance with an embodiment of the present invention;
FIG. 7 is a first diagram illustrating a method of reference path classification according to an embodiment of the present invention;
FIG. 8 is a schematic diagram of the variation of the heading angle and curvature, respectively, with path type according to an embodiment of the invention;
FIG. 9 is a second diagram illustrating a method of reference path classification according to an embodiment of the invention;
fig. 10 is a block diagram of a structure of a reference path planning apparatus according to an embodiment of the present invention;
FIG. 11 is a fourth flowchart of a method of reference path planning according to an embodiment of the present invention;
fig. 12 is a schematic diagram of the internal structure of a computer device according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
It should be noted that the terms "first", "second" and "third" related to the embodiments of the present invention only distinguish similar objects, and do not represent specific ordering for the objects, and the terms "first", "second" and "third" may be interchanged with specific order or sequence, where permitted. It is to be understood that the terms "first," "second," and "third" are used interchangeably where appropriate to enable embodiments of the present invention described herein to be practiced in sequences other than those illustrated or described herein. The terms "first," "second," and "third" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance.
The method for reference path planning provided by the present application can be applied to the application environment shown in fig. 1, where fig. 1 is an application environment schematic diagram of the method for reference path planning according to an embodiment of the present invention, where a vehicle-mounted computer terminal 102 communicates with a server 104 through a network. The vehicle-mounted computer terminal 102 acquires the path characteristics of the first reference path through the server 104, classifies the first reference path according to the path characteristics to obtain the path type of the first reference path, and performs interpolation calculation on each path type through an interpolation model corresponding to the path type to obtain a smoother second reference path. The server 104 may be implemented by an independent server or a server cluster composed of a plurality of servers, and the server 104 may acquire a first reference path matched with a High Definition Map (HD Map) through a Global Positioning System (GPS) or a BeiDou Navigation Satellite System (BDS). Optionally, the first reference path may also be a path from a GPS acquisition and preprocessing.
In an embodiment, a method for reference path planning is provided, and fig. 2 is a first flowchart of the method for reference path planning according to the embodiment of the present invention, as shown in fig. 2, the method includes the following steps:
step S202, obtaining a first reference path, and classifying the first reference path according to the path characteristics of the first reference path to obtain the path type of the first reference path. The first reference path is an original reference path obtained from a high-precision map, and may also be a path acquired through a GPS and subjected to preprocessing. The reference path refers to an input path for trajectory generation by the autonomous vehicle in the decision-making planning module, in a structured roadway, typically a portion or the entirety of a roadway or lane centerline, fig. 3 is a schematic illustration of a first reference path according to an embodiment of the present invention, as shown in fig. 3, the curve of the original reference path is not smooth enough, and is represented by the jump of curvature, and can not be used as the final reference path, the path characteristics comprise the path arc length, the position, the course angle, the curvature value, the curvature peak valley and the like of the reference path, the first reference path may be classified according to a comparison result of the path characteristics with a preset value, the path type includes a straight path, a deflected path, a special path, and the like, the deflection path comprises a turning path, a left turning path, a right turning path, an S-turn path, a lane changing path and the like, and the special path comprises a rotary island path and the like.
Step S204, an endpoint state space is obtained according to the path characteristics of the first reference path, and the first reference path is smoothed through an interpolation model according to the endpoint state space to obtain a second reference path, wherein the interpolation model corresponds to the path type. The endpoint state space is represented by [ x y θ κ ]]TWhere x and y are the positions of the first reference path in a Cartesian coordinate system, θ is the heading angle, and κ is the curvature, and the second reference path is a path with continuous curvature.
Through step S202 and step S204, after the first reference path is acquired, the first reference path is divided into a straight path, a deflection path, or a special path according to the path characteristics of the first reference path, and different interpolation models are used to interpolate the reference path for different path types. In the related art, the type of the first reference path is not distinguished, all types of paths are interpolated by adopting a polynomial convolution curve, a cubic spline or a quintic spline, and for simple paths such as straight paths, the interpolation calculation amount is large by the method. The method provided by the embodiment solves the problems that a large amount of numerical calculation is needed to obtain the optimal solution for the smoothing processing of the reference path, the calculation resources are consumed, and the solution efficiency is low, reduces the consumption of the calculation resources, and improves the calculation efficiency.
