Vehicle control method and device based on path planning, equipment, medium and product

Abstract

The embodiment of the invention provides a vehicle control method and device based on path planning, equipment, a medium and a product. The method comprises the following steps: determining an initial lane path on a target lane according to lane line data of the target lane; according to the bending degree between target lane points on the initial lane path and the constraint condition of the vehicle driving on the target lane, the target lane path of the vehicle driving on the target lane is determined, and the target lane points are at least four lane points adjacently arranged along the initial lane path; and when the vehicle enters the target road, controlling the vehicle to run according to the target lane path. The method is used for improving the driving stability of the vehicle and improving the driving experience effect of a user.

Description

Technical Field

[0001] The present application relates to the field of autonomous driving technology, and in particular to a vehicle control method, device, equipment, medium and product based on path planning. Background Art

[0002] Path planning is one of the core technologies in the field of autonomous driving. It involves calculating a feasible path from the starting point to the end point while avoiding obstacles to ensure safety and efficiency. This technology relies on path planning algorithms and combines sensor data processing and environmental modeling to achieve autonomous navigation in static or dynamic environments.

[0003] Currently, path planning technology integrates advanced sensors (such as lidar, cameras and radar) with high-precision map data and combines complex algorithms to calculate the optimal driving path. The vehicle's control system uses this information to make decisions in real time to ensure that obstacles and traffic conditions can be avoided efficiently and safely, thereby achieving autonomous navigation from the starting point to the destination.

[0004] However, when guiding a vehicle to travel along a planned path, existing vehicle control methods sometimes fail to perform sufficient path smoothing, which affects the vehicle's driving stability and the user's driving experience. Summary of the invention

[0005] The embodiments of the present application provide a vehicle control method, device, equipment, medium and product based on path planning, which are used to improve the driving stability of the vehicle and enhance the driving experience of the user.

[0006] In a first aspect, an embodiment of the present application provides a vehicle control method based on path planning, comprising:

[0007] Determine an initial lane path on the target lane according to lane line data of the target lane;

[0008] Determine a target lane path for the vehicle to travel in the target lane according to the curvature between the target lane points on the initial lane path and the constraint conditions for the vehicle to travel in the target lane, wherein the target lane points are at least four lane points adjacently arranged along the initial lane path;

[0009] When the vehicle enters the target road, the vehicle is controlled to travel according to the target lane path.

[0010] In a possible implementation, determining an initial lane path on the target lane according to lane line data of the target lane includes:

[0011] Determine lane centerline information of each lane on the target lane according to lane line data of the target lane;

[0012] According to the lane centerline information of each lane on the target lane, the centerlines of the lanes are spliced ​​to obtain the initial lane path on the target lane.

[0013] In one possible implementation, the degree of curvature between the target lane points on the initial lane path is determined based on a curvature loss of a first lane point relative to a second lane point among the target lane points, where the second lane point is a target lane point that is adjacent to the first lane point in front and behind on the initial lane path.

[0014] In a possible implementation, the curvature loss is determined according to a positional relationship between the first lane point and the second lane point in the first direction and a positional relationship in the second direction.

[0015] In one possible implementation, the curvature loss satisfies:

[0016] cost i =(x i+1 -3x i +3x i-1 +x i-2 ) 2 +(y i+1 -3y i +3y i-1 +y i-2 ) 2 ;

[0017] Among them, cost i Characterizes the curvature loss of the first lane point i; x i represents the position of the first lane point i in the first direction; x i+1 、x i-1 、x i-2 Respectively represent the positions of the second lane point i+1, the second lane point i-1, and the second lane point i-2 in the first direction; y i Represents the position of the first lane point i in the second direction; y i+1 ,y i-1 ,y i-2 Respectively represent the positions of the second lane point i+1, the second lane point i-1, and the second lane point i-2 in the second direction.

[0018] In a possible implementation, determining a target lane path for the vehicle to travel in the target lane according to the curvature between the target lane points on the initial lane path and the constraint condition for the vehicle to travel in the target lane includes:

[0019] According to the curvature between the target lane points on the initial lane path, a target matrix is ​​obtained, wherein the rows of the target matrix represent the curvature of the target lane points in the first direction, the columns of the target matrix represent the curvature of the target lane points in the second direction, and the elements in the target matrix represent the curvature of the target lane points;

[0020] According to the constraint conditions of the vehicle driving on the target lane, a constraint matrix, a linear cost vector, an upper constraint vector and a lower constraint vector are determined;

[0021] A lane path of a target lane for the vehicle is determined according to the target matrix, the constraint matrix, the linear cost vector, the upper bound constraint vector and the lower bound constraint vector.

[0022] In a possible implementation, determining a lane path of a target lane for a vehicle to travel in according to a target matrix, a constraint matrix, a linear cost vector, an upper constraint vector, and a lower constraint vector includes:

[0023] Adjust the format of the target matrix to obtain an initial sparse matrix table;

[0024] Determine a weight coefficient corresponding to a non-zero element in the initial sparse matrix table, the weight coefficient comprising at least one of a curvature loss weight, a total path length loss weight, and an offset loss weight;

[0025] According to the weight coefficients corresponding to the non-zero elements in the initial compression matrix, the initial sparse matrix table is adjusted to obtain a sparse matrix table;

[0026] The sparse matrix table, the constraint matrix, the linear cost vector, the upper bound constraint vector and the lower bound constraint vector are input into the optimization solver to obtain the lane path of the target lane of the vehicle.

[0027] In a possible implementation manner, the lane line data of the target lane is data obtained from a reference map.

[0028] In a second aspect, an embodiment of the present application provides a vehicle control device based on path planning, comprising:

[0029] A determination module, for determining an initial lane path on the target lane according to lane line data of the target lane;

[0030] A planning module, for determining a target lane path for the vehicle to travel in the target lane according to the curvature between the target lane points on the initial lane path and the constraint conditions for the vehicle to travel in the target lane, wherein the target lane points are at least four lane points adjacently arranged along the initial lane path;

[0031] The control module is used to control the vehicle to travel according to the target lane path when the vehicle enters the target road.

[0032] In a third aspect, an embodiment of the present application provides a vehicle control device, including: a memory, a processor;

[0033] Memory stores computer-executable instructions;

[0034] The processor executes the computer-executable instructions stored in the memory, so that the processor executes the above first aspect and / or various possible implementations of the first aspect.