In one embodiment, smoothing the first reference path by an interpolation model according to the endpoint state space to obtain a second reference path includes one of:
and under the condition that the first reference path is a straight line path, solving the first reference path through a straight line path interpolation model. The linear path interpolation model can be realized by a least square method or a two-point linear formula.
And when the first reference path is a deflection path, solving the first reference path through a deflection path interpolation model, wherein a polynomial with curvature of arc length for the deflection path is expressed by a high-order polynomial or a combination of a plurality of first-order polynomials, and the high-order polynomial is a polynomial of third order or more than third order. The deflection path interpolation model can be realized by combining a clothoid curve and a circular arc, wherein the clothoid curve is a curve with continuously changing curvature and is usually arranged between a straight line and a circular curve as transition or between two circular curves with different radiuses. For example, where the yaw path is a turn, FIG. 4 is a schematic diagram of the curvature of a turn path according to an embodiment of the present invention, the relationship of curvature to path arc length being given in the form of a piecewise function. As shown in fig. 4, the abscissa is the arc length of the turning path, and the ordinate is the curvature value, which can be obtained from the following equation 1:
where s is the arc length of the turn path, s1Arc length, s, of curve value in rising state in turning path2For arc lengths, s, in which the curvature values remain constant in the turning path3Arc length, a, for decreasing curvature values in the curve path1、b1Is a constant number, a1Without unit, in the case of arc length units of "meter", b1The unit of (2) is 1/m.
The heading angle can be obtained from equation 2 as follows:
the position coordinate x can be obtained by the following equation 3:
the position coordinate y can be obtained by the following equation 4:
the endpoint state space comprises an initial value state space and a terminal state space, wherein the initial value state space is the state space at the starting point of the turning path, and the terminal state space is the state space at the ending point of the turning path. After the endpoint state space is obtained, the parameter matrix is solved using newton's method. The solving method comprises the steps of firstly providing an initial estimated value of a terminal state space, setting an initial step length, calculating the deviation between a target terminal state space and the terminal state space, and then calculating a Jacobian (Jacobian) matrix of the terminal state space to a parameter vector as the gradient approaching the optimal solution. Assuming that a solution close to the optimal parameter is obtained in the step k, in order to find a closer solution, linearization is carried out in the step k to obtain a step length closer to the optimal parameter, and after the parameter vector is updated, iteration is carried out in the step k +1 until a parameter vector close to the optimal parameter is obtained.
Wherein the state space X is [ X y θ κ ]]TThe target state space is Xdes＝[xf yf θf κf]TVector of parametersIs p ═ b1 b2 s1 s2]TThe Jacobian matrix is
And obtaining a second reference path meeting the preset condition by a Newton method.
In the case where the first reference path is a special path, the first reference path is solved by a special path interpolation model, wherein the special path interpolation model can be generated by selecting a polynomial clothoid or a quintic-spline, and the curvature of the polynomial clothoid is a polynomial of third order or more. In the case of quintic spline curves, the coefficient matrix can be solved directly from the endpoint state space.
In the embodiment, different interpolation models are respectively adopted for calculating the reference paths of different path types, so that the consumption of calculation resources is reduced, the calculation efficiency is improved, and the situation of over-interpolation of the quintic spline curve in the process of interpolating a straight path and a deflection path is avoided.
In an embodiment, fig. 5 is a second flowchart of a method for reference path planning according to an embodiment of the present invention, and as shown in fig. 5, the method may further include the following steps:
step S502, when the first reference path is a special path and the complexity of the special path is greater than a decomposition threshold, the special path is decomposed into a combination of multiple path types to obtain multiple endpoint state spaces, where the special path includes a path with a complex curvature value change such as a roundabout. When the special path can be directly solved by using the special path interpolation model, decomposition is not needed, and when the special path can not be directly interpolated by using the special path interpolation model, the interpolation model is selected to interpolate the reference path according to the decomposed result. For example, if the first reference path is a roundabout, the roundabout path is classified and assigned to a special path, and the special path cannot be directly interpolated using a quintic spline, the roundabout path is decomposed into a combination of turn paths, or into a combination of turn paths and turn around paths, and if the roundabout path is decomposed into a combination of turn paths, the endpoint state space corresponding to each turn path is solved.