[0035] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, in which computer-executable instructions are stored. When the computer-executable instructions are executed by a processor, they are used to implement the first aspect above and / or various possible implementations of the first aspect.

[0036] In a fifth aspect, an embodiment of the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the above first aspect and / or various possible implementation methods of the first aspect.

[0037] The vehicle control method, device, equipment, medium and product based on path planning provided in the embodiments of the present application, after determining the initial path of the target lane, can more accurately predict and adjust the curvature change of the path by combining and analyzing the curvature of multiple groups of target lane points on the initial path and the constraints of the vehicle driving such as speed, acceleration and steering angle limits, thereby ensuring that the vehicle follows a gradual and smooth trajectory when turning, rather than suddenly changing direction to optimize the path. This allows the vehicle to travel stably according to the trajectory of the target lane path when entering the target road, thereby reducing unnecessary shaking and sharp turns, thereby improving the vehicle driving stability and enhancing the user's driving experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0039] Figure 1 A schematic diagram of the principle of an existing method for determining the degree of bending provided in an embodiment of the present application;

[0040] Figure 2 A schematic diagram of a scenario of a vehicle control method based on path planning provided in this application;

[0041] Figure 3 Schematic diagram of the process of the vehicle control method based on path planning provided in this application Figure 1 ;

[0042] Figure 4Schematic diagram of the process of the vehicle control method based on path planning provided in this application Figure 2 ;

[0043] Figure 5a A schematic diagram of a flow chart of another vehicle control method based on path planning provided in this application;

[0044] Figure 5b A simulation test diagram of the vehicle control method based on path planning provided in this application;

[0045] Figure 5c A comparison chart of the running times provided for this application;

[0046] Figure 5d A comparison chart of the curvature change rate provided for this application;

[0047] Figure 5e Comparison of driving trajectories provided for this application Figure 1 ;

[0048] Figure 5f Comparison of driving trajectories provided for this application Figure 2 ;

[0049] Figure 5g A comparison chart of the smoothing effect provided for this application;

[0050] Figure 5h Schematic diagram of the 0.5-meter burr noise smoothing effect provided for this application;

[0051] Figure 5i Schematic diagram of the 1-meter burr noise smoothing effect provided for this application;

[0052] Figure 5j A schematic diagram of the smoothing effect of a step noise point with a 0.5-meter offset provided in this application;

[0053] Figure 6 A schematic diagram of the structure of a vehicle control device based on path planning provided in this application;

[0054] Figure 7 A schematic diagram of the structure of the vehicle control device provided in this application.

[0055] The above drawings have shown clear embodiments of the present application, which will be described in more detail later. These drawings and text descriptions are not intended to limit the scope of the present application in any way, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0056] First, the terms involved in this application are explained:

[0057] Path planning is the process of calculating an optimal or satisfactory path from a starting point to a destination in a given environment. It is widely used in robot navigation, self-driving cars, drone flights and other fields to ensure that mobile entities can reach their destinations safely and efficiently.

[0058] A lane can refer to an area on a road that is specially divided for vehicle travel, usually clearly defined by markings (such as solid or dotted white or yellow lines). Each lane is designed to guide vehicles to travel in an orderly and safe manner and ensure smooth traffic flow. The parameters of different lane types are also different.

[0059] In the prior art, there are two important requirements for path planning in vehicle control methods: one is the requirement for the smoothness of the path curve, and the other is the requirement for the efficiency of path planning. Among them, the smoothness is directly related to the quality of the path curve; an unsmooth path curve may cause distortion in the lateral planning of the vehicle, and a sudden change in the curvature of the path curve will cause a sudden change in speed and severe braking in the longitudinal planning, ultimately resulting in an unsmooth driving performance, and may even require deceleration or parking. In terms of the requirement for path planning efficiency, if the path planning fails to complete the planning or planning fails in a short time, it will affect the normal driving of the vehicle, or even make it impossible to drive normally.

[0060] Figure 1 The schematic diagram of the principle of the existing method for determining the degree of curvature provided in the embodiment of the present application. Figure 1 As shown, in the path planning process, by selecting three consecutive points on the path (P i-1 , P i , P i+1 ) form a group to form an approximate parallelogram structure, and use the modulus of the diagonal vector to quantify the degree of curvature at each point. However, this method may cause the three points to be planned on a straight line, resulting in a broken line path when evaluating point by point. When a vehicle travels along such a path containing broken line segments, the driving performance may not be smooth enough, requiring frequent deceleration or stopping, which in turn affects the user's driving experience.

[0061] The vehicle control method, device, equipment, medium and product based on path planning provided in the embodiments of the present application can determine the degree of curvature by combining at least four adjacent lane points, so as to capture more complex curve features, and effectively avoid the collinearity problem that may occur when only three points are used, and reduce the risk of misjudging three points as straight lines, thereby ensuring that the path more truly reflects the actual shape and curvature of the road, and by combining and analyzing the curvature degrees of multiple groups of target lane points on the initial path and the constraints, it is ensured that the path follows a gradual and smooth trajectory when turning, rather than suddenly changing direction, thereby enabling the vehicle to achieve the effect of improving the driving stability of the vehicle and the driving experience of the user when driving.

[0062] Figure 2 A schematic diagram of a vehicle control method based on path planning provided in this application, such as Figure 2 As shown, the execution subject of a vehicle control method based on path planning provided in an embodiment of the present application may be a vehicle control system. Among them, the vehicle control system may be a server, and the server may be a device such as an on-board computer. This embodiment does not impose any special restrictions on the implementation method of the execution subject, as long as the execution subject can determine the initial lane path on the target lane based on the lane line data of the target lane; determine the target lane path of the target lane of the vehicle based on the degree of curvature between the target lane points on the initial lane path and the constraints of the vehicle driving on the target lane, and the target lane points are at least four lane points set adjacently along the initial lane path; when the vehicle enters the target road, control the vehicle to drive according to the target lane path.

[0063] The technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems are described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0064] Figure 3 Schematic diagram of the process of the vehicle control method based on path planning provided in this application Figure 1 ,like Figure 3 As shown, the method includes:

[0065] S301. Determine an initial lane path on the target lane according to lane line data of the target lane.

[0066] The target lane may refer to the lane where the vehicle is going to travel, and the lane may include part of the lane or all of the lane between the starting point and the end point of the vehicle.

[0067] Lane line data may refer to information about the geometry and position of a lane, and may include the centerline coordinates, width, and topological relationships of the lane. For example, lane line data may include: lane line type (such as solid line, dashed line, double solid line), color (such as white, yellow), width, length, and spacing (such as the specific size of dashed line segments and gaps), position (left, right, or center divider), and additional traffic markings (such as arrows, text, and symbols); in some embodiments, after determining the target lane, the lane line data of the target lane may be determined based on a high-precision map preload, or may be determined based on static lane data provided by the vehicle manufacturer.

[0068] Among them, the lane line data of the target lane determined by pre-loading the high-precision map can be: before the vehicle enters the target lane, the lane line data containing detailed lane line information (such as position, type, color and width) can be downloaded from the high-precision map according to the lane marking of the target lane. The high-precision map can be derived from road data collected by professional surveying and mapping vehicles through sensors such as lidar and cameras, combined with satellite remote sensing, aerial photography and official traffic information provided by the government, and the use of artificial intelligence technology to process and integrate these multi-source data. It is created and continuously updated by the map service provider to provide detailed lane lines, traffic signs and other accurate information.

[0069] Determining the lane line data of the target lane based on the static lane data provided by the vehicle manufacturer may refer to: the vehicle manufacturer may pre-install lane line data for a specific area or commonly used route when the vehicle leaves the factory, and this data is based on detailed road surveys and historical traffic information. When the vehicle is about to enter a known target lane, the lane line data of the target lane may be directly called and / or uploaded. The initial lane path may refer to a theoretical driving trajectory first planned based on the lane line data, and the initial lane path may reflect the centerline position and general direction of the target lane. The path may be composed of a series of points distributed along the lane, which is used to preliminarily define the ideal driving route of the vehicle within the lane.

[0070] In some embodiments, when a vehicle is about to enter the left fast lane of a highway, the vehicle control system can first extract the lane line coordinates (such as the position of the center line and the boundary line) of the lane from the high-precision map, and then generate a series of reference points evenly distributed along the center line of the lane through these coordinate points. For example, the vehicle control system can set a reference point every 10 meters and connect these points to form a smooth virtual path, which is the initial lane path.

[0071] Among them, in the embodiment of the present application, the method for determining the initial lane path on the target lane according to the lane line data of the target lane may include:

[0072] Determine lane centerline information of each lane on the target lane according to lane line data of the target lane;

[0073] According to the lane centerline information of each lane on the target lane, the centerlines of the lanes are spliced ​​to obtain the initial lane path on the target lane.

[0074] The lane centerline information may refer to a set of coordinate points of the lane centerline.

[0075] The splicing process may refer to a process of connecting multiple independent lane centerlines into a continuous path, and the splicing process may include processing operations involving removing overlapping parts and filling points.

[0076] After determining the lane centerline information of each lane on the target lane, the lane centerlines connected front and back on the lane can be spliced ​​according to the positional relationship of the lanes, so as to obtain the original reference line representing the initial lane path. In some embodiments, if there is an overlap between adjacent lane centerlines, the duplicate data points can be removed, and if the interval between adjacent lane centerlines is too large, points will be added between them to ensure the continuity of the path, so as to obtain the initial lane path.

[0077] For example, the coordinates of the lane centerline of the consecutive adjacent lanes A are A1, A2, ..., An; the coordinates of the lane centerline of lane B are B1, B2, ..., Bm. When the distance between An and B1 is less than the preset first distance, it is determined that An and B1 coincide, so that any coordinate point can be retained and connected between An-1 and B2. When the distance between An and B1 is greater than the preset second distance, C1, C2, ..., Ck can be inserted between An and B1, thereby obtaining the initial lane path.

[0078] S302. Determine a target lane path for the vehicle to travel in the target lane according to the degree of curvature between the target lane points on the initial lane path and the constraints on the vehicle traveling in the target lane, wherein the target lane points are at least four lane points adjacently arranged along the initial lane path.

[0079] The target lane point may refer to at least four adjacent lane points on the initial lane path, wherein the lane points on the initial lane path may be points where the initial lane path is divided according to a preset length. For example, when the path length of the initial lane path is 500m and the preset length is 10cm, there may be 5001 lane points on the initial lane path. When these points are D1, D2, ..., C5001, respectively, the target lane point may be D1 to D4, D2 to D5, ..., D4998 to D5001 of the 5001 lane points. The target lane point may also be D1 to D5, D2 to D6, ..., D4997 to D5001 of the 5001 lane points.

[0080] The degree of curvature between target lane points can refer to the change in the relative position between target lane points on the path. The degree of curvature is used to evaluate the curvature or the sharpness of the turning of the path. In other words, it can measure whether the path at a certain point forms an obvious turning point or turning. If the position of a target lane point changes greatly relative to the target lane points before and after it, it means that the path at that point is relatively curved; on the contrary, if the target lane points are almost on a straight line, it means that the path is relatively straight.

[0081] In this embodiment of the present application, the degree of curvature between the target lane points on the initial lane path is determined based on the curvature loss of the first lane point relative to the second lane point among the target lane points, and the second lane point is the target lane point adjacent to the first lane point on the initial lane path.

[0082] The first lane point is the middle lane point among at least four target lane points. For example, when there are four target lane points and the first lane point is lane point i, the second lane point may be lane point i-1, lane point i+1, lane point i+2; or lane point i-2, lane point i-1, lane point i+1. When there are five target lane points, the first lane point is the third lane point.

[0083] The curvature loss of the first lane point relative to the second lane point refers to the change in the curvature of the path at that point by comparing the geometric relationship between the first lane point and its adjacent second lane points before and after it in path planning. If the first lane point has a large deviation from the position of the adjacent second lane points before and after it, it indicates that the path has changed significantly or turned here, resulting in a higher curvature loss; on the contrary, if the first lane point almost forms a straight line with the two points before and after it, the curvature loss is low, indicating that the path is smoother.

[0084] The degree of curvature between target lane points can be characterized as the result of the curvature loss of the first lane point relative to the second lane point.

[0085] In this embodiment of the present application, the curvature loss is determined based on the positional relationship between the first lane point and the second lane point in the first direction and the second direction.