Step S504, according to the endpoint state space, performing interpolation on the special path to obtain a second reference path. For example, under the condition that the roundabout path is decomposed into the combination of the turning paths, the endpoint state space corresponding to each turning path is obtained, each endpoint state space is interpolated through the interpolation model, and then the multiple sections of reference paths obtained through interpolation are combined into the complete second reference path.
Through the steps S502 and S504, the special path with the high complexity is decomposed and then interpolated, compared with the method of directly interpolating the complicated special path, the decomposed interpolation calculation reduces the calculation amount, improves the calculation speed, and the decomposed path is interpolated, so that the curvature characteristics of the road can be captured more favorably, and the generated second reference path is smoother.
In an embodiment, fig. 6 is a flowchart three of a method for referring to path planning according to an embodiment of the present invention, as shown in fig. 6, the method may further include the following steps:
step S602, according to the path characteristics of the second reference path, the second reference path is evaluated by using a target loss function, so as to obtain a path error. And evaluating the reference path generated each time through the target loss function, and simultaneously ensuring that the deviation of the endpoint state space of the second reference path and the first reference path is restrained in an acceptable neighborhood. The target loss function is composed of a distance deviation term and a curvature deviation term, the distance deviation term represents the deviation degree of the second reference path and the first reference path, the curvature deviation term represents the smoothness degree of the curvature of the path, and in the iteration process, the path which is small in lateral error and smooth enough is selected while the end point state constraint condition is met.
The target loss function can be obtained from the following equation 5:
wherein k is1、k2Is a weight parameter.
The constraint is given by the following equation 6:
wherein p isiIs the end point position of the second reference path,
is the end position of the first reference path, θiIs the end point heading angle of the second reference path,
is the end point position of the first reference path, kiBeing the end point curvature of the second reference path,
is the end point curvature of the first reference path, δpTo preset a position deviation threshold value, deltaθFor a predetermined course angle deviation threshold value, deltaκIs a preset curvature deviation threshold.
Step S604, when the path error is not within the preset error range, the parameter value of the path feature is selected again within the preset deviation threshold of the path feature, so as to obtain the state space of the offset endpoint. The preset deviation threshold is adjusted according to the actual situation, and a preset parameter range is given in formula 6.
Step S606, according to the offset endpoint state space, a third reference path is selected through the objective loss function. And solving the path again after the endpoint state space is shifted, and finding the path with the minimum cost function through iteration, namely the path is the third reference path.
Through the steps S602 to S606, after the second reference path is obtained, the second reference path is evaluated through the target loss function, and meanwhile, the endpoint state space is adjusted within the constraint condition that the endpoint state space is satisfied, and the reference path with the minimum target loss function value is iteratively selected as the optimal reference path to be output, so that the finally output reference path is smoother, and the safety of automatic driving is improved.
In one embodiment, FIG. 7 is a first schematic diagram of a method for reference path classification according to an embodiment of the present invention, in this embodiment classifying according to path features of a first reference path, the path features including path arc length, heading angle, curvature value and curvature peak-valley, wherein, the course angle is an included angle between the direction of the vehicle mass center speed and the coordinate horizontal axis in the geodetic coordinate system, the smaller the course angle is, the closer the reference path is to a straight line, the curvature value can change along with the change of the curvature degree of the reference path, the higher the curvature degree of the reference path is, the higher the curvature value is, the closer the reference path is to the straight line, the lower the curvature value is, in the process that the curvature increases and then decreases or decreases and then increases along the path, the peak-valley is an image description of the curvature changing along the path, and the step of classifying the first reference path comprises one of the following steps:
step S702, when the absolute value of the course angle increment is smaller than a first preset course angle and the absolute value of the curvature value is smaller than a preset curvature value, determining that the first reference path is a straight path. In fig. 7, "| Δ heading angle |" represents an absolute value of the heading angle increment, and "| curvature value |" represents an absolute value of the curvature value.