[0086] Among them, the first direction and the second direction can refer to the lateral and longitudinal directions of the path, and the directions can be used to describe the relative positions between lane points. The curvature loss is determined based on the positional relationship between the first lane point and the second lane point adjacent to it in the lateral direction (first direction) and the longitudinal direction (second direction). The distance change of the target lane point in the forward direction and the lateral offset change can be taken into account. By comprehensively analyzing the position differences in these two directions, the degree of geometric mutation or turning of the path at a certain lane point can be more accurately quantified.

[0087] Among them, in the embodiment of the present application, the curvature loss satisfies:

[0088] cost i =(x i+1 -3x i +3x i-1 +x i-2 ) 2 +(y i+1 -3y i +3y i-1 +y i-2 ) 2 ;

[0089] Among them, cost i Characterizes the curvature loss of the first lane point i; x i represents the position of the first lane point i in the first direction; x i+1 、x i-1 、x i-2 Respectively represent the positions of the second lane point i+1, the second lane point i-1, and the second lane point i-2 in the first direction; y i Represents the position of the first lane point i in the second direction; y i+1 ,y i-1 ,y i-2 Respectively represent the positions of the second lane point i+1, the second lane point i-1, and the second lane point i-2 in the second direction.

[0090] The constraints of a vehicle driving on a target lane can refer to various restrictions and requirements imposed on vehicle behavior in order to ensure safe, compliant and efficient driving. These conditions may include, but are not limited to, speed limits, minimum and maximum turning radius, lane change rules, safe distance from the vehicle in front, etc. The constraints can be obtained by integrating laws and regulations, traffic rules, vehicle physical characteristics, and lane environment data. For example, the vehicle control system can extract information such as speed limits and lane use rules from high-precision maps and traffic databases, and combine the vehicle's own performance parameters (such as minimum turning radius and acceleration capability) to obtain constraints.

[0091] In the embodiment of the present application, the vehicle control system can first analyze the degree of curvature between each point on the initial lane path, evaluate the degree of curvature of the path to characterize the curvature change, and identify sharp turns and flat sections. Based on the above information, combined with the constraints of vehicle driving (such as speed limit, minimum turning radius and safety distance), by constructing mathematical models such as constraint matrices, linear cost vectors and upper and lower bound constraint vectors, the optimal target lane path can be determined.

[0092] S303: When the vehicle enters the target road, the vehicle is controlled to travel according to the target lane path.

[0093] Among them, when the vehicle enters the target road, since the method for determining the target lane path in the embodiment of the present application is based on the fastest calculation method of the discrete point smoothing algorithm and introduces curvature smoothing rate loss, the entire smoothing process takes a very short time and can run stably under simulation. There will be no problem of being unable to complete the calculation within the specified time. The vehicle control system can directly generate the target lane path and drive according to the target lane path.

[0094] In some embodiments, if the target lane path is generated by a server, the vehicle control system may receive the target lane path of the target road planned by the server and control the vehicle to travel according to the target lane path.

[0095] In some embodiments, the target lane path can be displayed in the navigation software or on the vehicle's large screen to show the driver.

[0096] The vehicle control method based on path planning provided in the embodiment of the present application, after determining the initial path of the target lane, can more accurately predict and adjust the curvature change of the path by combining and analyzing the curvature of multiple groups of target lane points on the initial path and the constraints of the vehicle's driving such as speed, acceleration and steering angle limits, thereby ensuring that the vehicle follows a gradual and smooth trajectory when turning, rather than suddenly changing direction to optimize the path. This allows the vehicle to travel stably according to the trajectory of the target lane path when entering the target road, thereby reducing unnecessary shaking and sharp turns, thereby improving the vehicle's driving stability and enhancing the user's driving experience.

[0097] Figure 4 Schematic diagram of the process of the vehicle control method based on path planning provided in this application Figure 2 ,like Figure 4 As shown, in this embodiment Figure 3 Based on the embodiment, the step of determining the target lane path of the target lane of the vehicle according to the curvature between the target lane points on the initial lane path and the constraint condition of the vehicle driving on the target lane in the vehicle control method based on path planning is described in detail. The method includes:

[0098] S401. A target matrix is ​​obtained according to the curvature between target lane points on the initial lane path, wherein the rows of the target matrix represent the curvature of the target lane points in a first direction, the columns of the target matrix represent the curvature of the target lane points in a second direction, and the elements in the target matrix represent the curvature of the target lane points.

[0099] Among them, the row of the target matrix represents the curvature of the target lane point in the first direction, which can be referred to as (x i+1 -3xi +3x i-1 +x i-2 ) 2 The row of the target matrix represents the curvature of the target lane point in the second direction, which can be referred to as (y i+1 -3y i +3y i-1 +y i-2 ) 2 The elements in the target matrix represent the curvature of the target lane point and can be referred to as (x i+1 -3x i +3x i-1 +x i-2 ) 2 +(y i+1 -3y i +3y i-1 +y i-2 ) 2 For example, the target matrix can be shown in Table 1:

[0100]

[0101]

[0102] Among them, the first row in the target matrix can include data of target lane point i-1, target lane point i, target lane point i+1, and target lane point i+2, and the second row in the target matrix can include data of target lane point i, target lane point i+1, target lane point i+2, and target lane point i+3.

[0103] In some embodiments, the target matrix can be a symmetric matrix (P matrix) represented as a quadratic programming (QP) used to define the coefficients of the quadratic terms in the objective function. In autonomous driving path planning, the P matrix can represent the curvature loss or other geometric characteristics between points on the path.

[0104] S402: Determine a constraint matrix, a linear cost vector, an upper constraint vector, and a lower constraint vector according to the constraint conditions of the vehicle traveling on the target lane.

[0105] Among them, the constraint matrix (A matrix) can characterize the relationship between the linear constraints in the constraints. Each row vector in the constraint matrix can represent a constraint, and the column vector corresponds to the decision variable (such as speed, acceleration, steering angle, etc.). The constraint matrix ensures that the vehicle's behavior complies with traffic regulations and physical limitations, such as speed limits, minimum turning radius, and safe distance. The linear cost vector (Q vector) can be used to describe the cost or priority of each decision variable, reflecting the linear terms that need to be minimized during the optimization process. The linear cost vector can be used to determine which behaviors are more favorable, such as a smoother path or lower speed changes.

[0106] The upper constraint vector (U vector) can be used to define the maximum allowed value for each decision variable. For example, the maximum speed, maximum acceleration, or maximum steering angle. These upper limits ensure that the vehicle does not exceed its physical capabilities or violate traffic laws. The lower constraint vector (I vector) corresponds to the upper constraint vector and can be used to define the minimum allowed value for each decision variable. For example, the minimum speed, minimum acceleration, or minimum steering angle. These lower limits also ensure the safety and legality of the vehicle's behavior.

[0107] For example, when the constraints are: maximum speed Vmax = 30m / s, minimum speed Vmin = 5m / s, maximum acceleration a max =2m / s 2 , minimum acceleration a max =2m / s 2 hour,

[0108] The constraint matrix can be a 4*2 matrix, where each row of the constraint matrix corresponds to a constraint condition, the first column represents the speed, and the elements of the first column can be 1, -1, 0, 0 in sequence, and the second column represents the acceleration, and the elements of the second column can be 0, 0, 1, -1 in sequence. The linear cost vector can be a 2*1 matrix, and the elements in the matrix can be 0, 1 in sequence. The upper bound constraint vector can be a 4*1 matrix, and the elements in the matrix can be 30, -5, 2, -2 in sequence; the lower bound constraint vector can be a 4*1 matrix, and the elements in the matrix can be 5, 30, -2, -2 in sequence.

[0109] S403, performing format adjustment processing on the target matrix to obtain an initial sparse matrix table;

[0110] S404, determining a weight coefficient corresponding to a non-zero element in the initial sparse matrix table, where the weight coefficient includes at least one of a curvature loss weight, a total path length loss weight, and an offset loss weight;

[0111] S405 . Adjust the initial sparse matrix table according to the weight coefficients corresponding to the non-zero elements in the initial compression matrix to obtain a sparse matrix table.

[0112] The sparse matrix table may refer to a matrix table constructed by P matrix coefficients under a new curvature loss expressed in a CSC (Compressed Sparse Column) compression format. Sparse matrices are used to efficiently store and process matrices with only a few non-zero values ​​in a large number of elements. The use of sparse matrix tables can facilitate the optimization calculation of the optimization solver in subsequent steps.

[0113] The curvature loss weight can be used to quantify the impact of the curvature of each point on the path on the optimization target. A higher curvature loss weight means that the system is more inclined to choose a smoother path with fewer turns to reduce the discomfort and potential risks caused by sharp turns. By adjusting this weight, it can be ensured that the vehicle moves along a relatively straight road as much as possible during driving, improving driving safety and comfort.

[0114] The total path length loss weight reflects the importance of path length to the optimization goal. A higher total path length loss weight will cause the vehicle control system to prioritize shorter paths to minimize driving distance and time. By adjusting this weight, the driving path can be shortened as much as possible while meeting other constraints.

[0115] The deviation loss weight can be used to measure the impact of the degree to which the vehicle deviates from the ideal lane centerline on the optimization target. A higher deviation loss weight prompts the system to keep the vehicle in the center of the lane as much as possible, avoiding frequent lane changes or lane deviations, thereby enhancing driving stability and safety.

[0116] In an embodiment of the present application, the format adjustment process of the target matrix to obtain the initial sparse matrix table may include: scanning the entire target matrix, identifying the positions of all non-zero elements and their corresponding values, wherein, since the matrices in most practical applications (such as the P matrix in path planning) are sparse, that is, most of the elements are zero, the storage and calculation burden can be significantly reduced; selecting the corresponding sparse matrix format, such as for path planning problems, a sparse matrix in CSC (Compressed Sparse Column) format can be selected. Create a sparse matrix structure and store non-zero elements and their position information.

[0117] In some embodiments, the target matrix is ​​also formatted and the method of obtaining the initial sparse matrix table can also be performed manually, that is, the target matrix is ​​displayed, and the initial sparse matrix table is obtained in response to the user's operation based on the target matrix.

[0118] The weight coefficient may be a preset weight coefficient. After obtaining the sparse matrix table, the initial sparse matrix table may be adjusted in response to the user's configuration operation to obtain the sparse matrix table. For example, the sparse matrix table may be as shown in Table 2:

[0119]

[0120]

[0121] S406: Input the sparse matrix table, the constraint matrix, the linear cost vector, the upper bound constraint vector and the lower bound constraint vector into the optimization solver to obtain a lane path of the target lane for the vehicle to travel.

[0122] Among them, the optimization solver can refer to a software tool or algorithm used to solve mathematical optimization problems, aiming to find the best solution that satisfies given constraints. In autonomous driving path planning, the optimization solver can receive input parameters such as sparse matrix tables, constraint matrices, linear cost vectors, upper bound constraint vectors, and lower bound constraint vectors, and solve the optimal path through efficient algorithms (such as OSQP, Gurobi, CPLEX, etc.).

[0123] In an embodiment of the present application, a sparse matrix table, a constraint matrix, a linear cost vector, an upper bound constraint vector and a lower bound constraint vector may be input into an optimization solver to solve an optimal path, thereby obtaining a lane path for the target lane of the vehicle.

[0124] The vehicle control method based on path planning provided in the embodiment of the present application achieves accurate, safe and efficient vehicle driving path planning by means of efficient sparse matrix processing and optimized solver application.

[0125] Figure 5a A flow chart of another vehicle control method based on path planning provided in this application is as follows: Figure 5a As shown, the method includes:

[0126] Obtain anchor points (lane points) by interpolation from reference points; obtain a list of anchor points; set anchor point boundaries; smooth splicing; trim historical trajectory boundaries and fix anchor points; perform a smoothing algorithm on the anchor points after trimming historical trajectory boundaries and fixing anchor points to obtain a reference line copy representing the planned path. If the validity of the reference line copy meets the validity requirements, control the vehicle based on the reference line copy.

[0127] Among them, in the embodiment of the present application, it is possible to first 1) determine the lane that the vehicle will pass through, usually by temporarily saving the data in the form of a lane topology number; 2) query the lane structure in the local map according to the lane topology number, obtain the lane centerline geometry points and segmented lane width list of the corresponding lane, and verify the front and rear topological relationship of the lane at the same time; 3) splice the front and rear connected lane centerlines, and the splicing process may involve additional work such as removing the overlapping parts of the adjacent lane centerlines and filling in the points when the interval between the adjacent lane centerlines is too large, to form an original reference line; 4) smooth and encrypt the points of the original reference line, and after processing, obtain a reference line that is smooth, sufficiently dense, and of a length that meets driving requirements that can be used for planning, and control the vehicle to drive according to the reference line.