Step S704, when the absolute value of the course angle increment is greater than or equal to the first preset course angle or the absolute value of the curvature value is greater than or equal to the preset curvature value, and the number of curvature peaks and valleys is 1 or 2, determining that the first reference path is a deflected path. Wherein the number of curvature peaks and valleys is the sum of the number of peaks and valleys. For example, in the case where the values of the steering angle and the curvature satisfy the preset conditions and the number of peaks and valleys is 1, the first reference path is considered to be a left turn or a right turn, and is attributed to the deflected path.
Step S706, when the absolute value of the course angle increment is greater than or equal to the first preset course angle or the absolute value of the curvature value is greater than or equal to the preset curvature value, and the number of the curvature peaks and valleys is greater than 2, determining that the first reference path is a special path.
FIG. 8 is a schematic diagram showing the variation of the course angle and the curvature respectively according to the path type, as shown in FIG. 8, the course angle of the straight path has small variation when the arc length of the path increases, and basically fluctuates in a small range, there are only peaks or only valleys in the left-turn path or the right-turn path, and there are two peaks in the u-turn path.
Through the steps S702 to S706, the first reference path is divided into a straight path, a deflected path or a special path according to the course angle, the curvature value and the curvature peak and valley, and the path classification is more definite due to the path characteristics including the path arc length, the course angle, the curvature value and the curvature peak and valley, so that after the path classification, interpolation calculation is performed by adopting different interpolation models according to different path types, and the calculation efficiency is improved.
In one embodiment, fig. 9 is a second schematic diagram of a method for reference path classification according to an embodiment of the present invention, and as shown in fig. 9, after determining that the first reference path is a deflection path, the method further includes one of the following steps:
step S902, when the value of the curvature peak-valley is 1 and the absolute value of the course angle increment is greater than the second preset course angle, determining that the first reference path is the u-turn path.
Step S904, when the value of the curvature peak-valley is 1, the absolute value of the course angle increment is less than or equal to the second preset course angle, and the curvature value is a positive value, determines that the first reference path is a left-turn path. In the present embodiment, the direction in which the curve curves is determined according to the positive or negative of the curvature value, and the left-turn path is set to the curvature positive direction in the present embodiment.
Step S906, when the value of the curvature peak-valley is 1, the absolute value of the course angle increment is less than or equal to the second preset course angle, and the curvature value is a negative value, determines that the first reference path is a right-turn path.
Step S908, when the value of the curvature peak-valley is 2, only a peak or only a valley exists in the curvature peak-valley, and the absolute value of the course angle increment is greater than a third preset course angle, determining that the first reference path is a u-turn path.
Step S910, under the condition that the value of the curvature peak valley is 2, only the peak or only the valley exists in the curvature peak valley, and the absolute value of the course angle increment is less than or equal to a third preset course angle, judging that the first reference path is a left-turning path or a right-turning path again according to the positive and negative of the curvature value. When the curvature value is positive, the first reference path is determined to be a left-turn path, and when the curvature value is negative, the first reference path is determined to be a right-turn path.
Step S912, when the curvature peak-valley value is 2, a peak and a valley exist in the curvature peak-valley, and the arc length of the path is greater than the preset arc length, it is determined that the first reference path is an S-turn path.
Step S914, when the curvature peak-valley value is 2, a peak and a valley exist in the curvature peak-valley, and the path arc length is less than or equal to the preset arc length, the first reference path is determined to be the lane change path.
Through the steps S902 to S914, the deflection paths are further classified, which is beneficial to performing interpolation calculation by using different interpolation models for different path types after path classification, and further improves the calculation efficiency.
In one embodiment, after the first reference path is obtained, the first reference path is processed offline to obtain the second reference path, and compared with the online processing of path planning in the related art, the offline processing is beneficial to saving resources of a vehicle-mounted computer and improving the operation efficiency.
In one embodiment, after the global path is changed, the vehicle-mounted computer sends a cloud re-planning instruction and acquires the reference path from the cloud again, so that the vehicle-mounted computer can timely respond to the situation that the existing path cannot pass through.
In one embodiment, the second reference path is obtained by segmenting the first reference path, and the segmentation is obtained by classification. For example, the first reference path is 100 meters, after classification, the first reference path is divided into two 40-meter straight paths and a 20-meter left-turn path, solution is performed according to an endpoint state space between the two straight paths and the 20-meter left-turn path, three segmented paths are obtained, the segmented paths are spliced to obtain a second reference path, but the second reference path may not be smooth enough, so that the endpoint state space needs to be shifted and then solved again, and a path with the minimum cost function is found through iterative computation to serve as a third reference path.