[0128] Among them, the process of executing the smoothing algorithm may include: reading smoother configuration parameters; calculating P matrix, A matrix, Q vector, U vector and I vector; setting OSQP solver configuration according to P matrix, A matrix, Q vector, U vector and I vector to perform OSQP solution; judging the solution state, and obtaining the solution result after the solution state meets the requirements, and if the solution result is regularized, calculating the reference line contour to generate a reference line copy.

[0129] That is, the anchor points at the corresponding positions can be sampled from the reference line according to the S coordinate value and a list of anchor points can be constructed; then, the anchor point boundaries are set. For the case of curve splicing, since the historical trajectory cannot be moved, the trajectory boundaries are additionally corrected, the left and right boundary movement ranges are set to 0 (the floating point value is less than a very small positive number), and hard constraint labels are added. Then, these data are input into the optimization algorithm based on OSQP. The algorithm reads the configuration parameters internally and calculates the required matrices and vectors such as P, A, q, u, l, etc., to construct a set of equations for the quadratic optimization problem consisting of losses and constraints, where the P matrix is ​​the target matrix in the embodiment of the present application. After calling OSQP for solution, the solution state is analyzed and different abnormal states are processed; for successful solutions, the result parameters are extracted, re-regularized into a two-dimensional shape point set, and a new reference line is recalculated based on the optimized shape points to distinguish it from the reference line before the original smooth optimization. Finally, the smoothed reference line is judged for rationality, such as whether it is far away from the original reference line or causes part of the vehicle body to exceed the original lane. After verification, it is output to the downstream system for specific real-time planning processing. Figure 5b The simulation test diagram of the vehicle control method based on path planning provided in this application is as follows: Figure 5b As shown, in the simulation test map, the starting point of each planning is fixed at a position below the navigation path, and the end point is on the ramp. The simulation route includes sections of different curvatures such as main road straights, main road curves, and ramp curves, and the original lanes of the map have multiple alternating connections of "straight-curved" and "curved-straight" sections.

[0130] Table 3 is a comparison table of subjective and objective indicators of different smoothing algorithms provided by this application, as shown in Table 3:

[0131]

[0132]

[0133] Among them, the average running time of the FemPos-ipopt method, CosTheta method, and Spiral method in Table 3 has exceeded the total time of the overall functional process of the planning algorithm in a general sense, and the comparative experiment is a simulation experiment conducted on an ordinary desktop platform. The actual on-board domain controller computing power will further decrease, so it is unacceptable from the perspective of computational efficiency. The problem reflected in the actual simulation process is the statistical situation of the rear planning timeout. The planning curve cannot be given in a short time, resulting in line flashing or missing lines. Generally, the ramp entrance has more complex lane curves and more road sections, which will lead to more serious smoothing timeout problems; secondly, in the algorithms FemPos-sqp method, FemPos-osqp method, QpSpline method, and FemPos-osqp-dkappa with acceptable computational efficiency, the average time used by the method in the embodiment of the present application and the FemPos-osqp method is less than 3 milliseconds, and the effect is the best.

[0134] Secondly, for the smoothness of the curve, it is mainly analyzed through indicators such as curvature change and planning starting point oscillation: in terms of the average curvature change rate, the method in the embodiment of the present application has greatly improved stability, and the maximum curvature change is reduced by nearly 70% (from 0.022 to 0.007) compared with the FemPos-osqp method, which also corresponds to a significant improvement in the curvature jump, greatly reducing the abnormal reduction of the maximum speed of the vehicle due to the curvature jump during normal driving, thereby causing the vehicle to decelerate abnormally or even stop abnormally. The lateral acceleration at the planning starting point is reduced by nearly 30% (from 0.004573149 to 0.003626685), which means that during normal driving along the lane, due to the influence of lane curvature, historical lane change trajectory, etc., the vehicle can enter the stable state of driving along the reference line faster from the unstable state close to the reference line, and reduce oscillation and overshoot. In terms of simulation effect, it is difficult to observe the left and right oscillations during driving.

[0135] For the method and FemPos-osqp method in the embodiments of the present application, Figure 5c The running time comparison chart provided for this application, Figure 5d The curvature change rate comparison diagram provided for this application is as follows: Figure 5c and Figure 5dAs shown, although the running speed of the method in the embodiment of the present application is slower than that of the FemPos-osqp method, the outliers are much smaller than those of the FemPos-osqp method. Therefore, it can be shown that the introduction of the loss of the curvature change rate can successfully avoid the problem of curvature jump caused by the reference line not being smooth enough, resulting in abnormal deceleration and stopping of the vehicle.

[0136] Figure 5e Comparison of driving trajectories provided for this application Figure 1 , Figure 5f Comparison of driving trajectories provided for this application Figure 2 ,like Figure 5e and Figure 5f As shown, since the driving routes are consistent and the driving time is similar, it can be considered that in the same frame, during the comparative experiment of the method in the embodiment of the present application and the FemPos-osqp method, the simulated vehicle is in a similar position and the driving state is also similar, which is convenient for comparison of control variables. Here, at the connection point from the straight road to the curve, the difference between the two methods under the extreme case of comparison can be determined, wherein the left side of the figure is a reference line (blue line) and a planning trajectory (black dot) drawing, the coordinates are the self-vehicle coordinate system, the origin (0, 0) represents the self-vehicle position, the upper right side is vt and st two longitudinal planning drawings, and the bottom is a lateral planning candidate trajectory map, mainly comparing the position of the left reference line: FemPos-osqp method, the smoothing process of the front and rear frames will cause a large reference line offset, which is greatly offset from the left side of the vehicle to the right side; the method in the embodiment of the present application is at the connection point where the straight road changes to the curve, and there is also a change from left to right, but the overall change of the reference line is small, and it is on the same side of the vehicle in a short time.