It should be understood that, although the steps in the flowcharts of fig. 2 to 9 are shown in sequence as indicated by the arrows, the steps are not necessarily performed in sequence as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least some of the steps in fig. 2-9 may include multiple sub-steps or multiple stages that are not necessarily performed at the same time, but may be performed at different times, and the order of performing the sub-steps or stages is not necessarily sequential, but may be performed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.
Corresponding to the above method for reference path planning, in this embodiment, a device for reference path planning is further provided, and the device is used to implement the above embodiments and preferred embodiments, and is not described again after having been described. As used below, the term "module" may be a combination of software and/or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, implementations in hardware or a combination of software and hardware are also possible and contemplated.
In an embodiment, a device for reference path planning is provided, and fig. 10 is a block diagram of a structure of the device for reference path planning according to an embodiment of the present invention, as shown in fig. 10, including: a classification module 1002 and an interpolation module 1004, wherein:
the classifying module 1002 is configured to classify the first reference path according to a path feature of the first reference path, so as to obtain a path type of the first reference path.
An interpolation module 1004, configured to obtain an endpoint state space according to the path feature of the first reference path, and smooth the first reference path through an interpolation model according to the endpoint state space to obtain a second reference path, where the interpolation model corresponds to the path type.
Through the classification module 1002 and the interpolation module 1004, after the classification module 1002 acquires the first reference path, the first reference path is divided into a straight path, a deflection path or a special path according to the path characteristics of the first reference path, and for different path types, the interpolation module 1004 performs interpolation on the reference path by using different interpolation models. In the related art, the type of the first reference path is not distinguished, all types of paths are interpolated by adopting a polynomial convolution curve, a cubic spline or a quintic spline, and for simple paths such as straight paths, the interpolation calculation amount is large by the method. The method provided by the embodiment solves the problems that a large amount of numerical calculation is needed to obtain the optimal solution for the smoothing processing of the reference path, the calculation resources are consumed, and the solution efficiency is low, reduces the consumption of the calculation resources, and improves the calculation efficiency.
In the following, embodiments of the present invention are described in detail with reference to an actual application scenario, and in a process of performing reference path planning, fig. 11 is a fourth flowchart of a method for reference path planning according to an embodiment of the present invention, as shown in fig. 11, the method includes the following steps:
step S1102 is to obtain a first reference path, and classify the first reference path according to the path characteristics of the first reference path to obtain the path type of the first reference path.
Step S1104, when the first reference path is a straight path or a deflection path, obtaining an endpoint state space according to a path feature of the first reference path, or when the first reference path is a special path and a complexity of the special path is greater than a decomposition threshold, decomposing the special path into a combination of multiple path types to obtain multiple endpoint state spaces.
Step S1106, according to the endpoint state space, performing path solution, and obtaining a second reference path through an interpolation model.
Step S1108, according to the path characteristics of the second reference path, the second reference path is evaluated by using a target loss function, so as to obtain a path error.
In step S1110, under the condition that the path error is not within the preset error range, the parameter value of the path feature is selected again within the preset parameter range of the path feature, so as to obtain the offset endpoint state space.
In step S1112, a third reference path is obtained according to the offset endpoint state space.
Through the above steps S1102 to S1112, the first reference path is divided into a straight path, a deflection path, or a special path according to the path characteristics of the first reference path, and the interpolation module 1004 performs interpolation calculation of the reference path by using different interpolation models for different path types. In the related art, a polynomial clothoid curve, a cubic spline or a quintic spline is adopted for all types of paths for interpolation, and for simple paths such as straight paths, the method provided by the embodiment with a large interpolation calculation amount solves the problems that a large amount of numerical calculation is needed for smoothing processing of a reference path to obtain an optimal solution, vehicle-mounted computing resources are consumed, and the solving efficiency is low, reduces the consumption of the vehicle-mounted computing resources, improves the computing efficiency, and avoids the situation of over-interpolation when the reference path is a straight path.
In one embodiment, a computer device is provided, which may be a terminal. The computer device includes a processor, a memory, a network interface, a display screen, and an input device connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a method of reference path planning. The display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, a key, a track ball or a touch pad arranged on the shell of the computer equipment, an external keyboard, a touch pad or a mouse and the like.
In one embodiment, fig. 12 is a schematic diagram of an internal structure of a computer device according to an embodiment of the present invention, and as shown in fig. 12, a computer device is provided, where the computer device may be a server, and the internal structure diagram may be as shown in fig. 12. The computer device includes a processor, a memory, a network interface, and a database connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The database of the computer device is used for storing data. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a method of reference path planning.
Those skilled in the art will appreciate that the architecture shown in fig. 12 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.
In one embodiment, a computer device is provided, which includes a collecting device, a memory, a processor, and a computer program stored on the memory and executable on the processor, where the collecting device is configured to collect road information and environment information, and the processor implements the steps in the reference path planning obtaining method provided in each of the above embodiments when executing the computer program.
In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method for reference path planning provided by the above-mentioned embodiments.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory, among others. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above examples only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (10)
1. A method of reference path planning, the method comprising:
classifying the first reference path according to the path characteristics of the first reference path to obtain the path type of the first reference path; wherein the path features include a path arc length, a heading angle, a curvature value, and a curvature peak-to-valley of the first reference path; classifying the first reference path according to a comparison result of the absolute value of the course angle increment and a first preset course angle, a comparison result of the absolute value of the curvature value and a preset curvature value, or classifying the first reference path according to a comparison result of the absolute value of the course angle increment and the first preset course angle, a comparison result of the absolute value of the curvature value and the preset curvature value and the number of curvature peaks and valleys, wherein the path type comprises at least one of a straight path, a deflection path and a special path;
obtaining an endpoint state space according to the path characteristics of the first reference path, and smoothing the first reference path through an interpolation model according to the endpoint state space to obtain a second reference path, wherein the interpolation model corresponds to the path type, and the endpoint state space is represented by [ x y θ κ ]]TWhere x and y are the positions of the first reference path in the Cartesian coordinate system, θ is the heading angle, and κ is the curvature.
2. The method of claim 1, wherein the smoothing the first reference path by an interpolation model according to the endpoint state space to obtain a second reference path comprises one of:
under the condition that the first reference path is a straight line path, solving the first reference path through a straight line path interpolation model;
under the condition that the first reference path is a deflection path, solving the first reference path through a deflection path interpolation model;
and under the condition that the first reference path is a special path, solving the first reference path through a special path interpolation model.
3. The method of reference path planning according to claim 1, wherein after said obtaining the type of the first reference path, the method further comprises:
under the condition that the first reference path is a special path and the complexity of the special path is greater than a decomposition threshold value, decomposing the special path into a combination of a plurality of types of paths to obtain a plurality of endpoint state spaces;
and interpolating the special path according to the endpoint state space to obtain a second reference path.
4. The method of reference path planning according to claim 1, wherein after said obtaining a second reference path, the method further comprises:
according to the path characteristics of the second reference path, evaluating the second reference path through a target loss function to obtain a path error;
under the condition that the path error is not within a preset error range, selecting the parameter value of the path characteristic again within a preset deviation threshold value of the path characteristic to obtain an offset endpoint state space;
and determining a third reference path through the target loss function according to the offset endpoint state space.
5. The method of reference path planning as claimed in claim 1, wherein the path features include path arc length, course angle, curvature value and curvature peak-valley, and the classifying the first reference path according to the path features of the first reference path to obtain the path type of the first reference path includes one of:
under the condition that the absolute value of the course angle increment is smaller than a first preset course angle and the absolute value of the curvature value is smaller than a preset curvature value, judging that the first reference path is a straight path;
under the condition that the absolute value of the course angle increment is greater than or equal to the first preset course angle or the absolute value of the curvature value is greater than or equal to the preset curvature value, and under the condition that the number of the curvature peaks and valleys is 1 or 2, the first reference path is judged to be a deflection path;
and under the condition that the absolute value of the course angle increment is greater than or equal to the first preset course angle or the absolute value of the curvature value is greater than or equal to the preset curvature value, and under the condition that the number of the curvature peaks and valleys is greater than 2, judging that the first reference path is a special path.