[0137] Figure 5g This is a comparison diagram of the smoothing effect provided in the present application. When comparing the smoothing effect before and after the improvement, it can be found that the smoothed curve in the method of the embodiment of the present application is closer to the original lane splicing curve in the case of a curved lane, while the FemPos-osqp method may obtain a curve with a large distance difference after smoothing, such as Figure 5g As shown, the red line is the spliced ​​original curve, the yellow line is the optimization algorithm output of the FemPos-osqp method, and the blue line is the optimization output of the method of the embodiment of the present application. It can be seen that the smoothed curve in the method of the embodiment of the present application is closer to the original curve. In this case, if there are noise points in the original curve, even if the smoothing is significantly improved compared to the algorithm before the improvement, it is impossible to absolutely determine that its filtering effect meets the requirements, so a separate supplementary experiment is required. Figure 5h This is a schematic diagram of the 0.5-meter burr noise smoothing effect provided for this application. Figure 5i This is a schematic diagram of the 1-meter burr noise smoothing effect provided for this application. Figure 5jThe schematic diagram of the smoothing effect of the step noise with a 0.5-meter offset provided in this application is as follows: Figure 5h , Figure 5i and Figure 5j As shown, the black dots are noise points added by manual modification, the red curve is the reference line shape before smoothing after the introduction of noise points, and the blue is the reference line shape after smoothing. It can be seen from the figure that under the current interference of the three types of noise points, the method of the embodiment of the present application can still meet the driving requirements.

[0138] Therefore, another vehicle control method based on path planning provided by the embodiment of the present application has an extremely fast calculation speed of the smoothing algorithm executed. Through the curvature smoothing rate loss, it is ensured that the entire smoothing process takes a very short time and runs stably, avoiding the problem of being unable to complete the calculation within the specified time; secondly, the curvature of different points is adjusted to an average consistent trend through the curvature change rate loss, and the proportion of different loss weights is dynamically adjusted, which can be flexibly optimized after the code is compiled. In particular, when dealing with problems such as vehicle lane changes and reference line splicing, the curvature jump at the connection point is successfully avoided, and the problem of abnormal vehicle deceleration caused by sudden changes in curvature is eliminated, providing a better reference line for the planning process, thereby greatly improving the planning effect and driving stability.

[0139] Figure 6 A schematic diagram of the structure of a vehicle control device based on path planning provided in this application, such as Figure 6 As shown, the vehicle control device 60 based on path planning provided in this embodiment includes:

[0140] A determination module 601 is used to determine an initial lane path on the target lane according to the lane line data of the target lane;

[0141] A planning module 602 is used to determine a target lane path for the vehicle to travel in the target lane according to the curvature between the target lane points on the initial lane path and the constraint conditions for the vehicle to travel in the target lane, wherein the target lane points are at least four lane points adjacently arranged along the initial lane path;

[0142] The control module 603 is used to control the vehicle to travel according to the target lane path when the vehicle enters the target road.

[0143] In a possible implementation, the determination module 601 may be specifically configured to:

[0144] Determine lane centerline information of each lane on the target lane according to lane line data of the target lane;

[0145] According to the lane centerline information of each lane on the target lane, the centerlines of the lanes are spliced ​​to obtain the initial lane path on the target lane.

[0146] In a possible implementation, the planning module 602 may be specifically configured to:

[0147] The curvature between the target lane points on the initial lane path is determined according to the curvature loss of the first lane point relative to the second lane point among the target lane points, and the second lane point is the target lane point adjacent to the first lane point in front and behind on the initial lane path.

[0148] In a possible implementation, the planning module 602 may be specifically configured to:

[0149] The curvature loss is determined according to a positional relationship between the first lane point and the second lane point in the first direction and a positional relationship in the second direction.

[0150] In a possible implementation, the planning module 602 may be specifically configured to: the curvature loss satisfies:

[0151] cost i =(x i+1 -3x i +3x i-1 +x i-2 ) 2 +(y i+1 -3y i +3y i-1 +y i-2 ) 2 ;

[0152] Among them, cost i Characterizes the curvature loss of the first lane point i; x i represents the position of the first lane point i in the first direction; x i+1 、x i-1 、x i-2 Respectively represent the positions of the second lane point i+1, the second lane point i-1, and the second lane point i-2 in the first direction; y i Represents the position of the first lane point i in the second direction; y i+1 ,y i-1 ,y i-2 Respectively represent the positions of the second lane point i+1, the second lane point i-1, and the second lane point i-2 in the second direction.

[0153] In a possible implementation, the planning module 602 may be specifically configured to:

[0154] According to the curvature between the target lane points on the initial lane path, a target matrix is ​​obtained, wherein the rows of the target matrix represent the curvature of the target lane points in the first direction, the columns of the target matrix represent the curvature of the target lane points in the second direction, and the elements in the target matrix represent the curvature of the target lane points;

[0155] According to the constraint conditions of the vehicle driving on the target lane, a constraint matrix, a linear cost vector, an upper constraint vector and a lower constraint vector are determined;

[0156] A lane path of a target lane for the vehicle is determined according to the target matrix, the constraint matrix, the linear cost vector, the upper bound constraint vector and the lower bound constraint vector.

[0157] In a possible implementation, the planning module 602 may be specifically configured to:

[0158] Adjust the format of the target matrix to obtain an initial sparse matrix table;

[0159] Determine a weight coefficient corresponding to a non-zero element in the initial sparse matrix table, the weight coefficient comprising at least one of a curvature loss weight, a total path length loss weight, and an offset loss weight;

[0160] According to the weight coefficients corresponding to the non-zero elements in the initial compression matrix, the initial sparse matrix table is adjusted to obtain a sparse matrix table;

[0161] The sparse matrix table, the constraint matrix, the linear cost vector, the upper bound constraint vector and the lower bound constraint vector are input into the optimization solver to obtain the lane path of the target lane of the vehicle.

[0162] In a possible implementation, the planning module 602 may be specifically configured to:

[0163] The lane line data of the target lane is obtained from the reference map.

[0164] The path planning-based vehicle control device provided in this embodiment can execute the method provided in the above method embodiment. Its implementation principle and technical effect are similar, and this embodiment will not be described in detail here.

[0165] Figure 7 This is a schematic diagram of the structure of the vehicle control device provided in this application. Figure 7 As shown, the vehicle control device 70 provided in this embodiment includes: at least one processor 701 and a memory 702. Optionally, the vehicle control device 70 also includes a communication component 703. The processor 701, the memory 702 and the communication component 703 are connected via a bus 704.

[0166] In a specific implementation process, at least one processor 701 executes the computer-executable instructions stored in the memory 702, so that at least one processor 701 executes the above method.

[0167] The specific implementation process of the processor 701 can be found in the above method embodiment, and its implementation principle and technical effect are similar, so this embodiment will not be repeated here.

[0168] In the above embodiments, it should be understood that the processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), etc. A general-purpose processor may be a microprocessor or any conventional processor. The steps of the method disclosed in the invention may be directly implemented as being executed by a hardware processor, or may be executed by a combination of hardware and software modules in the processor.

[0169] The memory may include a high-speed memory (Random Access Memory, RAM), and may also include a non-volatile memory (Non-volatile Memory, NVM), such as at least one disk memory.

[0170] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, the bus in the drawings of the present application is not limited to only one bus or one type of bus.

[0171] The present application also provides a computer program product, including a computer program, which implements the above method when executed by a processor.

[0172] The present application also provides a computer-readable storage medium, in which computer-executable instructions are stored. When a processor executes the computer-executable instructions, the above method is implemented.

[0173] The above-mentioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk. The readable storage medium can be any available medium that can be accessed by a general or special-purpose computer.

[0174] An exemplary readable storage medium is coupled to a processor so that the processor can read information from the readable storage medium and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can be located in an application specific integrated circuit (Application Specific Integrated Circuits, referred to as: ASIC). Of course, the processor and the readable storage medium can also exist in the device as discrete components.

[0175] The division of units is only a logical function division, and there may be other divisions in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interface, device or unit, which can be electrical, mechanical or other forms.

[0176] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0177] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0178] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the 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, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0179] Those skilled in the art can understand that all or part of the steps of implementing the above-mentioned method embodiments can be completed by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, the steps of the above-mentioned method embodiments are executed; and the aforementioned storage medium includes: ROM, RAM, disk or optical disk and other media that can store program codes.

Claims

1. A vehicle control method based on path planning, characterized in that: include: Determining an initial lane path on the target lane according to lane line data of the target lane; Determining a target lane path for the vehicle to travel in the target lane according to the curvature between the target lane points on the initial lane path and the constraint condition for the vehicle to travel in the target lane, wherein the target lane points are at least four lane points adjacently arranged along the initial lane path; When the vehicle enters a target road, the vehicle is controlled to travel according to the target lane path.

2. The method according to claim 1, characterized in that The determining, based on the lane line data of the target lane, an initial lane path on the target lane comprises: Determining lane centerline information of each lane on the target lane according to the lane line data of the target lane; According to lane centerline information of each lane on the target lane, the centerlines of the lanes are spliced ​​to obtain an initial lane path on the target lane.

3. The method according to claim 1, characterized in that The degree of curvature between the target lane points on the initial lane path is determined based on the curvature loss of a first lane point relative to a second lane point among the target lane points, where the second lane point is a target lane point adjacent to the first lane point in front and behind on the initial lane path.

4. The method according to claim 3, characterized in that The curvature loss is determined according to a positional relationship between the first lane point and the second lane point in a first direction and a positional relationship in a second direction.

5. The method according to any one of claims 3 to 4, characterized in that: The curvature loss satisfies: cost i =(x i+1 -3x i +3x i-1 +x i-2 ) 2 +(y i+1 -3y i +3y i-1 +y i-2 ) 2 ; Among them, the cost i Characterizes the curvature loss of the first lane point i; the x i represents the position of the first lane point i in the first direction; the x i+1 , the x i-1 , the x i-2 Respectively represent the positions of the second lane point i+1, the second lane point i-1, and the second lane point i-2 in the first direction; the y i Characterizes the position of the first lane point i in the second direction; the y i+1 , the y i-1 , the y i-2 Respectively represent the positions of the second lane point i+1, the second lane point i-1, and the second lane point i-2 in the second direction.

6. The method according to claim 1, characterized in that The step of determining a target lane path for the vehicle to travel in the target lane according to the curvature between the target lane points on the initial lane path and the constraint condition for the vehicle to travel in the target lane comprises: According to the curvature between the target lane points on the initial lane path, a target matrix is ​​obtained, wherein the rows of the target matrix represent the curvature of the target lane points in a first direction, the columns of the target matrix represent the curvature of the target lane points in a second direction, and the elements in the target matrix represent the curvature of the target lane points; Determining a constraint matrix, a linear cost vector, an upper constraint vector, and a lower constraint vector according to a constraint condition for the vehicle to travel on the target lane; A lane path for the vehicle to travel in the target lane is determined according to the target matrix, the constraint matrix, the linear cost vector, the upper bound constraint vector, and the lower bound constraint vector.

7. The method according to claim 6, characterized in that The determining, according to the target matrix, the constraint matrix, the linear cost vector, the upper bound constraint vector and the lower bound constraint vector, a lane path for the vehicle to travel in the target lane comprises: Performing format adjustment processing on the target matrix to obtain an initial sparse matrix table; Determine a weight coefficient corresponding to a non-zero element in the initial sparse matrix table, wherein the weight coefficient includes at least one of a curvature loss weight, a total path length loss weight, and an offset loss weight; Adjusting the initial sparse matrix table according to weight coefficients corresponding to non-zero elements in the initial compression matrix to obtain a sparse matrix table; The sparse matrix table, the constraint matrix, the linear cost vector, the upper bound constraint vector and the lower bound constraint vector are input into an optimization solver to obtain a lane path for the vehicle to travel in the target lane.

8. The method according to any one of claims 1 to 7, characterized in that The lane line data of the target lane is data obtained from a reference map.

9. A vehicle control device based on path planning, characterized in that: include: A determination module, configured to determine an initial lane path on the target lane according to lane line data of the target lane; A planning module, configured to determine a target lane path for the vehicle to travel in the target lane according to a degree of curvature between target lane points on the initial lane path and a constraint condition for the vehicle to travel in the target lane, wherein the target lane points are at least four lane points adjacently arranged along the initial lane path; The control module is used to control the vehicle to travel according to the target lane path when the vehicle enters the target road.

10. A vehicle control device, characterized in that: include: Memory, processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory, so that the processor performs the method according to any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, which are used to implement the method according to any one of claims 1 to 8 when executed by a processor.

12. A computer program product, characterized in that The invention comprises a computer program, which implements the method according to any one of claims 1 to 8 when being executed by a processor.

Images