6. The method of reference path planning according to claim 5, wherein after said determining that the first reference path is a deflected path, the method further comprises one of:
under the condition that the numerical value of the curvature peak valley is 1 and the absolute value of the course angle increment is greater than a second preset course angle, judging that the first reference path is a U-turn path;
under the condition that the numerical value of the curvature peak valley is 1, the absolute value of the course angle increment is smaller than or equal to the second preset course angle, and the curvature value is a positive value, the first reference path is judged to be a left-turning path;
when the value of the curvature peak valley is 1, the absolute value of the course angle increment is smaller than or equal to the second preset course angle, and the curvature value is a negative value, the first reference path is judged to be a right turning path;
under the condition that the numerical value of the curvature peak valley is 2, only a peak or only a valley exists in the curvature peak valley, and the absolute value of the course angle increment is greater than a third preset course angle, judging that the first reference path is a U-turn path;
under the condition that the numerical value of the curvature peak valley is 2, only a peak or only a valley exists in the curvature peak valley, and the absolute value of the course angle increment is smaller than or equal to the third preset course angle, judging that the first reference path is a left-turning path or a right-turning path according to the positive and negative of the curvature value;
when the curvature peak valley has a numerical value of 2, a peak and a valley exist in the curvature peak valley at the same time, and under the condition that the path arc length is greater than the preset arc length, the first reference path is judged to be an S-shaped path;
and when the curvature peak valley has a numerical value of 2, a peak and a valley exist in the curvature peak valley, and the first reference path is judged to be the lane change path under the condition that the path arc length is less than or equal to the preset arc length.
7. The method of reference path planning according to any one of claims 1 to 6, wherein prior to the path feature according to the first reference path, the method comprises:
and acquiring the first reference path, and downloading the first reference path and then performing offline processing.
8. An apparatus for reference path planning, the apparatus comprising a classification module and an interpolation module:
the classification module is used for classifying the first reference path according to the path characteristics of the first reference path to obtain the path type of the first reference path; the path characteristics comprise a path arc length, a course angle, a curvature value and a curvature peak valley of the first reference path, the first reference path is classified according to a comparison result of an absolute value of a course angle increment and a first preset course angle, a comparison result of the absolute value of the curvature value and a preset curvature value, or the first reference path is classified according to a comparison result of the absolute value of the course angle increment and the first preset course angle, a comparison result of the absolute value of the curvature value and a preset curvature value and the number of the curvature peak valleys, and the path type comprises at least one of a straight path, a deflection path and a special path;
the interpolation module is configured to obtain an endpoint state space according to the path feature of the first reference path, and smooth the first reference path through an interpolation model according to the endpoint state space to obtain a second reference path, where the interpolation model corresponds to the path type, and the endpoint state space is represented by [ xy θ κ ]]TWhere x and y are the positions of the first reference path in the Cartesian coordinate system, θ is the heading angle, and κ is the curvature.
9. A computer device, comprising an acquisition device, a memory, and a processor:
the memory stores a computer program;
the acquisition device is used for acquiring road information and environmental information;
the processor, when executing the computer program, realizes the steps of the method of any one of claims 1 to 7.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911425379.XA CN111177934B (en) | 2019-12-31 | 2019-12-31 | Method, apparatus and storage medium for reference path planning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911425379.XA CN111177934B (en) | 2019-12-31 | 2019-12-31 | Method, apparatus and storage medium for reference path planning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111177934A CN111177934A (en) | 2020-05-19 |
| CN111177934B true CN111177934B (en) | 2021-06-22 |
Family
ID=70647498
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911425379.XA Active CN111177934B (en) | 2019-12-31 | 2019-12-31 | Method, apparatus and storage medium for reference path planning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111177934B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111737636B (en) * | 2020-06-12 | 2024-02-02 | 北京百度网讯科技有限公司 | Path curve generation method, device, computer equipment and storage medium |
| CN111665845B (en) * | 2020-06-24 | 2023-09-22 | 阿波罗智能技术(北京)有限公司 | Method, apparatus, device and storage medium for planning path |
| CN112631306B (en) * | 2020-12-28 | 2021-12-14 | 深圳市普渡科技有限公司 | Robot moving path planning method and device and robot |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104007705A (en) * | 2014-05-05 | 2014-08-27 | 上海交通大学 | Prospective interpolation system for compressing and smoothening small segment paths |
| CN109540162A (en) * | 2018-11-12 | 2019-03-29 | 北京四维图新科技股份有限公司 | Processing method, acquisition methods, device and the mobile unit of ADAS map datum |
| CN109947101A (en) * | 2019-03-18 | 2019-06-28 | 北京智行者科技有限公司 | Path smooth processing method and processing device |
| CN109974724A (en) * | 2017-12-27 | 2019-07-05 | 湖南中车时代电动汽车股份有限公司 | A kind of paths planning method for intelligent driving system |
| CN110466527A (en) * | 2019-08-22 | 2019-11-19 | 广州小鹏汽车科技有限公司 | A kind of travel control method of vehicle, system and vehicle |
-
2019
- 2019-12-31 CN CN201911425379.XA patent/CN111177934B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104007705A (en) * | 2014-05-05 | 2014-08-27 | 上海交通大学 | Prospective interpolation system for compressing and smoothening small segment paths |
| CN109974724A (en) * | 2017-12-27 | 2019-07-05 | 湖南中车时代电动汽车股份有限公司 | A kind of paths planning method for intelligent driving system |
| CN109540162A (en) * | 2018-11-12 | 2019-03-29 | 北京四维图新科技股份有限公司 | Processing method, acquisition methods, device and the mobile unit of ADAS map datum |
| CN109947101A (en) * | 2019-03-18 | 2019-06-28 | 北京智行者科技有限公司 | Path smooth processing method and processing device |
| CN110466527A (en) * | 2019-08-22 | 2019-11-19 | 广州小鹏汽车科技有限公司 | A kind of travel control method of vehicle, system and vehicle |
Non-Patent Citations (1)
| Title |
|---|
| 《智能车决策路径规划的研究进展》;张鑫辰 等;《计算机科学》;20191031;第46卷(第10A期);第19-23页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111177934A (en) | 2020-05-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111177934B (en) | Method, apparatus and storage medium for reference path planning | |
| US20230244237A1 (en) | Optimal path library for local path planning of an autonomous vehicle | |
| EP3959576A1 (en) | Methods and systems for trajectory forecasting with recurrent neural networks using inertial behavioral rollout | |
| CN108519094B (en) | Local path planning method and cloud processing terminal | |
| CN110595475A (en) | Tracking path fitting method and device, intelligent automobile and storage medium | |
| CN111976718B (en) | Automatic parking control method and system | |
| CN111750861B (en) | Optimization path planning method and device and electronic equipment | |
| CN111121777A (en) | Unmanned equipment trajectory planning method and device, electronic equipment and storage medium | |
| CN113085895B (en) | Vehicle lane change track planning method, device, equipment, storage medium and vehicle | |
| CN110262506A (en) | Confirmation method, vehicle travel control method, device and equipment a little is taken aim in advance | |
| CN111126598A (en) | Automatic parking method, automatic parking device, computer equipment and storage medium | |
| CN114454888A (en) | Lane line prediction method and device, electronic equipment and vehicle | |
| CN117141520A (en) | Real-time track planning method, device and equipment | |
| CN114664094B (en) | Vehicle track prediction method, device, chip, terminal and computer equipment | |
| CN115218902A (en) | Trajectory planning method, device, equipment and storage medium | |
| CN116659539B (en) | Path planning method, path planning device and domain controller | |
| CN114114930B (en) | Method, device, equipment and medium for generating local reference path of automobile | |
| CN116659538B (en) | Road diameter changing planning method and device and vehicle-mounted equipment | |
| CN114490900B (en) | Multi-vehicle track data fusion method, device, equipment and storage medium | |
| EP4261738A1 (en) | Multi-target lane tracking | |
| CN115760909A (en) | Path generation method, path generation device, computer equipment, storage medium and program product | |
| CN115309145A (en) | Obstacle avoidance control method and device for vehicle, storage medium and terminal | |
| Zhao et al. | A Mean-variance Optimization DQN in Autonomous Driving | |
| CN114906152A (en) | Lane line construction method, lane line construction device, electronic device, and computer-readable storage medium | |
| CN117246359A (en) | Vehicle track prediction method, device, equipment, storage medium and product |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |